If you have a solid understanding of kinematics and fourier transforms it makes this stuff easier to grasp, but that is honestly more esoteric than what most people need to actually apply this in practice. This is BY FAR one of the best intuitive explanations of how phase works with sound reinforcement that I have ever come across. It's something I've understood but always had a hard time explaining in detail because it's such a math heavy topic to really get in to. Great job bridging that gap! Would love to see more.
Inexpensive marginal test gear has released a tidal wave of inaccurate, maybe, and outright made-up-shit from folks focused more on ad revenue than knowing what the phuk they are spewing.
He really isnt. You just need to broaden your horizons a bit. the videos by xiph are superb. there are many tacking it at the level of PCB design for audio as well.
@@jeremyglover5541you're not wrong, but it's also kind of telling that your example is xiph, because Monty's videos are uniquely the best resource on their topics despite being around 15 years old now.
I believe there's a term called Jerk when decribes the rate of cahnge of acceleration. Also wonder how damping factor and power bandwidth comes into play. Awesome video, thanks!
Some other terms I may have made up - bounce (sharktooth), ramp (sawtooth), step (square wave), impulse, doublet (N-wave), jerk, etc. Not many are taught the elementary waveforms.
We may be jerks, but we also have a sense of humor, the next three derivatives are snap, crackle, and pop. Although, back when I was playing with this stuff, "jounce" was the derivative of jerk.
That was one of the best descriptions I've ever seen for the behaviors of the "signal, driver, soundwave, mic" relationship. It's a very complicated process, and this is a great way to break it down and understand each component. For your next challenge, you should try to do the same type of breakdown for a bass-reflex cabinet. That is one of the most important and least understood elements of speaker/cabinet design. I've been struggling to wrap my head around it for ages.
Wait, that was 30 minutes? I just restarted the video to read the comments. I was too busy soaking in the info to read comments or notice half an hour passed on my first time through
Amazing! Amazing amazing. This is the kind of content I've always wished existed. Now we just need a pipeline for absolute beginners to get to this point
The phase relationship between the input signal and the output from the speakers is actually much more complicated in practice. The reason is that the equation m.a+c.v+k.x=F(x) applies to established sinusoidal stimuli. In practice, the musical signal is not like that. It is rather a transient process in which the behavior of the speaker is determined by an inhomogeneous system of differential equations, which makes the issue of instantaneous phase shift quite complicated. Thank a lot for video.
There's a small mistake on the "AC coupled" waveform drawing (graph on the right) at 1:55 , after the "DC" portion of the signal, the AC coupled signal should not have any positive slope since there is no positive slope in the original signal. (like when the DC battery is released at 2:12) Thanks for the nice thorough video again :)
As the first waveform is passed intact, this is not a large enough time snippet to say if the flat section is even DC or just a bit of lower frequency out of context. Most likely the flat "DC" bit would come through DC blocking with a little slope and a slow rebound if the pulse is not followed shortly by a matching negative excursion.
In the near-field of a dipole, which is where you're measuring with your microphone, the pressure and the velocity are 180 degrees out of phase. The driver cone determines the air velocity and the microphone picks up the pressure field. That's where your phase shift at 14:00 comes from.
Could you, please, decipher your thought for us? The wave of 20Hz is about 55 feet long.How will a microphone position offset by a feet or two affect the result?
I don't know much about audio engineering, but I believe he's saying the driver creates an inverse pressure wave with an in phase velocity. Imagine that the biggest change in pressure doesn't happen when the speaker goes from rest/neutral to outermost, but actually when it goes from outermost to innermost positions. The hz doesn't so much matter as the speed of sound itself, too. A very low frequency wouldn't be captured on a mic further away, unless the velocity was astronomical. Edit: I realized right after posting that my explanation would result in 90 degrees in shift, and also I had paused right before he started explaining velocity being pressure's derivative, which isn't what this poster was talking about, I don't think. The video's explanation at around 24:00 sums it up, no?
@@LuxLucidOfficial Yes, the explanation at 23-25m is a good summary how the air pressure goes in the near field. My question was about "The driver determines cone velocity and the mic picks up the pressure.." To me it sounded like the driver throws apples and the mic picks up oranges, so what's the rationale here?
It is not a dipole. It is a baffled driver. Pressure and velocity are not generally 180 degrees out of phase, since that will depend on the acoustic environment.
@@Rene_Christensen In the very near-field it doesn't matter if it is baffled or not. The moving membrane displaces the air in a way that requires a dominant dipole moment of the multipole series. And the near-field does not know anything about the acoustic environment, by definition. Check out the derivation of the velocity potential of an oscillating solid sphere, in which case you get a pure dipole. The pressure is the negative time derivative of the velocity potential and the velocity of the object coincides with the field velocity at the surface, making both the spatial derivative of the velocity potential. You can then determine yourself that they are exactly pi out of phase near the membrane. Only at a distance that is significantly greater than the membrane radius you will start to see a significant deviation from this relationship. You can find the relevant derivations here for example: www.math.fsu.edu/~hju/cht12.htm
Love it when i can feel like i understand a complex subject without actually understanding the any of it 😅. But in all honesty, you did an excellent job of explaining things i know little about without losing me and the length of the video perfectly pushed my brain all the way out to its limits!
When i was at university we learned that particle velocity and sound pressure are out of phase in the nearfield and in phase in the far field. And considering that the Driver stimulate the particles directly, resulting in particle velocity and the microphone converting sound pressure to electrical current, your results would make sense (basically also what you explained at 23:55). It would be interesting to see the same measurements in the far field.
I can already tell you what happens in the Farfield because we did do those measurements, they just didn’t make it in this video. What happens is that the mic measurement remains +180 degrees out of polarity with the laser but the mic actually gets slightly ahead of the laser by about 1’ or so. This shift happens very quickly, within a few feet of the driver, and remains consistent all the way out to 200’ and applies at all frequencies. We realized this has to do with the fact that the acoustic center of the driver is actually about 1 foot in front of it. Might make a whole video on this topic.
@@devinlsheets_alphasound That would be great! Because heard and read alot about phase effects of drivers and sound but never really understood what it all means physically. I love how your experiment does exactly that: what is moving and how. Maybe you could even try to visualize how the air moves with dry ice or small plastic balls on strings? just a thought
@@devinlsheets_alphasound "We realized this has to do with the fact that the acoustic center of the driver is actually about 1 foot in front of it" Yes, it varies for different cabinets. Would be nice to do the measurements to verify (by multiple independent entities!!!) the model(s)/math.
@@devinlsheets_alphasound That is as it should be. For somewhat free-field conditions (I realize you are in a room, but still), the pressure is in-phase with acceleration of the piston (not the acceleration where the pressure is measured!, for which in the far-field the velocity indeed is in-phase with pressure for pure propagation, as mentioned in some posts here), because the acoustic environment seen from the piston is mass-like at lower frequencies for a typical cone size. It is not quite a perfect match between piston acceleration and pressure, but for example for a flat 18 inch piston in a baffle at 1 m distance, the two are only 5 degrees out of phase at 300 Hz and even closer at lower frequencies. With acceleration being 180 degrees out of phase with displacement, what you see in your measurements is what you should see. I should say that the phase coming the distance travelled from the piston to the microphone (linear phase) needs to be removed in these calculations, so in your measurement setup, you need to either input this distance in some setting, or manually get rid of it.
In a car, what pushes you (pressure) back in the seat? Acceleration. Not speed and not position. Similarly the air in front of the speaker cone becomes pressurised by the acceleration of the cone, not the speed or the position. This should also explain why high frequencies have more acoustic power thab lows at the same driver displacment - acceleration is higher. EDIT: In hindsight I misspoke. Pressure is not created by acceleration, but sound (modulation of pressure) is. Thus g-force felt in a car is analogous to sound energy.
No, pressure is caused by the motion of the cone. By your logic, a skydiver would feel no air pressure once he reached terminal velocity. you are confusing the potential energy of sitting in a car seat at constant velocity, with the kinetic energy of a column of air being moved by a cone.
@joshua43214 You make a great point! I don't think it's a perfect analogy. However I would note that sound is not a constant pressure (if the cone were moving at a constant velocity) because that is DC. Sound is changes in pressure (cone changing velocity or accel / decel) and in truth this is a continuous cycle of conversion between potential energy (compressed / rarefied medium) and kinetic energy (medium in motion expanding / contracting). In the case of a sky diver the medium is in motion past the body (wind) but sound is more like the force felt by the sky diver who will indeed feel weightless at terminal velocity. Your comment made me think hard, thanks!
@@Audio_Simon I think you have mistaken the skydiver example, the diver doesn't feel weightlessness at terminal velocity. It's only just at the moment of jumping. At terminal velocity, you feel your weight supported by the air.
@tusharjamwal You are quite right that air will resist the motion of the falling body in the form of drag. That said, from an object in free-fall is generally considered effectively weightless because gravity is pulling down but there is no reaction force from the ground. There is quite a fun thought experiment along these lines to do with Einstein's theory of relativity, I think Veritasium made a video about it.
A saw cutting wood might be a better analogy. Same length of stroke, same number of teeth, you'll be expending far more energy cutting at 10Khz vs 60hz. Part of that energy goes towards the friction of the cutting. Part of it counteracting the momentum of the saw itself. The energy lost to counter momentum increases with the mass of the saw. Subwoofers are far larger and heavier than tweeters, ala more energy is lost purely in the motion irrespective of how much actual cutting (sound generated) occurs.
Hands down the best video I've seen describing audio phase relationships both in thoroughness and in critical thinking. I can only hope for more people to enjoy this!
I totally love the detailed technical explanation. Fantastic video. Little brain numb (in a good way) after watching it with high focus. I learned quite a lot from this video. Thanks for the great work! I really didn't expect to get info on both the advanced electronics I expected, but on fluid dynamics as well. I never thought about that before, but it really does make sense having to factor air pressure, displacement etc. And that's all Fluid Dynamics/Mechanics. Totally awesome.
one thing I can add , as I know a sound person or two, is that delay lines are often used with big systems to ensure all drivers are in phase with each other. And if any are at a different distance either ahead or behind the main drivers, then the sound from them is phase corrected with what is comming from ahead or behind them
Delay lines use delays to make the sound sources in "time" with each other, not in phase with each other which would be impossible to achieve being that they are radiating from 2 different locations. Without the delay you would hear a very distinct echo from hearing the delay speaker first and then the main speakers some time after that. It is very disorientating. That echo starts when the two sources are more then about 40 ms apart in time (about 40 feet of distance). If 2 sound sources are less then 40 ms apart then they will sound as one signal. This is called the "Haas Effect". Now what is interesting is that the brain will locate the sound from what sound signal it hears first even if the second signal is somewhat louder then the first. For a delay line, as long as the delay speaker is "slightly" behind in time ~20ms of the main speakers, your brain will still think the source of the sound is from the main speaker (stage) even though the delay speaker is louder and may be off to the side or above you or even slightly behind you.
Came here looking for a comment like this. Maybe you can remove the permanent magnet part of the speaker and drive the air coil in front of the mic to check.
Wish i’d found your channel earlier. I just finished my bachelor degree in music engineering and your videos would’ve made my time much much easier lol. your videos are so detailed and intuitive please keep making these.
Great work! This is the level of detail I would love to see from all audio hardware testing. The only mistake I noticed is that you didn't deinterlace the video around 16:30 which results in horizontal comb artefacts in video. Another example of missing deinterlacing can be seen around 18:22. I'd recommend using ffmpeg for deinterlacing because it has resulted in best quality for me but other options do exist, too.
This must be one of the most beautiful videos I have seen. I watched it 3 times already and will keep doing it. It explains so much and rises so many new questions...
AC coupling is just a lower cutoff frequency low cut filter... oh hey you said it yourself. I better stop commenting. Oh yeah by the way, absolute phase does not matter as much (no human can hear absolute phase differences from a single source). Of course group delay is more important (yes they are related by frequency), but even then up until a certain point mostly unnoticable in a PA kind of situation (room modes and accoustics in general will make sure of that). I still respect this kind of research and compilation of knowledge.
Amazing video, really helps understand all those microphone measurements! Would be interesting to see how a cardiod condenser microphone(capsule with 2 exposed capacitive membranes) would measure, instead of an omnidirectional one, because it would capture a difference in pressure between two sides of the capsule, and not average pressure around it.
Regarding the 180 phase shift between the laser sensor and the microphone at 13.30. The first is a position sensor, the second a pressure sensor. The pressure at the speakers is proportional and in phase with the acceleration of the membrane. There is a mathematical relationship between position, velocity and acceleration of a quantity varying according to the sin law. relationship position - velocity d(t)sin(x)=cos(x) - they are 90 degrees apart. velocity - acceleration d(t)cos(x)= - sin(x) - they are again at 90 degrees. Position-acceleration relationship d(t)d(t)sin(x)= - sin(x) - they are 180 degrees out of phase.
The issue is that most audio engineers are not really engineers / scientists, which is fine. But they cannot interpret measurements correctly. So, people create these very strong believes about how things work based on what they "measured" themselves. Interpreting experimental data and performing proper measurements is difficult, and it is something that people go to school for for many years. So, if you have not, do not expect to be able to properly measure things and interpret the data correctly.
@ 23:20 interesting how due to the phaseshift there is a 2nd order distortion to the laser measured output at the begining cycles. In this tone output it settles down to low distortion in a few cycles I see but with music signal that is not a constant tone it would be more or less all the time distortion and imd. Amplifier damping factor would influence this to an extent I think, high damping = quick recovery but larger amplitude distortion, low damping = slow recovery but less distortion of amplitude. Very interesting topic, thank you so much for making this higly educational video and thank Filipo @ B&C.
This video is brilliant. Taking advantage of the fact that your branch is also music and not just a sound scientist, make experiences with real music (even if you don't have the ability to explain what you see, and therefore it would be much more fragile conjectures). They are all individual or isolated signals, when in reality we work with complex signals. It is true that pink noise is a complex signal, but it still lacks the transient component, which is a key attribute in music.
The current lags behind the voltage in an inductive AC line. Since speaker lines are more or less varied-voltage AC, and you're coiling it around that ferrite ring, you're creating inductance. That inductance will affect the phase and frequency of the output. This is why we use coil inductors to make low-pass circuits. And that's not to mention all the other interference from various amp stages and whatever else. So long as the delay is no more than ~2ms it's fine, humans generally can't hear intervals that small anyway.
I have to say that watching this video in the middle of the night accompanied by a glass of whiskey is a wonderful experience. Some parts you understand, while others seem familiar, just like the alcohol.
Immediately subscribed! The fact YT just recommended your video to me tells me its algorithm isn't as good as it should be. Awesome work! This must have taken a lot of time!
Absolutely INCREDIBLE content and presentation - lifetime producer and audio & physics enthusiast here. You are one of the top UA-camr's I've come across mate
Brilliant stuff. I'd love to see the comparisson between the laser and the mic with a complex wave, to see if the various consituent frequencies all track at 180 degrees to each other, or if there's a compounding effect.
Absolutely fantastic. Add me to the fan base. Could you PLEASE do a video on what’s going on *inside a sealed speaker enclosure with attention to energy that is forced back through the driver diaphragm-especially regarding how this effect is or is not captured but standard measurement specs? I think that would be highly illuminating for many in the loudspeaker design community. Thanks for your work.
Excellent video! Loved the bit about sound wave trough created at speaker movement crest (sound crest 180° out of phase with speaker movement crest). And how for 30 Hz and 200 Hz, speaker movement crest is 90°-180° out of phase with the driving electrical crest (although I'm not sure I grasped why for that part).
Great Video, The laser measures position of the speaker membrane. But position is not what's making sound, that's probably speed or even acceleration. Which is why the Freq. Response of the laser position took a nose dive. Because position is the integral of speed, and speed is the integral of acceleration. And an integration is actually a low pass filter operation. So, you have 1 or 2 low pass filters to compare against sound measurement.
18:50 I think the delay is caused by the inertia of the moving parts of the speaker. And that cannot get worse than 180 degrees because if it were delayed more, it would catch the next incoming electrical wave and that would result in effective speed-up of the movement reducing the delay to less than 180 degrees again. If you have constant latency, it's caused by some kind of processing, not by physical movement of the speaker.
I think it would be nice to watch more of these in action. How would it be form the floor? how woul it be from above? Beautiful, I was thinking abount these set up for a while and here we are! Cheers, Congratulations for your knowledge and imagination and dedication and sharing!!!
While nothing new to the folks designing such hardware, this is an awesome first dive into this topic. Great work! Interesting things I've noticed: - As pointed out in 2:30 1st order HPF will create a 90° phase shift. A 2nd order filter will create a 180° phase shift. - In 17:10 you can see the driver reach 180° phase shift. At 180° the speakers output becomes pretty useless. This seems very similar in behavior to the Gain Bandwidth Limit(GBW) of an OpAmp. (The GBW basically dictates how much of your maximum amplification you can use for a certain frequency). When driving the speaker at a higher volume, I would expect the amplitude to drop even quicker, but the phase behavior to remain the same. - I would love to see a plot of actual time delay instead of phase
Wow this all makes sooo much sense. Wish I found your channel sooner. This has answered alot of questions I've had when figuring out phase relationships between multiple driver setups. I can actually hear phase differences in drivers after some listening. I went and bought rta mics to see if I could see what I was hearing. Turns out I was right. This video puts everything in perspective for me fromy own experience. It all makes perfect sense. Finally a channel that makes my brain tingle! I love it. ❤
I'm only partway through the video, but I gotta say I'm impressed by this video a lot. I'm an electronics engineer and took a course on audio engineering in college (the kind where we talk about speaker low frequency dynamics, Thiel-Small models, psychoacoustics, etc) and seeing the same content from an actual audio engineer's perspective is really fresh and interesting. 13:29 -- There is a phase shift between the position and pressure waveform because of several factors actually; I believe because the acceleration of the movement of the speaker cone and therefore the air molecules it is pushing against is the 2nd derivative of the position, and since one derivative imparts 90 degree of phase advance (think about the derivative of sin(x) being cos(x), which is 90 degrees up), you will immediately see a 180 degree phase advance of the acceleration. Roughly speaking force = acceleration, and force is pressure * area, so pressure (that's SPL) and acceleration is in-phase. In the Thiel-Small model it talks about volume velocity; the relationship between volume velocity and particle velocity is analogous to pressure and force.
I all my life play Viola in symphony orchestras, really good ones. I sit where Bach, Mozart and Beethoven sat in the middle voices and directly in front of the conductor. I sit where the listener sits and I worked in public radio as a classical music host and announcer. I could hear what a lot pretended to hear.
great video. Let me make a wee observation. ALL microphones have HPF. They need to equalise the pressure or they would break in a plane or with changes of atmospheric pressure. Typically this is tuned around 1-3 Hz in lab mics
Hi, first of all, amazing video, thank you! question: why are you using interlaced video in 2024? I can see the interlaced artifact in several parts of the video.
@@devinlsheets_alphasound It might seem like a meaningless detail but you effectively converted pristine 60 fps footage to 30 fps which additionally now contains a bunch of combing artifacts. It distracts the audience from the content of your video and is completely unnecessary
Amazingly well presented explanation of the very interesting behaviour! 5/5 will view again 'cuz I'm not entierly sure I got it nailed down on first attempt.
Phase shift does not alter how we hear a single frequency per se... It alters the relationship between two or more frequencies. Thus, while it alters the timing of peaks at one frequency, this is of little consequence. What is of significant consequence is that phase shifting alters the timing of two or more frequencies differently. It mis-aligns their timing in relation to each other. This alters how you hear the combination. Because remember that our ear only hears one change in air pressure at any given moment. That one change in air pressure is the combination of all the frequencies it can hear. We can't hear 4 separate keys on a piano; we hear the combination of those 4 vibrating strings. All the rest is an illusion of pure brain processing. The more steep the phase shift, the more this relationship is altered, and the more you can hear it compared to the reference sound. And like the microphone, we cannot hear the absolute pressure of one singular moment. We can only hear the change in air pressure between two or more moments. I think I've belabored the point too much already. Phase shift is inevitable. It's nearly impossible to eliminate without major compromises elsewhere. But we can try to make it as gradual as we can. The point is to maintain the relationship between frequencies across the entire audio spectrum. The human ear hears just like the microphone. It's a change in air pressure that creates an electrical signal that goes to the brain; not a measurement of absolute pressure. So the microphone is actually a good representation of the human ear. Eardrum and diaphragm alike, must be in motion to produce a signal. Perhaps someday, we will figure out how to bypass the eardrum entirely and induce electrical signals directly to the brain - producing the absolute perfect recreation. No more speaker design compromises, no more air pressure delays or external noise, no more hearing deficiencies.... just pure sound exactly as the artist intended.
The effects of phase shift are not widely agreed upon in sound reproduction. I can tell from my personal experiments, building a high end electrostatic speakers, that minimum phase does not sound very much different to variable phase. Only a microphone will pick it up... Phase really starts to matter mostly in bass frequencies where cancellations become very audible.
@@Munakas-wq3gp It matters a lot in the upper mid range frequencies where our brains use time delays and amplitude differences for location information. Worst case scenario, all location information is lost or misrepresented.
This is normal since the light is faster than sound ;) Assuming the electronics measuring this do not add any/same phase shift or time delay. Well done!
Thank you very much! You put so much passion and work in your investigations and the video. Awesome setup, respect! All of your argumentations sound logic to me. The only thing I need to think about again in detail is what happens here with the acoustic nearfield/farfield with the longitudinal waves we produce. With the mic distance we should be in nearfield with 90 degrees phaseshift between pressure and velocity but with 200Hz not. I wonder if it matters, because what sensor is capturing what? Mic: captures pressure Laser: captures excursion and translate it to voltage, what represents pressure at the source without nearfield/farfield acoustic effects…good reference btw! …so it is possible, that we also see some of these acoustic effects, what you mentioned and explained with the mic excursion.
You should do a laplace transform of the pulse response of the speaker input to laser output. This will give you the transfer function, and you have all the ansswers you need.
Interesting insight, this raises many questions in what is the objective of sound reproduction, and what matters and what does not. The basis of sound reproduction is for our ears to hear the same sound i.e. original sound versus sampling the original sound and attempting to reproduce the same experience of the listener at a later time. Do the very low frequencies matter and if so is phase significant or how much is the question.
Really love your technical videos, especially with practical examples such as this one and the line array effect! You've confirmed that the pressure/location is allways 180 degrees phase shifted, regardless of frequency. But it might be interesting to mention that, above FS, the input voltage and location are also 180 degrees phase shifted, which results that the microphone phase matches the amplifer voltage in phase! In a 3way speaker, the midrange and tweeter are (/should be) used above their FS which means these frequencies should be in phase on the microphone compared with the amplifier voltage. After delay compensation of course... Also, crossover may (will) mess up stuf depending on implementation but its a fun fact that after all the phase stuff it ends up back were it started for a large part of the frequency range. Sadly it is hard to show this in measurements, because any starting/stopping a sine has additional frequency contents that may (will) fal outside of this frequency range where above remarks are valid...
Another thing you should take into account is that phase delays and out of phase speaker combinations may have very desirable audio effects as well, things related to reverberations, harmonic and spatial effects that people want. Technically correct performance in terms of engineering often don't sound that good, but the "bad" design ends being what sounds the best, not always, but quite often.
Makes sense since the transfer function of a speaker can be modelled as resistor, inductor and capacitor. To counter this, you'd need use an amplifier with the inverse transfer function like that used in Yamaha's servo technology. Cheers
This is something that's well known by anybody with a basic electronics education. If fact, most of this stuff has been known in the early 1900s already.
@@Max24871 Yes. The internet has opposite effect on general public level of knowledge. Negative effect. They know less about the world. They need a UA-cam video on BASIC subjects, phenomena. Like there was no books (you know, real paper ones) etc. Pathetic...
Anyone else find interlaced video to be distracting for some reason? This is super interesting.. I wonder how hard it would be to create a frequency based delay filter that would compensate for the delay, and how that would sound Thanks for making this! I'm gunna have to look at your other videos, Subscribed!
Why would you assume the physical position of the dust cap or any particular spot on the diaphragm matches the transverse waveform? It's not necessarily a phase shift because that implies a comparative source. You are seeing a delay. If you now consider 2 transducers and the amount of delay is the same, they are considered in phase at the chosen frequency.
Is the « slowness » of the system at 200hz due to the resistance of the surround and the weight of the moving parts or does back emf go up with frequency. I’m not sure but faradays law states that back emf is given by the rate of change of the magnetic flux (Wikipedia). At a higher frequency I would say that the rate of change of magnetic flux is higher since the coil is moving in and out of the magnet faster to try and match the input signal. I’m sure the answer is somewhere in the middle, the mechanical resistance and inertia of the moving parts will impead the movement of the cone at higher frequency’s, but I would also guess emf has something to do with it. Lovely video though, it’s so nice to have people to break down the fundamental phenomena at play in speakers, thank you and well done 😊
FACT: - its not the signal (voltage) that moves the speaker. - The speaker moved by magnetic flux, a result of electric current flowing in the coil. - Higher frequency's result in to more current delay inside the coil. (this is exactly the part missing from this video) There for the speaker output is a combination of mechanical and Emf delay. agree: Lovely video though, it’s so nice to have people to break down the fundamental phenomena at play in speakers, thank you and well done 😎
I can't help but think of the AFBS (Acoustic Feedback System) on my 80s Aiwa hifi that uses microphones directly in front of the woofers to ensure what is coming out of them is the same as what is going in (for increased bass from small speakers). It might be a simpler process than I thought if the comparison signal is already more or less inverted, although I am sure there is a lot more thought and design that has gone into it. Japan in the 80s!
My god, this is the most physically correct description I have ever seen of the microphone - speaker interaction. Do you plan on solving the differential equations for the speaker-air-microphone-system? This should give a nice analytical solution, where the theoretical phase should be easy to calculate. (Because if you ain't going to do so, I feel like I want to do it ^^)
@19:00 Speakers are an inductive load. At Dc they are shorts, at higher and higher frequencies they "open up". This is due to the physical sizing of the inductor and how the core saturates. Inductors also introduce a phase lag in the current signal which is why at higher frequencies you are seeing 1 pi radians of shift.
What a great video! I feel like sticking to the driver movement when thinking of the sound is misleading. When making sounds what is characteristic is the change in pressure or the frequency of it. Now I am wondering about the electric energy used for acceleration and deceleration of the driver and of the air which by the way is a proportional and not a derivative of some degree. Maybe there is some hidden magic to get more from a speaker in its confined movement range.
The transfer from speaker input voltage to cone excursion acts as a 2nd order high pass filter, hence gets 180 degr phase shift above its corner freq. (which equals the resonance freq. of the cone).
Amazing video. The mic is measuring the acceleration of the mic’s diaphragm not the acceleration of the speaker driver right? Still a second derivative but for different reasons (and subject to different resonant characteristics)
Awesome video! For those like me questioning now the whole sense of measurements - I guess still the "truth" is what the microphone reads, as this is what our ears hear, isn't it? I mean - we don't hear the exact movement of the cone, or even the movement of the air (velocity). What we do hear is the pressure difference = 2nd derivative of the cone motion, am I right?
I suggest you dive into the concept of loudspeaker drivers behaving as minimum phase systems over most of their frequency range before breakup. Loudspeakers do store energy based on their transfer function just like capacitors and inductors do. The loudspeaker phase response can be decomposed into three components: A frequency independent time delay corresponding to the distance between the baffle and the microphone, the frequency dependent phase related to the magnitude response and frequency independent excessive delay. This excessive delay corresponds to the delay between the virtual source being behind the baffle. In other words, there is an additional frequency independent part on top of what you measure with a ruler from e.g. the dustcap or the baffle to the microphone. Now this frequency independent excessive delay must be compensated for in the microphone signal and it must be so much that the measured phase in the pass band is basically identical to the calculated minimum phase from the amplitude response. (Hilbert Transformation). Higher frequencies do always experience some phase shift due to the inductive behavior of the voice coil. It is an inductor after all. All of this snaps into place once the minimum phase concept is clear.
Please, please do a phase analysis of audiophile crossover components versus cheap crossover caps, resistors, and inductors. Also compare to active crossovers. Want to see two sine sweeps simultaneously 1, 2 or 3 octaves apart and see how far the phase shift is
wow! Congratulations to this absolutely amazing Video. This is peak educational content. I would intuitively think that it would make a difference in phase shift if you'd used a mic with a pressure gradient capsule instead of this measurement mic, wouldn't it? Maybe this would be an interesting follow up Video.
Dang, very interesting to see. I've spent years with these as mental exercises. I came from the side of why do certain very simple tube amps have so much more detail in the mid ranges. Most simply say it's "just tube amps add distortion" but often not in the levels detectable by the human ear. I know that muddy bass they can produce, some might like it, I don't, and that's not what many people are talking about. I have future plans ( I have a workshop clean full of projects, and I"m prone to severe sidetracking) to use a laser or capacitive sensor to resolve speaker position. Nothing was off the shelf, or within budget, last time I looked. So a DIY project planned. But needs to get into the mid ranges as well. But my interest is that single ended tube amps have great highs and mids that I love, but bass is muddy. I believe that is due to the fact that they are more like current sources than voltage sources. I got a book on current source driving loudspeakers, they math is above my skill level, but comes to the conclusion that trying to control a large bass speaker with current control is nearly a lost cause, but very effective at frequencies above that. I believe the author to be correct in his conclusions. One issue is that testing with sine waves only doesn't catch the complex interactions of actual music. Between the forces of the surround and the velocity of the cone what happens when you try to impose a second signal. Instruments(snares come to mind) that have frequencies across the spectrum also seem to sound very good on low or no feedback tube amps. I once heard a wise old HiFi guy say if you see the speaker move it's no longer HiFi, I laughed, then slowly I've started to agree. Started to think that's why I like ESL panels so much. But being unwealthy I have some ugly home built speakers with a couple 6" woofers and AMT tweeters on a simple crossover to avoid phasing issues with a equally ugly hacked together amplifier. People say it sounds great, or I have nice friends, IDK. Sometime soon I'm building another amp to biamp the speakers with a digital crossover. Anyway, great video and very educational, I have nothing to disagree with.
One small comment around 11:20 you show a current plot referenced to the voltage with phase. If you would look at the impedance, that is of course V/I (instead of I/V what is shown here). Because of division of complex numbers give subtraction of phase, the phase of the impedance is exactly flipped (so as to be +90 deg in the ideal coil region). Phase is always a struggle. Reference microphones from B&K/G.R.A.S. are most of the time inverted phase, because the preamp has an inverted stage there but it is not really advertised. Same way when you use accelerometers or vibrometers which measure velocity that you have to add +90 deg every time you make a derivative of the displacement.
You've gotta be one of the most passionate, most interested, most curious, and smartest people in audio on the Internet right now.
If you have a solid understanding of kinematics and fourier transforms it makes this stuff easier to grasp, but that is honestly more esoteric than what most people need to actually apply this in practice. This is BY FAR one of the best intuitive explanations of how phase works with sound reinforcement that I have ever come across. It's something I've understood but always had a hard time explaining in detail because it's such a math heavy topic to really get in to. Great job bridging that gap! Would love to see more.
Inexpensive marginal test gear has released a tidal wave of inaccurate, maybe, and outright made-up-shit from folks focused more on ad revenue than knowing what the phuk they are spewing.
You are the only one in the web teaching actual science with audio engineering. Things most audio “engineers” have no clue about. Thank you
He really isnt. You just need to broaden your horizons a bit. the videos by xiph are superb. there are many tacking it at the level of PCB design for audio as well.
I'm not saying this isnt a good video, because it is, but it isnt uniquely good by any means.
@@jeremyglover5541 Are you talking about the Audio University channel, when you say "xiph"?
www.youtube.com/@AudioUniversity/videos
@@jeremyglover5541you're not wrong, but it's also kind of telling that your example is xiph, because Monty's videos are uniquely the best resource on their topics despite being around 15 years old now.
@@jeremyglover5541do you have any recommendations for people doing audio PCB design?
For the first time ever, I finally understand the affects that phase response has on a system because of this demo. Thank you!
sorry to be that guy, but effects*
I believe there's a term called Jerk when decribes the rate of cahnge of acceleration. Also wonder how damping factor and power bandwidth comes into play. Awesome video, thanks!
Yes, physics behaves just like jerks! 😂👍 Grinds the hecks outta ya, for shiz'n'giggles 😃👌
Some other terms I may have made up - bounce (sharktooth), ramp (sawtooth), step (square wave), impulse, doublet (N-wave), jerk, etc. Not many are taught the elementary waveforms.
We may be jerks, but we also have a sense of humor, the next three derivatives are snap, crackle, and pop. Although, back when I was playing with this stuff, "jounce" was the derivative of jerk.
The derivatives go: position, velocity, acceleration, jerk, snap crackle, pop
True. And snap is the ROC of jerk, crackle the ROC of snap and pop the ROC of crackle.
This is now one of my favourite videos on tech. It's just... the applied sound engineering and exploration, I had to rewatch it to gain insight
That was one of the best descriptions I've ever seen for the behaviors of the "signal, driver, soundwave, mic" relationship. It's a very complicated process, and this is a great way to break it down and understand each component. For your next challenge, you should try to do the same type of breakdown for a bass-reflex cabinet. That is one of the most important and least understood elements of speaker/cabinet design. I've been struggling to wrap my head around it for ages.
An insane amount of useful information and I’m all here for it! The visualizations are super helpful too, thank you so much for the effort!
Most educational half hour I've spent in a long time. Awesome. Thanks!
Wait, that was 30 minutes? I just restarted the video to read the comments. I was too busy soaking in the info to read comments or notice half an hour passed on my first time through
Love it, this reminds me greatly of my electrical engineering classes at uni. You're like connecting all the dots when it comes to audio applications
Amazing! Amazing amazing. This is the kind of content I've always wished existed. Now we just need a pipeline for absolute beginners to get to this point
nothing is in phase my friend
Except the grid
Actually....lasèrs...they got their shit together by definition.
😆
😂
reminds me of Rick & Morty with "true level" experience
Great video. I really like how you've taken something that can very dry, mathematical and abstract and make it understandable in an intuitive way
The phase relationship between the input signal and the output from the speakers is actually much more complicated in practice. The reason is that the equation m.a+c.v+k.x=F(x) applies to established sinusoidal stimuli. In practice, the musical signal is not like that. It is rather a transient process in which the behavior of the speaker is determined by an inhomogeneous system of differential equations, which makes the issue of instantaneous phase shift quite complicated. Thank a lot for video.
There's a small mistake on the "AC coupled" waveform drawing (graph on the right) at 1:55 , after the "DC" portion of the signal, the AC coupled signal should not have any positive slope since there is no positive slope in the original signal. (like when the DC battery is released at 2:12)
Thanks for the nice thorough video again :)
TRUE lol that’s what happens when I make quick and dirty graphics with PowerPoint on long airplane rides.
And btw DC is blocked because running 2 kilowatts of DC to a speaker will ruin your day very fast and you won't hear a thing.
As the first waveform is passed intact, this is not a large enough time snippet to say if the flat section is even DC or just a bit of lower frequency out of context. Most likely the flat "DC" bit would come through DC blocking with a little slope and a slow rebound if the pulse is not followed shortly by a matching negative excursion.
In the near-field of a dipole, which is where you're measuring with your microphone, the pressure and the velocity are 180 degrees out of phase. The driver cone determines the air velocity and the microphone picks up the pressure field. That's where your phase shift at 14:00 comes from.
Could you, please, decipher your thought for us? The wave of 20Hz is about 55 feet long.How will a microphone position offset by a feet or two affect the result?
I don't know much about audio engineering, but I believe he's saying the driver creates an inverse pressure wave with an in phase velocity. Imagine that the biggest change in pressure doesn't happen when the speaker goes from rest/neutral to outermost, but actually when it goes from outermost to innermost positions. The hz doesn't so much matter as the speed of sound itself, too. A very low frequency wouldn't be captured on a mic further away, unless the velocity was astronomical.
Edit: I realized right after posting that my explanation would result in 90 degrees in shift, and also I had paused right before he started explaining velocity being pressure's derivative, which isn't what this poster was talking about, I don't think. The video's explanation at around 24:00 sums it up, no?
@@LuxLucidOfficial Yes, the explanation at 23-25m is a good summary how the air pressure goes in the near field. My question was about "The driver determines cone velocity and the mic picks up the pressure.." To me it sounded like the driver throws apples and the mic picks up oranges, so what's the rationale here?
It is not a dipole. It is a baffled driver. Pressure and velocity are not generally 180 degrees out of phase, since that will depend on the acoustic environment.
@@Rene_Christensen In the very near-field it doesn't matter if it is baffled or not. The moving membrane displaces the air in a way that requires a dominant dipole moment of the multipole series. And the near-field does not know anything about the acoustic environment, by definition. Check out the derivation of the velocity potential of an oscillating solid sphere, in which case you get a pure dipole. The pressure is the negative time derivative of the velocity potential and the velocity of the object coincides with the field velocity at the surface, making both the spatial derivative of the velocity potential. You can then determine yourself that they are exactly pi out of phase near the membrane. Only at a distance that is significantly greater than the membrane radius you will start to see a significant deviation from this relationship.
You can find the relevant derivations here for example: www.math.fsu.edu/~hju/cht12.htm
Love it when i can feel like i understand a complex subject without actually understanding the any of it 😅. But in all honesty, you did an excellent job of explaining things i know little about without losing me and the length of the video perfectly pushed my brain all the way out to its limits!
When i was at university we learned that particle velocity and sound pressure are out of phase in the nearfield and in phase in the far field. And considering that the Driver stimulate the particles directly, resulting in particle velocity and the microphone converting sound pressure to electrical current, your results would make sense (basically also what you explained at 23:55). It would be interesting to see the same measurements in the far field.
I can already tell you what happens in the Farfield because we did do those measurements, they just didn’t make it in this video. What happens is that the mic measurement remains +180 degrees out of polarity with the laser but the mic actually gets slightly ahead of the laser by about 1’ or so. This shift happens very quickly, within a few feet of the driver, and remains consistent all the way out to 200’ and applies at all frequencies. We realized this has to do with the fact that the acoustic center of the driver is actually about 1 foot in front of it. Might make a whole video on this topic.
@@devinlsheets_alphasound That would be great! Because heard and read alot about phase effects of drivers and sound but never really understood what it all means physically. I love how your experiment does exactly that: what is moving and how. Maybe you could even try to visualize how the air moves with dry ice or small plastic balls on strings? just a thought
@@devinlsheets_alphasound "We realized this has to do with the fact that the acoustic center of the driver is actually about 1 foot in front of it"
Yes, it varies for different cabinets.
Would be nice to do the measurements to verify (by multiple independent entities!!!) the model(s)/math.
@@devinlsheets_alphasound That is as it should be. For somewhat free-field conditions (I realize you are in a room, but still), the pressure is in-phase with acceleration of the piston (not the acceleration where the pressure is measured!, for which in the far-field the velocity indeed is in-phase with pressure for pure propagation, as mentioned in some posts here), because the acoustic environment seen from the piston is mass-like at lower frequencies for a typical cone size. It is not quite a perfect match between piston acceleration and pressure, but for example for a flat 18 inch piston in a baffle at 1 m distance, the two are only 5 degrees out of phase at 300 Hz and even closer at lower frequencies. With acceleration being 180 degrees out of phase with displacement, what you see in your measurements is what you should see. I should say that the phase coming the distance travelled from the piston to the microphone (linear phase) needs to be removed in these calculations, so in your measurement setup, you need to either input this distance in some setting, or manually get rid of it.
In a car, what pushes you (pressure) back in the seat? Acceleration. Not speed and not position.
Similarly the air in front of the speaker cone becomes pressurised by the acceleration of the cone, not the speed or the position.
This should also explain why high frequencies have more acoustic power thab lows at the same driver displacment - acceleration is higher.
EDIT: In hindsight I misspoke. Pressure is not created by acceleration, but sound (modulation of pressure) is. Thus g-force felt in a car is analogous to sound energy.
No, pressure is caused by the motion of the cone.
By your logic, a skydiver would feel no air pressure once he reached terminal velocity.
you are confusing the potential energy of sitting in a car seat at constant velocity, with the kinetic energy of a column of air being moved by a cone.
@joshua43214 You make a great point! I don't think it's a perfect analogy. However I would note that sound is not a constant pressure (if the cone were moving at a constant velocity) because that is DC. Sound is changes in pressure (cone changing velocity or accel / decel) and in truth this is a continuous cycle of conversion between potential energy (compressed / rarefied medium) and kinetic energy (medium in motion expanding / contracting).
In the case of a sky diver the medium is in motion past the body (wind) but sound is more like the force felt by the sky diver who will indeed feel weightless at terminal velocity.
Your comment made me think hard, thanks!
@@Audio_Simon I think you have mistaken the skydiver example, the diver doesn't feel weightlessness at terminal velocity. It's only just at the moment of jumping. At terminal velocity, you feel your weight supported by the air.
@tusharjamwal You are quite right that air will resist the motion of the falling body in the form of drag. That said, from an object in free-fall is generally considered effectively weightless because gravity is pulling down but there is no reaction force from the ground. There is quite a fun thought experiment along these lines to do with Einstein's theory of relativity, I think Veritasium made a video about it.
A saw cutting wood might be a better analogy.
Same length of stroke, same number of teeth, you'll be expending far more energy cutting at 10Khz vs 60hz.
Part of that energy goes towards the friction of the cutting. Part of it counteracting the momentum of the saw itself.
The energy lost to counter momentum increases with the mass of the saw.
Subwoofers are far larger and heavier than tweeters, ala more energy is lost purely in the motion irrespective of how much actual cutting (sound generated) occurs.
Hands down the best video I've seen describing audio phase relationships both in thoroughness and in critical thinking. I can only hope for more people to enjoy this!
I totally love the detailed technical explanation. Fantastic video.
Little brain numb (in a good way) after watching it with high focus.
I learned quite a lot from this video. Thanks for the great work!
I really didn't expect to get info on both the advanced electronics I expected, but on fluid dynamics as well. I never thought about that before, but it really does make sense having to factor air pressure, displacement etc. And that's all Fluid Dynamics/Mechanics. Totally awesome.
one thing I can add , as I know a sound person or two, is that delay lines are often used with big systems to ensure all drivers are in phase with each other. And if any are at a different distance either ahead or behind the main drivers, then the sound from them is phase corrected with what is comming from ahead or behind them
Delay lines use delays to make the sound sources in "time" with each other, not in phase with each other which would be impossible to achieve being that they are radiating from 2 different locations. Without the delay you would hear a very distinct echo from hearing the delay speaker first and then the main speakers some time after that. It is very disorientating. That echo starts when the two sources are more then about 40 ms apart in time (about 40 feet of distance). If 2 sound sources are less then 40 ms apart then they will sound as one signal. This is called the "Haas Effect". Now what is interesting is that the brain will locate the sound from what sound signal it hears first even if the second signal is somewhat louder then the first. For a delay line, as long as the delay speaker is "slightly" behind in time ~20ms of the main speakers, your brain will still think the source of the sound is from the main speaker (stage) even though the delay speaker is louder and may be off to the side or above you or even slightly behind you.
Have you thought about the mic being too close to the coil and reading magnetic force instead of air pressure?
Good thought! I wonder flux extends far enough though but still, interesting thought!
Came here looking for a comment like this. Maybe you can remove the permanent magnet part of the speaker and drive the air coil in front of the mic to check.
I changed my mind after watching the complete video, it all makes sense now. Microphone is reporting acceleration, not position.
Wish i’d found your channel earlier. I just finished my bachelor degree in music engineering and your videos would’ve made my time much much easier lol. your videos are so detailed and intuitive please keep making these.
Great work! This is the level of detail I would love to see from all audio hardware testing. The only mistake I noticed is that you didn't deinterlace the video around 16:30 which results in horizontal comb artefacts in video. Another example of missing deinterlacing can be seen around 18:22. I'd recommend using ffmpeg for deinterlacing because it has resulted in best quality for me but other options do exist, too.
This must be one of the most beautiful videos I have seen. I watched it 3 times already and will keep doing it. It explains so much and rises so many new questions...
I look at those Yamaha Dante interfaces everyday,This a great video. Great channel
That's an incredibly inciteful tutorial. I had no idea how much went in in the journey of music through my sound system! Terrific demonstration!
I'm not sure I'll ever need any of this knowledge, but it was super interesting to watch.
Your explanations are very detailed, yet easy to grasp.
AC coupling is just a lower cutoff frequency low cut filter... oh hey you said it yourself. I better stop commenting. Oh yeah by the way, absolute phase does not matter as much (no human can hear absolute phase differences from a single source). Of course group delay is more important (yes they are related by frequency), but even then up until a certain point mostly unnoticable in a PA kind of situation (room modes and accoustics in general will make sure of that). I still respect this kind of research and compilation of knowledge.
The animation 24 minutes in was such an awesome lightbulb moment, thank you!! Epically useful video!
Amazing video, really helps understand all those microphone measurements!
Would be interesting to see how a cardiod condenser microphone(capsule with 2 exposed capacitive membranes) would measure, instead of an omnidirectional one, because it would capture a difference in pressure between two sides of the capsule, and not average pressure around it.
Regarding the 180 phase shift between the laser sensor and the microphone at 13.30. The first is a position sensor, the second a pressure sensor. The pressure at the speakers is proportional and in phase with the acceleration of the membrane. There is a mathematical relationship between position, velocity and acceleration of a quantity varying according to the sin law. relationship position - velocity d(t)sin(x)=cos(x) - they are 90 degrees apart. velocity - acceleration d(t)cos(x)= - sin(x) - they are again at 90 degrees. Position-acceleration relationship d(t)d(t)sin(x)= - sin(x) - they are 180 degrees out of phase.
The issue is that most audio engineers are not really engineers / scientists, which is fine. But they cannot interpret measurements correctly. So, people create these very strong believes about how things work based on what they "measured" themselves. Interpreting experimental data and performing proper measurements is difficult, and it is something that people go to school for for many years. So, if you have not, do not expect to be able to properly measure things and interpret the data correctly.
@ 23:20 interesting how due to the phaseshift there is a 2nd order distortion to the laser measured output at the begining cycles. In this tone output it settles down to low distortion in a few cycles I see but with music signal that is not a constant tone it would be more or less all the time distortion and imd. Amplifier damping factor would influence this to an extent I think, high damping = quick recovery but larger amplitude distortion, low damping = slow recovery but less distortion of amplitude. Very interesting topic, thank you so much for making this higly educational video and thank Filipo @ B&C.
This video is brilliant. Taking advantage of the fact that your branch is also music and not just a sound scientist, make experiences with real music (even if you don't have the ability to explain what you see, and therefore it would be much more fragile conjectures). They are all individual or isolated signals, when in reality we work with complex signals. It is true that pink noise is a complex signal, but it still lacks the transient component, which is a key attribute in music.
The current lags behind the voltage in an inductive AC line. Since speaker lines are more or less varied-voltage AC, and you're coiling it around that ferrite ring, you're creating inductance. That inductance will affect the phase and frequency of the output. This is why we use coil inductors to make low-pass circuits.
And that's not to mention all the other interference from various amp stages and whatever else.
So long as the delay is no more than ~2ms it's fine, humans generally can't hear intervals that small anyway.
I have to say that watching this video in the middle of the night accompanied by a glass of whiskey is a wonderful experience. Some parts you understand, while others seem familiar, just like the alcohol.
Immediately subscribed! The fact YT just recommended your video to me tells me its algorithm isn't as good as it should be. Awesome work! This must have taken a lot of time!
Absolutely INCREDIBLE content and presentation - lifetime producer and audio & physics enthusiast here. You are one of the top UA-camr's I've come across mate
Brilliant stuff. I'd love to see the comparisson between the laser and the mic with a complex wave, to see if the various consituent frequencies all track at 180 degrees to each other, or if there's a compounding effect.
Absolutely fantastic. Add me to the fan base. Could you PLEASE do a video on what’s going on *inside a sealed speaker enclosure with attention to energy that is forced back through the driver diaphragm-especially regarding how this effect is or is not captured but standard measurement specs? I think that would be highly illuminating for many in the loudspeaker design community. Thanks for your work.
Excellent video! Loved the bit about sound wave trough created at speaker movement crest (sound crest 180° out of phase with speaker movement crest). And how for 30 Hz and 200 Hz, speaker movement crest is 90°-180° out of phase with the driving electrical crest (although I'm not sure I grasped why for that part).
Love it, appreciate your curiousity and determination to understand things. Some nuggets of gold in here for those into audio system measurements
Great Video,
The laser measures position of the speaker membrane.
But position is not what's making sound, that's probably speed or even acceleration.
Which is why the Freq. Response of the laser position took a nose dive.
Because position is the integral of speed, and speed is the integral of acceleration.
And an integration is actually a low pass filter operation.
So, you have 1 or 2 low pass filters to compare against sound measurement.
18:50 I think the delay is caused by the inertia of the moving parts of the speaker. And that cannot get worse than 180 degrees because if it were delayed more, it would catch the next incoming electrical wave and that would result in effective speed-up of the movement reducing the delay to less than 180 degrees again. If you have constant latency, it's caused by some kind of processing, not by physical movement of the speaker.
I think it would be nice to watch more of these in action. How would it be form the floor? how woul it be from above? Beautiful, I was thinking abount these set up for a while and here we are! Cheers, Congratulations for your knowledge and imagination and dedication and sharing!!!
While nothing new to the folks designing such hardware, this is an awesome first dive into this topic. Great work!
Interesting things I've noticed:
- As pointed out in 2:30 1st order HPF will create a 90° phase shift. A 2nd order filter will create a 180° phase shift.
- In 17:10 you can see the driver reach 180° phase shift. At 180° the speakers output becomes pretty useless. This seems very similar in behavior to the Gain Bandwidth Limit(GBW) of an OpAmp. (The GBW basically dictates how much of your maximum amplification you can use for a certain frequency). When driving the speaker at a higher volume, I would expect the amplitude to drop even quicker, but the phase behavior to remain the same.
- I would love to see a plot of actual time delay instead of phase
Wow this all makes sooo much sense. Wish I found your channel sooner. This has answered alot of questions I've had when figuring out phase relationships between multiple driver setups. I can actually hear phase differences in drivers after some listening. I went and bought rta mics to see if I could see what I was hearing. Turns out I was right. This video puts everything in perspective for me fromy own experience. It all makes perfect sense. Finally a channel that makes my brain tingle! I love it. ❤
Fascinating.
You were able to come up with great ways to show what I have suspected for years.
god damn, it took me years to understand but this video helped ma A LOT!
I love audio but this is totally over my head.
It’s awesome! Thank you for your videos.
21:00 This was the most interesting part for me! Great work explaining the microphone behavior.
I'm only partway through the video, but I gotta say I'm impressed by this video a lot. I'm an electronics engineer and took a course on audio engineering in college (the kind where we talk about speaker low frequency dynamics, Thiel-Small models, psychoacoustics, etc) and seeing the same content from an actual audio engineer's perspective is really fresh and interesting.
13:29 -- There is a phase shift between the position and pressure waveform because of several factors actually; I believe because the acceleration of the movement of the speaker cone and therefore the air molecules it is pushing against is the 2nd derivative of the position, and since one derivative imparts 90 degree of phase advance (think about the derivative of sin(x) being cos(x), which is 90 degrees up), you will immediately see a 180 degree phase advance of the acceleration. Roughly speaking force = acceleration, and force is pressure * area, so pressure (that's SPL) and acceleration is in-phase.
In the Thiel-Small model it talks about volume velocity; the relationship between volume velocity and particle velocity is analogous to pressure and force.
I all my life play Viola in symphony orchestras, really good ones. I sit where Bach, Mozart and Beethoven sat in the middle voices and directly in front of the conductor. I sit where the listener sits and I worked in public radio as a classical music host and announcer. I could hear what a lot pretended to hear.
Only five minutes in but this is sick! Love it!
great video. Let me make a wee observation. ALL microphones have HPF. They need to equalise the pressure or they would break in a plane or with changes of atmospheric pressure. Typically this is tuned around 1-3 Hz in lab mics
Thanks for this insight!
Exactly the content for my morning coffee (and then I have to watch it another 5 times to somehow comprehend 😅)
Hi, first of all, amazing video, thank you!
question: why are you using interlaced video in 2024? I can see the interlaced artifact in several parts of the video.
It was set to interlaced somehow when I exported in Davinci and I didn’t catch it because I’m an audio guy and literally pay no attention to video LOL
@@devinlsheets_alphasound It might seem like a meaningless detail but you effectively converted pristine 60 fps footage to 30 fps which additionally now contains a bunch of combing artifacts. It distracts the audience from the content of your video and is completely unnecessary
Amazingly well presented explanation of the very interesting behaviour! 5/5 will view again 'cuz I'm not entierly sure I got it nailed down on first attempt.
Even for a layman like myself this was incredibly insightful!
yup. particle velocities ... pressure ... reactive near-field acoustic energy (air) flowa ... mechanical impedances.
great video.
Phase shift does not alter how we hear a single frequency per se... It alters the relationship between two or more frequencies. Thus, while it alters the timing of peaks at one frequency, this is of little consequence. What is of significant consequence is that phase shifting alters the timing of two or more frequencies differently. It mis-aligns their timing in relation to each other. This alters how you hear the combination. Because remember that our ear only hears one change in air pressure at any given moment. That one change in air pressure is the combination of all the frequencies it can hear. We can't hear 4 separate keys on a piano; we hear the combination of those 4 vibrating strings. All the rest is an illusion of pure brain processing. The more steep the phase shift, the more this relationship is altered, and the more you can hear it compared to the reference sound. And like the microphone, we cannot hear the absolute pressure of one singular moment. We can only hear the change in air pressure between two or more moments. I think I've belabored the point too much already.
Phase shift is inevitable. It's nearly impossible to eliminate without major compromises elsewhere. But we can try to make it as gradual as we can. The point is to maintain the relationship between frequencies across the entire audio spectrum.
The human ear hears just like the microphone. It's a change in air pressure that creates an electrical signal that goes to the brain; not a measurement of absolute pressure. So the microphone is actually a good representation of the human ear. Eardrum and diaphragm alike, must be in motion to produce a signal. Perhaps someday, we will figure out how to bypass the eardrum entirely and induce electrical signals directly to the brain - producing the absolute perfect recreation. No more speaker design compromises, no more air pressure delays or external noise, no more hearing deficiencies.... just pure sound exactly as the artist intended.
The effects of phase shift are not widely agreed upon in sound reproduction. I can tell from my personal experiments, building a high end electrostatic speakers, that minimum phase does not sound very much different to variable phase. Only a microphone will pick it up... Phase really starts to matter mostly in bass frequencies where cancellations become very audible.
@@Munakas-wq3gp It matters a lot in the upper mid range frequencies where our brains use time delays and amplitude differences for location information. Worst case scenario, all location information is lost or misrepresented.
This is normal since the light is faster than sound ;) Assuming the electronics measuring this do not add any/same phase shift or time delay. Well done!
This changed my life
It was all an illusion...
Thank you very much! You put so much passion and work in your investigations and the video. Awesome setup, respect!
All of your argumentations sound logic to me.
The only thing I need to think about again in detail is what happens here with the acoustic nearfield/farfield with the longitudinal waves we produce. With the mic distance we should be in nearfield with 90 degrees phaseshift between pressure and velocity but with 200Hz not. I wonder if it matters, because what sensor is capturing what?
Mic: captures pressure
Laser: captures excursion and translate it to voltage, what represents pressure at the source without nearfield/farfield acoustic effects…good reference btw!
…so it is possible, that we also see some of these acoustic effects, what you mentioned and explained with the mic excursion.
You should do a laplace transform of the pulse response of the speaker input to laser output. This will give you the transfer function, and you have all the ansswers you need.
Thank you for this excellent presentation of complex topic!
Interesting insight, this raises many questions in what is the objective of sound reproduction, and what matters and what does not. The basis of sound reproduction is for our ears to hear the same sound i.e. original sound versus sampling the original sound and attempting to reproduce the same experience of the listener at a later time. Do the very low frequencies matter and if so is phase significant or how much is the question.
This content was very nice and blissfully pleasant to watch! Thanks
Really love your technical videos, especially with practical examples such as this one and the line array effect!
You've confirmed that the pressure/location is allways 180 degrees phase shifted, regardless of frequency. But it might be interesting to mention that, above FS, the input voltage and location are also 180 degrees phase shifted, which results that the microphone phase matches the amplifer voltage in phase! In a 3way speaker, the midrange and tweeter are (/should be) used above their FS which means these frequencies should be in phase on the microphone compared with the amplifier voltage. After delay compensation of course... Also, crossover may (will) mess up stuf depending on implementation but its a fun fact that after all the phase stuff it ends up back were it started for a large part of the frequency range. Sadly it is hard to show this in measurements, because any starting/stopping a sine has additional frequency contents that may (will) fal outside of this frequency range where above remarks are valid...
Another thing you should take into account is that phase delays and out of phase speaker combinations may have very desirable audio effects as well, things related to reverberations, harmonic and spatial effects that people want.
Technically correct performance in terms of engineering often don't sound that good, but the "bad" design ends being what sounds the best, not always, but quite often.
Makes sense since the transfer function of a speaker can be modelled as resistor, inductor and capacitor. To counter this, you'd need use an amplifier with the inverse transfer function like that used in Yamaha's servo technology. Cheers
man your research is underrated, this must be put on papers❤
This is something that's well known by anybody with a basic electronics education. If fact, most of this stuff has been known in the early 1900s already.
@@Max24871 Ah. ty then
@@Max24871 Yes. The internet has opposite effect on general public level of knowledge. Negative effect. They know less about the world. They need a UA-cam video on BASIC subjects, phenomena. Like there was no books (you know, real paper ones) etc. Pathetic...
Anyone else find interlaced video to be distracting for some reason?
This is super interesting.. I wonder how hard it would be to create a frequency based delay filter that would compensate for the delay, and how that would sound
Thanks for making this! I'm gunna have to look at your other videos, Subscribed!
First time viewer. Awesome stuff. Would love to see a part ii of what this means in practice
Why would you assume the physical position of the dust cap or any particular spot on the diaphragm matches the transverse waveform? It's not necessarily a phase shift because that implies a comparative source. You are seeing a delay. If you now consider 2 transducers and the amount of delay is the same, they are considered in phase at the chosen frequency.
Is the « slowness » of the system at 200hz due to the resistance of the surround and the weight of the moving parts or does back emf go up with frequency.
I’m not sure but faradays law states that back emf is given by the rate of change of the magnetic flux (Wikipedia). At a higher frequency I would say that the rate of change of magnetic flux is higher since the coil is moving in and out of the magnet faster to try and match the input signal. I’m sure the answer is somewhere in the middle, the mechanical resistance and inertia of the moving parts will impead the movement of the cone at higher frequency’s, but I would also guess emf has something to do with it.
Lovely video though, it’s so nice to have people to break down the fundamental phenomena at play in speakers, thank you and well done 😊
FACT: - its not the signal (voltage) that moves the speaker.
- The speaker moved by magnetic flux, a result of electric current flowing in the coil.
- Higher frequency's result in to more current delay inside the coil.
(this is exactly the part missing from this video)
There for the speaker output is a combination of mechanical and Emf delay.
agree: Lovely video though, it’s so nice to have people to break down the fundamental phenomena at play in speakers, thank you and well done 😎
Curiosity, realworld orginal tests, weird data that makes it more interesting. very very informative.. thanks man, I was full today too...
I can't help but think of the AFBS (Acoustic Feedback System) on my 80s Aiwa hifi that uses microphones directly in front of the woofers to ensure what is coming out of them is the same as what is going in (for increased bass from small speakers). It might be a simpler process than I thought if the comparison signal is already more or less inverted, although I am sure there is a lot more thought and design that has gone into it. Japan in the 80s!
My god, this is the most physically correct description I have ever seen of the microphone - speaker interaction. Do you plan on solving the differential equations for the speaker-air-microphone-system? This should give a nice analytical solution, where the theoretical phase should be easy to calculate. (Because if you ain't going to do so, I feel like I want to do it ^^)
@19:00 Speakers are an inductive load. At Dc they are shorts, at higher and higher frequencies they "open up". This is due to the physical sizing of the inductor and how the core saturates. Inductors also introduce a phase lag in the current signal which is why at higher frequencies you are seeing 1 pi radians of shift.
Wow that was an excellent presentation! I have one question. How did account for amplifier damping factor influence on the speaker?
Congrats on the bump in views. Great content.
This is like music to my ears after taking Magu's Meyer Fundamentals training :-) ty
What a great video! I feel like sticking to the driver movement when thinking of the sound is misleading. When making sounds what is characteristic is the change in pressure or the frequency of it.
Now I am wondering about the electric energy used for acceleration and deceleration of the driver and of the air which by the way is a proportional and not a derivative of some degree.
Maybe there is some hidden magic to get more from a speaker in its confined movement range.
The transfer from speaker input voltage to cone excursion acts as a 2nd order high pass filter, hence gets 180 degr phase shift above its corner freq. (which equals the resonance freq. of the cone).
Amazing video. The mic is measuring the acceleration of the mic’s diaphragm not the acceleration of the speaker driver right? Still a second derivative but for different reasons (and subject to different resonant characteristics)
Awesome video! For those like me questioning now the whole sense of measurements - I guess still the "truth" is what the microphone reads, as this is what our ears hear, isn't it? I mean - we don't hear the exact movement of the cone, or even the movement of the air (velocity). What we do hear is the pressure difference = 2nd derivative of the cone motion, am I right?
Wow This was so much fun stepping this out the way you have.. Learning tons.. Love it...
This is good content. Thanks for the effort!
I suggest you dive into the concept of loudspeaker drivers behaving as minimum phase systems over most of their frequency range before breakup. Loudspeakers do store energy based on their transfer function just like capacitors and inductors do. The loudspeaker phase response can be decomposed into three components: A frequency independent time delay corresponding to the distance between the baffle and the microphone, the frequency dependent phase related to the magnitude response and frequency independent excessive delay. This excessive delay corresponds to the delay between the virtual source being behind the baffle. In other words, there is an additional frequency independent part on top of what you measure with a ruler from e.g. the dustcap or the baffle to the microphone. Now this frequency independent excessive delay must be compensated for in the microphone signal and it must be so much that the measured phase in the pass band is basically identical to the calculated minimum phase from the amplitude response. (Hilbert Transformation). Higher frequencies do always experience some phase shift due to the inductive behavior of the voice coil. It is an inductor after all. All of this snaps into place once the minimum phase concept is clear.
Crazy setup. Put a lot of effort into this !
Fascinating.. as well as a sound guy and electronics engineer..😊
Please, please do a phase analysis of audiophile crossover components versus cheap crossover caps, resistors, and inductors. Also compare to active crossovers. Want to see two sine sweeps simultaneously 1, 2 or 3 octaves apart and see how far the phase shift is
wow! Congratulations to this absolutely amazing Video. This is peak educational content.
I would intuitively think that it would make a difference in phase shift if you'd used a mic with a pressure gradient capsule instead of this measurement mic, wouldn't it?
Maybe this would be an interesting follow up Video.
Hi what DATS version do you have?, and can you measure a impedance sweep that ranges over 20kHz?
its funny, I didn't realize it was you channel. Your topics are just too interesting for me to pass up.
Dang, very interesting to see. I've spent years with these as mental exercises. I came from the side of why do certain very simple tube amps have so much more detail in the mid ranges. Most simply say it's "just tube amps add distortion" but often not in the levels detectable by the human ear. I know that muddy bass they can produce, some might like it, I don't, and that's not what many people are talking about. I have future plans ( I have a workshop clean full of projects, and I"m prone to severe sidetracking) to use a laser or capacitive sensor to resolve speaker position. Nothing was off the shelf, or within budget, last time I looked. So a DIY project planned. But needs to get into the mid ranges as well. But my interest is that single ended tube amps have great highs and mids that I love, but bass is muddy. I believe that is due to the fact that they are more like current sources than voltage sources. I got a book on current source driving loudspeakers, they math is above my skill level, but comes to the conclusion that trying to control a large bass speaker with current control is nearly a lost cause, but very effective at frequencies above that. I believe the author to be correct in his conclusions. One issue is that testing with sine waves only doesn't catch the complex interactions of actual music. Between the forces of the surround and the velocity of the cone what happens when you try to impose a second signal. Instruments(snares come to mind) that have frequencies across the spectrum also seem to sound very good on low or no feedback tube amps. I once heard a wise old HiFi guy say if you see the speaker move it's no longer HiFi, I laughed, then slowly I've started to agree. Started to think that's why I like ESL panels so much. But being unwealthy I have some ugly home built speakers with a couple 6" woofers and AMT tweeters on a simple crossover to avoid phasing issues with a equally ugly hacked together amplifier. People say it sounds great, or I have nice friends, IDK. Sometime soon I'm building another amp to biamp the speakers with a digital crossover. Anyway, great video and very educational, I have nothing to disagree with.
One small comment around 11:20 you show a current plot referenced to the voltage with phase. If you would look at the impedance, that is of course V/I (instead of I/V what is shown here). Because of division of complex numbers give subtraction of phase, the phase of the impedance is exactly flipped (so as to be +90 deg in the ideal coil region).
Phase is always a struggle. Reference microphones from B&K/G.R.A.S. are most of the time inverted phase, because the preamp has an inverted stage there but it is not really advertised. Same way when you use accelerometers or vibrometers which measure velocity that you have to add +90 deg every time you make a derivative of the displacement.