Hi Paul! I really enjoy listening to and seeing you here in Portugal and I have learned a lot about HiFi from you. If my CEO knew as much about our "products" as Paul knows about his area, it would be a blessing. Stay healthy in the company of family and employees. Thank you very much! Fernando
Absolutely brilliant - I had often wondered how a speaker accomplishes making multiple frequencies simultaneously. Thanks a million for explaining it so well 🙂
The fundamentals of this is that the diaphragm (to the best of its ability) just follows the drive signal (whatever it might be) - show the signal. Thanks, Paul, for another great video.
taking the loudspeaker components , the speaker fixed magnet , the voice coil suspension flexible holder , the voice coil attached to the speaker cone , taking delta T sec . , time constant , the voice coil , will oscillates the summations of all the fundamental frequencies , thus emitting all the fundamental frequencies from speaker cone , the overall effects is the integral summations of the audio complex wave over a unit time .
That would be something for these high-speed camera freaks, to film a loudspeaker with thousands of images every second and to "fire" the loudspeaker first with one sound (one frequency), then two, and so on .. ;)
we can not get enough of this question, i need all the angles to wrap my head around the answer. one way to explain it: imagine the outside of your house, all the sounds all around you, they are so many frequencies there and there are only one eardrum to pick it up. all the sounds are merging together making a complex wave
Inside your ear though, there are tiny hairs of various lengths, which resonate at particular frequencies, and that is actually how we hear sound as a mixture of frequencies. The idea that sound is composed of different frequencies mostly comes from how we perceive it biologically. The mathematical technique of Fourier transform basically does what our ear does--that is, convert a complex wave into its composite frequencies and amplitudes (just like our ear!)
Great question and answer, Paul. I've been listening to reproduced music all my life and even used to build speakers in the '70's and now I know why I never got the sound I was looking for. I didn't completely understand how the component speakers/horns did their work and why crossover design and construction is so critical to how they perform. Thanks!
Intermod is just two steady state sine waves. Again not accurately representing an actual typical sound signal. But way back in the day test and measurement equipment was very limited and adding two signals together was an advancement from using only one, like THD. Later changing the 2nd signal to a squarewave for TIM. Transient Intermodulation Distortion.
Until know I never asked myself that question. However this is very interesting to know. Each time we learn something we become richer persons. Thanks!
My realization of the doppler effect came from being in a 'garage band'. What happened was plugging the mic into a bass guitar amp produced wild as heck modulations to the singers voice because the vocal frequencies were 'riding' on the much slower bass movements of the speaker cone. It was similar to the sound you can make whilst gargling and humming at the same time.
Your eardrums, speakers, cables, signal paths through electronics and microphones all transfer music as a sum of waves represented as air pressure, mechanical motion (e.g. in a speaker cone), magnetic field (e.g. in a voice coil or cartridge), electric field (e.g. in electrostatic speakers) or voltage. The true magic is how our hearing works with the cochlea in our inner ear sort of disassembling the sum of sound waves to individual nerve pulses that our brain can process as those same separate frequencies that went into the music to begin with.
INteresting. SO this then raises the question if your ear sums the signal anyways is having the signal broken up by different drivers really better than a single driver producing an already summed pressure wave ?
@@Canadian_Eh_I Technically a single point sources is always best. But as Paul states, speakers not being point sources have other issues. In order for a speaker to produce lows it has to have some combination of larger size and/or greater excursions. To produce highs it has to be smaller and/or able to move very slightly but accurately without producing that signal in multiple spots at different phases. Thus very different design parameters for each.
@@glenncurry3041 I have a pair of KEF LS60 in one room that I find quite awesome with near single point performance. Right now I'm listening in my home office setup using a pair of KEF LSX II speakers connected via USB to my iMac 27" and a reworked/modified Velodyne servo subwoofer (LSXII includes subwoofer output) giving a very clean sub bass flat down to 20Hz. The Uni-Q drivers in my KEF speakers are 2-way and probably the closest you can get to a single point source. The subwoofer in this room blends in perfectly and people I've demoed this office setup tell me the bass is amazing and when I tell them it's thanks to a servo subwoofer, please find it, they didn't find it. It's a bummer servo subwoofers are so rare.
Additive and subtractive mixing on the recording means the speaker travels the mean of all of the frequencies mixed together at any point in time. Watching music on an oscilloscope will give you a visual representation of what is going on.
That's why we have intermodulation with full range speakers, separating drivers helps to avoid that, but i think there are some speakers manufactured in a way that at a certain point the higher frequencies just oscillates at the center, they use different materials and shapes to provoke a mechanical behavior.
Waveforms can be added or subtracted. When you put two waveforms of different frequencies together, you’ll get their sum/difference as some will cause constructive or destructive interference and your resulting waveform will look like the combination of these two. Same way your hearing works. And if you look at that resulting waveform with the scope, and use Fourier transformations (decomposing the waveform into its components) you’ll get the original two waveforms. The speaker moves the same way as the combination of the two waveforms. Your eardrum will also do the same thing if you listened to two waveforms from two separate speakers. If you shine a laser onto your eardrum to reflect onto a surface, you’ll see the combined waveform of the two individual waveforms coming from two separate speakers being combined by your ear drum. Kind of like colors of light, your eye sees as combined but prism can separate them into its components. Actually eyes see differences between red/green/blue and the differences between the intensities of those three are interpreted by our brain as a gamut of colors.
@@larazss3254 Superposition holds, so all waves are added together, and their sum can be decomposed into a resulting frequency spectrum. This spectrum assumes steady-state conditions, so while the music is limited in time, the components started infinite time ago, so it is abstraction whenever we go to the frequency domain.
@@larazss3254 you get bass frequencies go to subwoofer, mids go to woofer and highs go to tweeter. That is usually done in the cross over. Sometimes the frequencies are separated in the DSP, and go to individual amplifiers to power the different drivers like in a bi-amp application. But still you can have multiple frequencies of mid range or low or highs getting to one mid or low or high range speaker as they may be in the same group as cross over separates by groups which are ranges like bass 20-125Hz, mid 125-3000Hz, high 3000-20,000Hz. So you may end up with two 5000Hz and 6000Hz going to the same speaker which would play the sum of both waves. If you have 30Hz and 6000Hz, these two would go to separate speakers separated by the cross over. However, when they arrive at your ear or microphone, they will sum up and your ear drum or microphone membrane will vibrate summing both of them. It would look like a big wave of 30Hz with small wave on top of the large wave. Kind of like if you got large waves on a lake and when quick wind gust blow over the lake, you get little tiny waves in top of the large waves looking like little shivers. I hope that kind of helps visualizing it.
@@christianratajczak2809 Thank you, that is exactly what Paul explained. But it is still hard to imagine, how analog crossovers can precisely identify and carve out these small waves on top of the large ones...
I think it is easier to say that the speaker responds to a varying voltage that is a composite of all the frequencies in a given signal. A properly operating speaker is merely reproducing the waveform it is presented with.
Thanks for answering my question! I was aware of how the diaphragm moves at multiple frequencies so I was more so asking how that's possible when there is only one magnetic field at a time. Based on what you said, it sounds like the answer is because the magnetic field changes so quickly. Thanks!
The field driving the coil has the overlapped frequencies already composited into the voltage it gets, by the audio circuitry. Whats also interesting is how your ear takes the single overlapping wave and splits it out into frequencies using a biological structure.
@@KK-dv3wh The human brain is just wild when it comes to things like this. 😂 I'm not sure if I should call it an issue or be thankful that the alternative isn't happening, but all the rooms reverb is compiled to one (muddy) signal by the brain as well (instead of multiple clear echos). Room treatment for the win! ✌️😄
Simple answer is we typically do not listen to a single repetitive sine wave nor combinations of a couple. It becomes unlocking your mind from that constraint. Music is constantly changing, non-repetitive complex wave.
Thanks , Paul ! I had always thought of multiple frequencies simultaneously as intermodulation distortion , but Doppler does seem a more accurate descriptor . I imagine a line array with 20 drivers and 19 crossovers would be silly to keep each driver in a "sweet" half octave . That would be a lot of expensive capacitors !
treating it in a mathematical equations , a complex wave is a summations of individual frequencies ,alternatively known as Fourier Series a summations of sine 2*pi*f or cosine 2 *pi* f , where f is the individual frequencies . ( followed physical explanations - later explained ) .
the explanations of the multi frequency audio radiations of a loudspeaker , is best described by , arithmetic summations of the various audio fundamental frequencies , as expressed in the Fourier Series expressions of the Complex Audio Signal with the respect to Time .
I figured it works like runners around a track: runners on the inside of the track will circle faster than the outside runner if they didn’t start staggered. The higher frequencies are coming off the area of the cone nearest to the middle, lower frequencies from the outer closest to the edge.
I JUST spoke about this in my recent videos too... related to complex music.. and quick changes.. transient response.. articulate movement.. damping.. etc etc.. A pretty good explanation. I think this needs a 3d animation to show how these waves are simultaneously present.. Its harder to speak to and easier to show a graphic. Very good and important for people to understand.. say... reasons why we have...2way.. 3way.. 4 way..5 way.. etc... Also... would be nice to see how these frequency and signals are processed in the cross overs.. as in.. how the electrical signals / Frequencies are ' separated ' in the wires / caps / resistors. 🙏🙏🙏🙏🙏🙏 Thanks, Paul. ..Education is our Primary step in any journey. (Also.. Great to meet Chris at Axpona - we spoke briefly about the FR5 - " issue " / demo model performance)
Thanks! Could you please explain further how one can tell the difference in instrument (like an oboe vs. a violin playing the same note) when played by the same driver?
This question is a prime example of "a picture is worth a thousand words. A simple view of the audio signal on an oscilloscope would answer the question. DRIVERS CAN PLAY MULTIPLE FREQUENCIES AT THE SAME TIME.
I like the explanation for the question. Re single drivers and Doppler effect, don't buy it given the small diver excursion and frequencies involved. Show the math, please. Have a single driver speaker myself -- PearlAcoustics Sibelius SG -- and love it. Exquisitely detailed and clean sound with wonderful soundstage. As there is no need for a cross over, and there is a single source point for all sound solves a lot of other problems. KISS principle comes to mind.
Doppler distortion is largely inconsequential at the excursion levels of a well designed full range unit. Research has shown that if kept within 2cm it can be largely disguarded as a problem. If you think about a passing ambulance it has to travel some distance before we hear the drop in frequency caused by the doppler effect. All coil drivers exhibit the same behavior with respect to their bandwidth,where the lower limit is set by either higher mass and/ or compliance from their suspension (box alignment accepted) and the upper frequencies being a result of flexing within the cone structure around it's centre.Even the exotic rigid attempt materials do this and it's thanks to that behaviour that their higher frequency extension is simultaneously attainable.
2cm wavelength is 80Khz. So yes if your single driver design is flat, linear out to 80Khz,.... otherwise operating in non-linear areas causes distortions.
@@glenncurry3041Hi Glen. On the doppler effect topic,the modulation is proportional to the speed of the cone at it's low frequency excursion where the travel is greatest. A 2cm excursion (not wavelength) is quite alot of travel and it will be lower in the majority of units,full range or otherwise at " typical " listening levels,so can be largely discounted. The 80khz frequency mentioned will require extremely low mass,and/or dimensions and excursion, not to mention how far that frequency is beyond audibility, so has no relevance at all with respect to the doppler effect.
A voicecoil can vibrate at multiple frequencies. Old style "full range" speakers use to have small secondary cones glued to the dust cap to better capture and expel high frequencies from the voicecoil.
Very good explanation, thank you. Paul, a question... I know that the human ear is capable of hearing a range of frequencies that goes from 20 Hz to 20 KHz. I also know that these are four speakers (Tweeter, MidRange, MidBass and Subwoofer) that, “together”, are capable of reproducing the entire range of frequencies mentioned. But I ask... Each of the four speakers mentioned reproduces, individually, which range of frequencies? In other words, the Tweeter is responsible for reproducing, individually, which range of frequencies? What about MidRange? What about MidBass? And the subwoofer? I would just like to have a general idea of the frequency range that each speaker works individually. It's a generic question, as I deduce that there is no fixed rule for this. But I would just like to have a rough idea.
I remember back in the late 80’s I think, that Pioneer patented and launched a set of speakers that used some new technology to ensure that a speaker could output different frequencies without distortion. I have searched for years for these speakers or any reviews, but to no avail.
I think woofer beaming should be addressed: above a certain frequency, the sound is coming straight at you, and has poor dispersion. The tweeter still has wide dispersion. I became aware of a new Mission speaker that uses a 2.9 kHz crossover point with an 8 inch woofer, while MoFi tweeters use 1.6 kHz. Wouldn't this create a problem for the listener if they're not seated right in front of the Missions? The part I'm not sure about is, the effect of the driver's dust cap: would that improve the dispersion, making the traditional calculation between the woofer diameter and the wavelength unreliable? Someone in a forum had brought this up recently. I had been using Brent Butterworth's article about what is the ideal speaker driver configuration as a guide, previously.
You forgot to mention that cone design factor and the physical attributes of said cone . Higher frequencies in a driver due to speed are released from center out . Hence putting a whizzer cone at dust cap center . The center moves faster than the outer edge because there is less area . A good /bad cone design is essential to its ability to produce proper frequencies thought a certain spectrum .
Expressing a time signal (that's what's actually reproduced) in terms of a sine of sums is simply a mathematical transformation expressing tne signal in a new basis. That snusoids (in fact, you also need phases) can form a basis (i.e. any sine+base be converted to time and vice-versa) is the actual physical property. We use sines bec of the invention of the Fourier transform to make this change-of-basis possible and bec we find frequency domain analysis useful. Kinda makes one wonder whether math is invented or discovered.... Did we invent sinusoids or discover them?
The ear responds directly to what the driver is doing. The eardrum can be in only one place at a time, just like the driver. The bigger question is how the human ear can perceive many different frequencies simultaneously, but the answer is the same.
That's fascinating to know, especially when I can consider I listened to Paul say that through my new Beyerdynamic headphones. Each driver, 5Hz - 40,000kHz! I'm assuming headphone drivers can do this because of how light they are?
Great explanation, I used to buy dirt cheap 8" full range speakers from Radio Shack in 70's. Put them in wooden cabinets in a vehicle and they sounded better than car speakers of the day. Were they fantastic,no.
So I wonder IF it could be ideal (if possible), to have like 5, 7, or 33 broken down circuits with the same number of different sized speakers? So, normally the most I've seen I think is 4 way speakers with FOUR different frequency ranges and 4 different speakers, one to handle each. Would it "cost" too much power to separate it this much? Or would it actually potentially benefit? I'm guessing you COULD use a good midrange speaker type to handle each of many frequency ranges, right? So, let's call one speaker type a Sony 22m 2.5". You could have maybe 7 of these on a box or even panel, each one handling a separate frequency range. Would this actually produce technically "better" sound? Or would the difference between that and only 3-way be negligible?
I’m guessing that each speaker driver only plays its filtered portion of the incoming signal & if designed correctly the resultant sound of the all speaker driver is an accurate facsimile of the input signal. A simple example is the playback of a low frequency square wave which contains many (sine wave) frequency components thus is output through all of the speaker drivers (to varying degrees).
The problem with Fourier transforms is they require time accumulation. Which requires some level of steady state. The uncertainty tradeoff between temporal and spectral resolution (Gabor's principle) limits the minimum duration required for accurate pitch identification or discrimination. e.g. a half cycle is needed to detect frequency. The initial leading edge transient does not contain that much info.
My problem understanding is what about playing all the freqs at the same time..some freqs are gonna be in phase with each other and some wont.and every percent of cancelation and addition at the same time..always changing with the signal ..voltage and amperage. ( current) .. Its more complex than most even admit.
The speaker cone will move like that with 20 Hz and 20 kHz frequencies only if the amplitude of the 20 Hz is much higher than the amplitude of the 20 kHz. In the opposite scenario the movement will be quite different.
That's the problem with audio measurement specs. They are all based on steady state typically sine waves. That is because way back in the day that was the best we could produce for testing. Even producing a good sinewave was a challenge and a device to measure the results require a continually repeating signal. But music is seldom a boring single or even multiple steady state signal. It's well beyond a combined 20Khz and 20Hz steady state, It's an almost completely non-repetitive series of transients and harmonics combined into a very complex signal. If perfect, the speaker just moves based on whatever voltage is presented to it at that instant and it is constantly changing. Forget everything about individual frequencies, sine waves or simple combinations of. Think constantly changing non-repeating voltages that represent every sound at that point in time combined. Compared water flowing through a hose which is measurable in steady state laminar flow to a waterfall.
The way I see it is that you can have a thousand different frequencies playing simultaneously but they all superimpose to just one signal. The speaker is playing that one signal.
there had to be cancelation of motion of the cone according to this explanation . and this had to cause silencing effect of one frequency over the other . but it doesnt happen .
Imagine a system whereby each speaker can only produce sine waves of specific frequencies. That would be weird! And you would need a whole lot of speakers.
Depends on the Driver Quality (largely, the magnetic Strength.. which = greater control / accuracy). It seems that higher frequency drivers are far less prone to distortions, because they are lighter, and can move much faster. Where as accelerating a larger woofer, is much more challenging. If the woofer cant keep up with the acceleration demands.. you get micro-distortions (muddy-ness, where instruments and vocals blend in together, losing separation and details). I had a pair of 90s era "Techniques" speakers. They were 12" 3-ways. The tweeter and mids on them, were very low quality... but did enough to get the job done. The woofer was also low quality... with a small magnet. Id later picked up a 70's era set of "EPI 100v" speakers. The EPIs are a 2-way, with an 8" woofer... and the cabinet volume is like half the size of the Techniques. Yet... the EPIs utterly DESTROY the Techniques, in every single Metric. The EPIs have a deeper, and punchier bass (due to the woofers having a much larger and stronger magnet + stronger coils)... and the bass itself sounds far more "Musically Accurate". The EPIs inverted tweeter, is pure Magic. It produces a near Holographic 3D Soundstage / Image. The speakers sort of Vanish.. and its like the sounds are just coming out of thin air, without an actual Source projector. Its almost as if a band was in the actual room with you, playing Live. Now... the only drawback, is that the EPIs can not play quite as Loud. But when you hear how good the EPIs sound, you dont really care about the loss of a few Decibels. Its sort of like the difference between eating at a High end restaurant vs a cheap all-you-can-eat buffet. The buffet food is just edibly "OK"... where as the fancy restaurant dishes.. are "To die for". Pure Nirvana, in your Mouth. You would never know, that the EPIs lacked a separate midrange.. as the tweeter is so Magical, and Accurate. In fact, Ive heard NEW things in music that Id listened to my entire life. Such as... for the very first time... being able to fully understand what Lyrics the singer was singing (where as on other speakers, it gets too distorted to actually make out). Also, Ive heard speculation, that additional crossover components can alter and or distort the original signals. According to the designer of the EPI speakers... he used heavier gauge wire on the woofer coils.. to automatically filter out the highs, without the need for an additional crossover component... citing that reason. It may have also reduced the overall cost, too.. as there is less parts, and no need to solder / wire them.
Your eardrums cover all frequencies with a single diaphragm, a microphone covers all frequencies with a single diaphragm, and headphones cover all frequencies with a single diaphragm, so there is no reason speaker driver cant do the same.
I think that the maximum excursion is at the root of this. The eardrum and microphone diaphragm only move a small amount, so the doppler will be minimal, and the similarity in size and excursion means that it is unimportant. For a loudspeaker, the excursion can be significant. As I understand it, that is an advantage of a horn loaded driver as the driver diaphragm is working against a high acoustical load, this is akin to a transformer. If I am spouting garbage, please help me out here.
@Mantalban. Their vocal cords would have to be as smooth as rich corinthian leather. Singing extremely low frequencies and very high frequencies at the very same time?
I’m still kinda confused.. how can the speaker do 2 things at once (the 20hz and 20khz example) while u said it could only do one thing at a time.. If I try to understand how it works I guess a musical song contains ALOT of moving data huh? 🤔😅 Btw this animation is quite clear (although I still don’t completely understand how it can produce different frequencies at the same time). ua-cam.com/video/RxdFP31QYAg/v-deo.htmlsi=jO8iN-l1jBUhqIi3
My PearlAcoustics Sibelius single driver work just fine, sound excellent to me. The absence of a cross over with all sorts of artifact possibilities, and the coherence of the sound is a major plus for single driver speakers. From a fellow Swiss, Jä e Basler Bebbi.
@@danielgeiger7739 Auch ein Breitbänder benötigt ein elektrisches Korrekturnetzwerk um zumindest den Bafflestep zu kompensieren. Vielleicht hat dein PerlAcoustics Sibelius sogar so eines verbaut und du weisst es nur nicht? ;-) Grüsse aus dem Bernbiet.
@@SinusPrimus Amp - binding post - voice coil. That's the signal path in the Sibelius. I have not cut the speaker open, but at least that is what website indicates as well as Harley's very nice videos on UA-cam (Pearlacoustics channel). Very much worth the time. Re baffle step (roll off of low frequencies), Sibelius uses as Voight pipe design of the cabinet. You could consider that the "correction network" but it is not in the signal path, it is not electronic. It is careful internal cabinet design.
@@danielgeiger7739 Danke für den Hinweis. Theoretisch wäre eine rein akustische Kompensation ideal. In der Praxis setzt der Bafflestep aber nicht erst bei sehr tiefen Frequenzwein ein, sondern bei Schallwänden dieser Grösse bzw. Schmalheit im Bassmitteltonbereich. Also kann diese interne Schallführung das Problem nicht genügend oder nur ansatzweise beheben. Vielleicht schaue ich noch nach, ob es Messdiagramme gibt, die das belegen...
This is one of Paul's greatest answers, and he's had so many great ones. Well done!
That question has plagued me for 50 years. Thanks for solving the mystery for me. 👍👍👍👍👍👍👍👍
you've addressed this eleventeen times.... but THIS analogy _finally_ unlocked the door for my brain. thank you.
Hi Paul!
I really enjoy listening to and seeing you here in Portugal and I have learned a lot about HiFi from you. If my CEO knew as much about our "products" as Paul knows about his area, it would be a blessing.
Stay healthy in the company of family and employees. Thank you very much! Fernando
Most likable guy on UA-cam. Great video as usual!
Loved this explanation, one mystery 99.99% solved. Best, J
Well answered Paul ... you started to struggle at first 🤣.. but you nailed it 👍
So glad this was asked and answered
I've been wondering about this for decades. Thanks to both the question asked and the answer!
Always a pleasure to hear a clear, simple explanation of a sound device!
Absolutely brilliant - I had often wondered how a speaker accomplishes making multiple frequencies simultaneously. Thanks a million for explaining it so well 🙂
Paul, I thought you did a great job explaining this, as I understood it perfectly!
The fundamentals of this is that the diaphragm (to the best of its ability) just follows the drive signal (whatever it might be) - show the signal. Thanks, Paul, for another great video.
taking the loudspeaker components , the speaker fixed magnet , the voice coil suspension flexible holder , the voice coil attached to the speaker cone , taking delta T sec . , time constant , the voice coil , will oscillates the summations of all the fundamental frequencies , thus emitting all the fundamental frequencies from speaker cone , the overall effects is the integral summations of the audio complex wave over a unit time .
Medical student here, currently studying ENT. This question was bugging me so much. Thanks a lot Mr. McGowan!
That would be something for these high-speed camera freaks, to film a loudspeaker with thousands of images every second and to "fire" the loudspeaker first with one sound (one frequency), then two, and so on .. ;)
Paging @theslowmoguys to the white courtesy phone.
Absolutely brilliant! Fascinating question tackled in your inimitable way. Thanks so much!
we can not get enough of this question, i need all the angles to wrap my head around the answer.
one way to explain it: imagine the outside of your house, all the sounds all around you, they are so many frequencies there and there are only one eardrum to pick it up.
all the sounds are merging together making a complex wave
This is how I have understood it as well. This is why you get a single waveform in a music file even when the sounds are quite complex.
Inside your ear though, there are tiny hairs of various lengths, which resonate at particular frequencies, and that is actually how we hear sound as a mixture of frequencies. The idea that sound is composed of different frequencies mostly comes from how we perceive it biologically. The mathematical technique of Fourier transform basically does what our ear does--that is, convert a complex wave into its composite frequencies and amplitudes (just like our ear!)
Great question and answer, Paul. I've been listening to reproduced music all my life and even used to build speakers in the '70's and now I know why I never got the sound I was looking for. I didn't completely understand how the component speakers/horns did their work and why crossover design and construction is so critical to how they perform. Thanks!
Well said Paul. It’s something I had guessed but never really knew for sure. Thanks!
One of the most enlightening explanations you have provided. Thanks Paul!
I honestly thought you were gonna crash n burn with this one but you pulled it out.
Respect. I was going to go in the direction of intermod.
Intermod is just two steady state sine waves. Again not accurately representing an actual typical sound signal. But way back in the day test and measurement equipment was very limited and adding two signals together was an advancement from using only one, like THD. Later changing the 2nd signal to a squarewave for TIM. Transient Intermodulation Distortion.
Great explanation, I had the same question and did not understand how it could do both. Thank You very much.
Paul, you did it, I had this question too. I had suspicion of the answer as you had put it but was just guessing. Thank you for that answer.
Until know I never asked myself that question. However this is very interesting to know. Each time we learn something we become richer persons.
Thanks!
My realization of the doppler effect came from being in a 'garage band'. What happened was plugging the mic into a bass guitar amp produced wild as heck modulations to the singers voice because the vocal frequencies were 'riding' on the much slower bass movements of the speaker cone. It was similar to the sound you can make whilst gargling and humming at the same time.
Your eardrums, speakers, cables, signal paths through electronics and microphones all transfer music as a sum of waves represented as air pressure, mechanical motion (e.g. in a speaker cone), magnetic field (e.g. in a voice coil or cartridge), electric field (e.g. in electrostatic speakers) or voltage.
The true magic is how our hearing works with the cochlea in our inner ear sort of disassembling the sum of sound waves to individual nerve pulses that our brain can process as those same separate frequencies that went into the music to begin with.
INteresting. SO this then raises the question if your ear sums the signal anyways is having the signal broken up by different drivers really better than a single driver producing an already summed pressure wave ?
Size and mass make no difference?
@@Canadian_Eh_I Technically a single point sources is always best. But as Paul states, speakers not being point sources have other issues. In order for a speaker to produce lows it has to have some combination of larger size and/or greater excursions. To produce highs it has to be smaller and/or able to move very slightly but accurately without producing that signal in multiple spots at different phases. Thus very different design parameters for each.
@@glenncurry3041 I have a pair of KEF LS60 in one room that I find quite awesome with near single point performance. Right now I'm listening in my home office setup using a pair of KEF LSX II speakers connected via USB to my iMac 27" and a reworked/modified Velodyne servo subwoofer (LSXII includes subwoofer output) giving a very clean sub bass flat down to 20Hz. The Uni-Q drivers in my KEF speakers are 2-way and probably the closest you can get to a single point source. The subwoofer in this room blends in perfectly and people I've demoed this office setup tell me the bass is amazing and when I tell them it's thanks to a servo subwoofer, please find it, they didn't find it. It's a bummer servo subwoofers are so rare.
For example the driver is vibrating .01mm for the high frequency as it travels the 10mm for the lower frequency.
Additive and subtractive mixing on the recording means the speaker travels the mean of all of the frequencies mixed together at any point in time.
Watching music on an oscilloscope will give you a visual representation of what is going on.
Been looking for an answer to this obvious question for over a quarter of a century.
Thanks.
That's why we have intermodulation with full range speakers, separating drivers helps to avoid that, but i think there are some speakers manufactured in a way that at a certain point the higher frequencies just oscillates at the center, they use different materials and shapes to provoke a mechanical behavior.
Waveforms can be added or subtracted. When you put two waveforms of different frequencies together, you’ll get their sum/difference as some will cause constructive or destructive interference and your resulting waveform will look like the combination of these two. Same way your hearing works. And if you look at that resulting waveform with the scope, and use Fourier transformations (decomposing the waveform into its components) you’ll get the original two waveforms. The speaker moves the same way as the combination of the two waveforms. Your eardrum will also do the same thing if you listened to two waveforms from two separate speakers. If you shine a laser onto your eardrum to reflect onto a surface, you’ll see the combined waveform of the two individual waveforms coming from two separate speakers being combined by your ear drum. Kind of like colors of light, your eye sees as combined but prism can separate them into its components.
Actually eyes see differences between red/green/blue and the differences between the intensities of those three are interpreted by our brain as a gamut of colors.
When you have a multiple driver system, where does the decoupling into different waves lengths take place?
Hi Christian. Nice explanation.
@@larazss3254 Superposition holds, so all waves are added together, and their sum can be decomposed into a resulting frequency spectrum. This spectrum assumes steady-state conditions, so while the music is limited in time, the components started infinite time ago, so it is abstraction whenever we go to the frequency domain.
@@larazss3254 you get bass frequencies go to subwoofer, mids go to woofer and highs go to tweeter. That is usually done in the cross over. Sometimes the frequencies are separated in the DSP, and go to individual amplifiers to power the different drivers like in a bi-amp application. But still you can have multiple frequencies of mid range or low or highs getting to one mid or low or high range speaker as they may be in the same group as cross over separates by groups which are ranges like bass 20-125Hz, mid 125-3000Hz, high 3000-20,000Hz. So you may end up with two 5000Hz and 6000Hz going to the same speaker which would play the sum of both waves. If you have 30Hz and 6000Hz, these two would go to separate speakers separated by the cross over. However, when they arrive at your ear or microphone, they will sum up and your ear drum or microphone membrane will vibrate summing both of them. It would look like a big wave of 30Hz with small wave on top of the large wave. Kind of like if you got large waves on a lake and when quick wind gust blow over the lake, you get little tiny waves in top of the large waves looking like little shivers. I hope that kind of helps visualizing it.
@@christianratajczak2809 Thank you, that is exactly what Paul explained. But it is still hard to imagine, how analog crossovers can precisely identify and carve out these small waves on top of the large ones...
Thanks Paul this really helped my understanding of speakers 😊
I think it is easier to say that the speaker responds to a varying voltage that is a composite of all the frequencies in a given signal. A properly operating speaker is merely reproducing the waveform it is presented with.
I think he should have drawn two sine waves, added them up and show the speaker only has to respond to the sum of both waves.
Great explanation - that issue has baffled (pun) me for ages
Thanks for answering my question! I was aware of how the diaphragm moves at multiple frequencies so I was more so asking how that's possible when there is only one magnetic field at a time. Based on what you said, it sounds like the answer is because the magnetic field changes so quickly. Thanks!
The field driving the coil has the overlapped frequencies already composited into the voltage it gets, by the audio circuitry. Whats also interesting is how your ear takes the single overlapping wave and splits it out into frequencies using a biological structure.
@@KK-dv3wh The human brain is just wild when it comes to things like this. 😂
I'm not sure if I should call it an issue or be thankful that the alternative isn't happening, but all the rooms reverb is compiled to one (muddy) signal by the brain as well (instead of multiple clear echos).
Room treatment for the win! ✌️😄
Simple answer is we typically do not listen to a single repetitive sine wave nor combinations of a couple. It becomes unlocking your mind from that constraint. Music is constantly changing, non-repetitive complex wave.
That is a hell of a question
Great question and answer to something I myself was pondering just a few days ago
Audio itself explained. Thank you as always Paul. : )
Thanks , Paul ! I had always thought of multiple frequencies simultaneously as intermodulation distortion , but Doppler does seem a more accurate descriptor .
I imagine a line array with 20 drivers and 19 crossovers would be silly to keep each driver in a "sweet" half octave . That would be a lot of expensive capacitors !
Wow I always feel like I’m in school with these questions, only now I really like it 😆✌🏻📀! Thanks Paul
treating it in a mathematical equations , a complex wave is a summations of individual frequencies ,alternatively known as Fourier Series a summations of sine 2*pi*f or cosine 2 *pi* f , where f is the individual frequencies . ( followed physical explanations - later explained ) .
the explanations of the multi frequency audio radiations of a loudspeaker , is best described by , arithmetic summations of the various audio fundamental frequencies , as expressed in the Fourier Series expressions of the Complex Audio Signal with the respect to Time .
I figured it works like runners around a track: runners on the inside of the track will circle faster than the outside runner if they didn’t start staggered. The higher frequencies are coming off the area of the cone nearest to the middle, lower frequencies from the outer closest to the edge.
Well done! Better understanding of how that all works.
I JUST spoke about this in my recent videos too... related to complex music.. and quick changes.. transient response.. articulate movement.. damping.. etc etc..
A pretty good explanation. I think this needs a 3d animation to show how these waves are simultaneously present.. Its harder to speak to and easier to show a graphic.
Very good and important for people to understand.. say... reasons why we have...2way.. 3way.. 4 way..5 way.. etc...
Also... would be nice to see how these frequency and signals are processed in the cross overs.. as in.. how the electrical signals / Frequencies are ' separated ' in the wires / caps / resistors. 🙏🙏🙏🙏🙏🙏
Thanks, Paul. ..Education is our Primary step in any journey.
(Also.. Great to meet Chris at Axpona - we spoke briefly about the FR5 - " issue " / demo model performance)
Thanks! Could you please explain further how one can tell the difference in instrument (like an oboe vs. a violin playing the same note) when played by the same driver?
Excellent explanation Sir👍
This question is a prime example of "a picture is worth a thousand words. A simple view of the audio signal on an oscilloscope would answer the question. DRIVERS CAN PLAY MULTIPLE FREQUENCIES AT THE SAME TIME.
I like the explanation for the question. Re single drivers and Doppler effect, don't buy it given the small diver excursion and frequencies involved. Show the math, please. Have a single driver speaker myself -- PearlAcoustics Sibelius SG -- and love it. Exquisitely detailed and clean sound with wonderful soundstage. As there is no need for a cross over, and there is a single source point for all sound solves a lot of other problems. KISS principle comes to mind.
Doppler distortion is largely inconsequential at the excursion levels of a well designed full range unit. Research has shown that if kept within 2cm it can be largely disguarded as a problem. If you think about a passing ambulance it has to travel some distance before we hear the drop in frequency caused by the doppler effect. All coil drivers exhibit the same behavior with respect to their bandwidth,where the lower limit is set by either higher mass and/ or compliance from their suspension (box alignment accepted) and the upper frequencies being a result of flexing within the cone structure around it's centre.Even the exotic rigid attempt materials do this and it's thanks to that behaviour that their higher frequency extension is simultaneously attainable.
2cm wavelength is 80Khz. So yes if your single driver design is flat, linear out to 80Khz,.... otherwise operating in non-linear areas causes distortions.
@@glenncurry3041Hi Glen. On the doppler effect topic,the modulation is proportional to the speed of the cone at it's low frequency excursion where the travel is greatest. A 2cm excursion (not wavelength) is quite alot of travel and it will be lower in the majority of units,full range or otherwise at " typical " listening levels,so can be largely discounted. The 80khz frequency mentioned will require extremely low mass,and/or dimensions and excursion, not to mention how far that frequency is beyond audibility, so has no relevance at all with respect to the doppler effect.
always wondered this, thanks!
If you open an audio file in an audio editor (like Audacity) and you zoom in, you get a pretty good idea of the speaker movement.
Doppler distortion hadn't crossed my mind, but makes complete sense.......
Particularly noticeable when whizzer cones are used.
A voicecoil can vibrate at multiple frequencies. Old style "full range" speakers use to have small secondary cones glued to the dust cap to better capture and expel high frequencies from the voicecoil.
Great explanation!!!
Very good explanation, thank you. Paul, a question... I know that the human ear is capable of hearing a range of frequencies that goes from 20 Hz to 20 KHz. I also know that these are four speakers (Tweeter, MidRange, MidBass and Subwoofer) that, “together”, are capable of reproducing the entire range of frequencies mentioned. But I ask... Each of the four speakers mentioned reproduces, individually, which range of frequencies? In other words, the Tweeter is responsible for reproducing, individually, which range of frequencies? What about MidRange? What about MidBass? And the subwoofer? I would just like to have a general idea of the frequency range that each speaker works individually. It's a generic question, as I deduce that there is no fixed rule for this. But I would just like to have a rough idea.
I remember back in the late 80’s I think, that Pioneer patented and launched a set of speakers that used some new technology to ensure that a speaker could output different frequencies without distortion. I have searched for years for these speakers or any reviews, but to no avail.
Yep, that does help.
thanks Paul
I think woofer beaming should be addressed: above a certain frequency, the sound is coming straight at you, and has poor dispersion. The tweeter still has wide dispersion. I became aware of a new Mission speaker that uses a 2.9 kHz crossover point with an 8 inch woofer, while MoFi tweeters use 1.6 kHz. Wouldn't this create a problem for the listener if they're not seated right in front of the Missions? The part I'm not sure about is, the effect of the driver's dust cap: would that improve the dispersion, making the traditional calculation between the woofer diameter and the wavelength unreliable? Someone in a forum had brought this up recently. I had been using Brent Butterworth's article about what is the ideal speaker driver configuration as a guide, previously.
Fascinating!
Great explanation. That big magnet made me wonder whether bigger is always better with speakers. Are there downsides to the largest magnets?
You forgot to mention that cone design factor and the physical attributes of said cone . Higher frequencies in a driver due to speed are released from center out . Hence putting a whizzer cone at dust cap center . The center moves faster than the outer edge because there is less area . A good /bad cone design is essential to its ability to produce proper frequencies thought a certain spectrum .
Expressing a time signal (that's what's actually reproduced) in terms of a sine of sums is simply a mathematical transformation expressing tne signal in a new basis. That snusoids (in fact, you also need phases) can form a basis (i.e. any sine+base be converted to time and vice-versa) is the actual physical property. We use sines bec of the invention of the Fourier transform to make this change-of-basis possible and bec we find frequency domain analysis useful. Kinda makes one wonder whether math is invented or discovered.... Did we invent sinusoids or discover them?
The ear responds directly to what the driver is doing. The eardrum can be in only one place at a time, just like the driver. The bigger question is how the human ear can perceive many different frequencies simultaneously, but the answer is the same.
That's fascinating to know, especially when I can consider I listened to Paul say that through my new Beyerdynamic headphones. Each driver, 5Hz - 40,000kHz! I'm assuming headphone drivers can do this because of how light they are?
Please empty that C: Drive D: ! Love you Paul
Just like the air in the room with live music can hold many frequencies at the same time.
Good answer
Great explanation, I used to buy dirt cheap 8" full range speakers from Radio Shack in 70's. Put them in wooden cabinets in a vehicle and they sounded better than car speakers of the day. Were they fantastic,no.
So I wonder IF it could be ideal (if possible), to have like 5, 7, or 33 broken down circuits with the same number of different sized speakers? So, normally the most I've seen I think is 4 way speakers with FOUR different frequency ranges and 4 different speakers, one to handle each. Would it "cost" too much power to separate it this much? Or would it actually potentially benefit? I'm guessing you COULD use a good midrange speaker type to handle each of many frequency ranges, right? So, let's call one speaker type a Sony 22m 2.5". You could have maybe 7 of these on a box or even panel, each one handling a separate frequency range. Would this actually produce technically "better" sound? Or would the difference between that and only 3-way be negligible?
How do these questions get asked? Do we just drop them here?
I’m guessing that each speaker driver only plays its filtered portion of the incoming signal & if designed correctly the resultant sound of the all speaker driver is an accurate facsimile of the input signal. A simple example is the playback of a low frequency square wave which contains many (sine wave) frequency components thus is output through all of the speaker drivers (to varying degrees).
I think Fourier explained it better!
Looks like the be PS Audio Sub is getting close 👍🏻
The diaphragm produces a waveform that is broadband. The answer requires an understanding of the Fourier expansion.
The problem with Fourier transforms is they require time accumulation. Which requires some level of steady state. The uncertainty tradeoff between temporal and spectral resolution (Gabor's principle) limits the minimum duration required for accurate pitch identification or discrimination. e.g. a half cycle is needed to detect frequency. The initial leading edge transient does not contain that much info.
Cheers from Portugal 🎉😂🔈🎸🎶
My problem understanding is what about playing all the freqs at the same time..some freqs are gonna be in phase with each other and some wont.and every percent of cancelation and addition at the same time..always changing with the signal ..voltage and amperage. ( current) .. Its more complex than most even admit.
The speaker cone will move like that with 20 Hz and 20 kHz frequencies only if the amplitude of the 20 Hz is much higher than the amplitude of the 20 kHz. In the opposite scenario the movement will be quite different.
Superimpose the waveforms of two sine waves at different frequencies on an oscilloscope. Done
That's the problem with audio measurement specs. They are all based on steady state typically sine waves. That is because way back in the day that was the best we could produce for testing. Even producing a good sinewave was a challenge and a device to measure the results require a continually repeating signal.
But music is seldom a boring single or even multiple steady state signal. It's well beyond a combined 20Khz and 20Hz steady state, It's an almost completely non-repetitive series of transients and harmonics combined into a very complex signal. If perfect, the speaker just moves based on whatever voltage is presented to it at that instant and it is constantly changing.
Forget everything about individual frequencies, sine waves or simple combinations of. Think constantly changing non-repeating voltages that represent every sound at that point in time combined.
Compared water flowing through a hose which is measurable in steady state laminar flow to a waterfall.
frequency addition.
The way I see it is that you can have a thousand different frequencies playing simultaneously but they all superimpose to just one signal. The speaker is playing that one signal.
The best answar must be i'ts mowing like our eardrum😊
there had to be cancelation of motion of the cone according to this explanation . and this had to cause silencing effect of one frequency over the other . but it doesnt happen .
The same way that the audio signal driving the speaker contains multiple frequencies. This is college level math: Fourier Transform.
his hands explain it all
Imagine a system whereby each speaker can only produce sine waves of specific frequencies. That would be weird! And you would need a whole lot of speakers.
intermodulation distortion
That is why we need it 3 way speakers .
Surprised that PS Audio makes 2 way speakers for HI End Audio .
Depends on the Driver Quality (largely, the magnetic Strength.. which = greater control / accuracy). It seems that higher frequency drivers are far less prone to distortions, because they are lighter, and can move much faster. Where as accelerating a larger woofer, is much more challenging. If the woofer cant keep up with the acceleration demands.. you get micro-distortions (muddy-ness, where instruments and vocals blend in together, losing separation and details).
I had a pair of 90s era "Techniques" speakers. They were 12" 3-ways. The tweeter and mids on them, were very low quality... but did enough to get the job done. The woofer was also low quality... with a small magnet. Id later picked up a 70's era set of "EPI 100v" speakers. The EPIs are a 2-way, with an 8" woofer... and the cabinet volume is like half the size of the Techniques. Yet... the EPIs utterly DESTROY the Techniques, in every single Metric.
The EPIs have a deeper, and punchier bass (due to the woofers having a much larger and stronger magnet + stronger coils)... and the bass itself sounds far more "Musically Accurate". The EPIs inverted tweeter, is pure Magic. It produces a near Holographic 3D Soundstage / Image. The speakers sort of Vanish.. and its like the sounds are just coming out of thin air, without an actual Source projector. Its almost as if a band was in the actual room with you, playing Live.
Now... the only drawback, is that the EPIs can not play quite as Loud. But when you hear how good the EPIs sound, you dont really care about the loss of a few Decibels. Its sort of like the difference between eating at a High end restaurant vs a cheap all-you-can-eat buffet. The buffet food is just edibly "OK"... where as the fancy restaurant dishes.. are "To die for". Pure Nirvana, in your Mouth.
You would never know, that the EPIs lacked a separate midrange.. as the tweeter is so Magical, and Accurate. In fact, Ive heard NEW things in music that Id listened to my entire life. Such as... for the very first time... being able to fully understand what Lyrics the singer was singing (where as on other speakers, it gets too distorted to actually make out).
Also, Ive heard speculation, that additional crossover components can alter and or distort the original signals. According to the designer of the EPI speakers... he used heavier gauge wire on the woofer coils.. to automatically filter out the highs, without the need for an additional crossover component... citing that reason. It may have also reduced the overall cost, too.. as there is less parts, and no need to solder / wire them.
Implementation is key here. Better a good implemented one driver than bad multi. The Sibelius speakers for instance look promising.
Your eardrums cover all frequencies with a single diaphragm, a microphone covers all frequencies with a single diaphragm, and headphones cover all frequencies with a single diaphragm, so there is no reason speaker driver cant do the same.
I think that the maximum excursion is at the root of this. The eardrum and microphone diaphragm only move a small amount, so the doppler will be minimal, and the similarity in size and excursion means that it is unimportant.
For a loudspeaker, the excursion can be significant. As I understand it, that is an advantage of a horn loaded driver as the driver diaphragm is working against a high acoustical load, this is akin to a transformer.
If I am spouting garbage, please help me out here.
Terribly hard to explain.
Looking at an audio waveform on an oscilloscope helps a little, but Only a little.
Maybe those of us who sing, can someday learn to sing baritone AND falsetto at the very same time. If speakers can do it why can't our vocal cords?
Actually, some people can. It's called overtone
@Mantalban. Their vocal cords would have to be as smooth as rich corinthian leather.
Singing extremely low frequencies and very high frequencies at the very same time?
I’m still kinda confused.. how can the speaker do 2 things at once (the 20hz and 20khz example) while u said it could only do one thing at a time..
If I try to understand how it works I guess a musical song contains ALOT of moving data huh? 🤔😅
Btw this animation is quite clear (although I still don’t completely understand how it can produce different frequencies at the same time).
ua-cam.com/video/RxdFP31QYAg/v-deo.htmlsi=jO8iN-l1jBUhqIi3
That‘s one reason why fullrange speakers don‘t work. Greetz from 🇨🇭
My PearlAcoustics Sibelius single driver work just fine, sound excellent to me. The absence of a cross over with all sorts of artifact possibilities, and the coherence of the sound is a major plus for single driver speakers. From a fellow Swiss, Jä e Basler Bebbi.
@@danielgeiger7739 Auch ein Breitbänder benötigt ein elektrisches Korrekturnetzwerk um zumindest den Bafflestep zu kompensieren. Vielleicht hat dein PerlAcoustics Sibelius sogar so eines verbaut und du weisst es nur nicht? ;-) Grüsse aus dem Bernbiet.
@@SinusPrimus Amp - binding post - voice coil. That's the signal path in the Sibelius. I have not cut the speaker open, but at least that is what website indicates as well as Harley's very nice videos on UA-cam (Pearlacoustics channel). Very much worth the time.
Re baffle step (roll off of low frequencies), Sibelius uses as Voight pipe design of the cabinet. You could consider that the "correction network" but it is not in the signal path, it is not electronic. It is careful internal cabinet design.
@@danielgeiger7739 Danke für den Hinweis. Theoretisch wäre eine rein akustische Kompensation ideal. In der Praxis setzt der Bafflestep aber nicht erst bei sehr tiefen Frequenzwein ein, sondern bei Schallwänden dieser Grösse bzw. Schmalheit im Bassmitteltonbereich. Also kann diese interne Schallführung das Problem nicht genügend oder nur ansatzweise beheben. Vielleicht schaue ich noch nach, ob es Messdiagramme gibt, die das belegen...
"Mahopac" -- pronounced "May-o-Pak".
Its all just noise. Its your ears and brain hat separates out the different sounds
This is silly. Everyone knows a single woofer diaphragm can't produce two sounds at the same time. It's why the IRS V needs so many drivers...
JK 😛
Pronounced Mayopack
First
ua-cam.com/video/cH8iv4n2kco/v-deo.html