If you're working with regular PCM signals, then half the signal is (roughly) -6dB. If you're working with squared quantities (eg power, RMS) then half is -3dB (essentially all we're doing is dividing the result by 2 to undo the exponent once in log-domain, which saves having to take a square root first). This way you get the same dB values independent of whether you squared your signal first. Since this probably refers to the problem of adjusting and/or measuring regular signals (as they are sent to an audio interface) the 20log10 or ~6dB rule is what you want.
20log10(2)= 6.02 dB, for voltage level in the mixer, as it is easy to verify summing two equal signals at -6 dB and getting 0 dBFs on the master bus. Thinkink to the pressure, and so for the power, for the same math it is to see a factor 3 instead of 6 about the dB spl. ;-)
Wytse, I really dig your channel but you need to be corrected here. 3db means doubled performance, 6db means doubled SPL and depending on the frequency content 10db means doubled perceived Loudness for the human ear (which isn't reliable as so many human factors come into play)
A couple of quick points. As dB is a ratio scale, -3dB on your DAW will always equal -3dB on your console (or any other equipment that that the same signal passes through). It doesn't matter which scale (dBV, dbFS etc) that the equipment uses for this, as it is just halving the volume of the signal, which all pieces of equipment will measure as -3dB. ... Also 0 dbSPL is supposed to be the quietest thing that a human can hear, which is the source of that scale. This was set at 20 micro pascals by some audio boffins once upon a time. Talking of humans, we perceive +10dB as doubling the volume and =10dB as halving it. This means that you have to up the volume by around 10x to double the perceived volume and this is why you need LOTS of power for PA type sound. To double the loudness of a 200W system, you need 2000W! Hope this helps someone. If not, meh.
hi! there are few things to lightly correct here. double loudness is not increase of 3dB. It is actually increase of +10db. you used the wording that if we want to hear something "twice as loud". so in your case it would mean 90db is "twice as loud" as 80dB. +3dB is double power as in watts (used with dBm measurement) and +6dB is double amplitude as in voltage (used in dBV and dBu). also, reference for dbu is 0.775 volts. dbSPLC is also used in measurements. its curve is largely flat as apposed to dbSPLA which mimics human ear.
As long as signal stays inside the computer it can go over to plus dB FS. There is no clipping inside most daws and generous headroom in many encoding formats.
j7ndominica0 I should have read this before commenting. It’s basically the same as my point with the caveat that you are recording in floating point. In finite/fixed point, 0 is the maximum, no exceptions.
Thanks a lot for the explanation. How do I do I use compressors to get the best performance in a live outdoor setting? Better yet, how exactly should I connect it in my sound rack? I have an analogue mixer, Amplifiers, equalizer for each amplifier(bass amp, low-mid amp, high-mid amp etc.) and a BEHRINGER ULTRAMIZER PRO. A video explaining all these would be nice. Don't mind an email either.
Kajene Dawkins If your mixer has insert points use it as an insert effect. If your mixer doesn't put it right after the external mic preamp. If you use the mic preamps of your mixer and it doesn't provide insert jacks you have no chance to use compression on each channel. The only way to use a compressor in this case is to use it as a master compressor right after the main outs and before the PA-amp.
dB(a) is often used in building engineering for measuring sound insulation in buildings. I.e. Building regulations state the sound insulation performance between two rooms (or houses) in dB(a). For appartment bulding the requirement is +5 dB. I didn’t know there where so many different dB scales, I only knew of the A weighted scale.
Thanks, clear and useful. Just one thing, for pro studio application you speak about mode A in SPL measure, but I knew for monitor calibration it is used C curve (an Slow time constant), isn't it ? :-)
I'm still confused! I read that when dealing with power, measured typically in watts, then a 3db increase yields a doubling of power. But when dealing with voltage or SPL a 6db increase doubles voltage or sound pressure.
For floating point, non-clipped audio (up to 0 dBFS) is in the range from -1 to 1. 0 is silence. Adding 1 dB is multiplying by ~1.122. Subtracting 1 dB is multiplying by ~0.891.
Hello dude, One of the first thing who surprise me when i was at audio school, it's the differents kinds of db... i was thinking that there is only one Bel meter. Lol. Thanks for this interresting point in vidéo. encouragement from France. If i well remenber, -18 dbfs digital = 0dbu Analog? huhu disturbing rules. Finaly the good master level of the software, and the master level of the console is? (i know it's depend of kind of music, and no real rule on it but...)
how are the differents types of dB useful in mixing and mastering, is it any difference between the two? cause i see lots of mastering plug ins use different types of measurements than in the mixing stage, any input about this topic?
If you're working with regular PCM signals, then half the signal is (roughly) -6dB. If you're working with squared quantities (eg power, RMS) then half is -3dB (essentially all we're doing is dividing the result by 2 to undo the exponent once in log-domain, which saves having to take a square root first). This way you get the same dB values independent of whether you squared your signal first. Since this probably refers to the problem of adjusting and/or measuring regular signals (as they are sent to an audio interface) the 20log10 or ~6dB rule is what you want.
20log10(2)= 6.02 dB, for voltage level in the mixer, as it is easy to verify summing two equal signals at -6 dB and getting 0 dBFs on the master bus. Thinkink to the pressure, and so for the power, for the same math it is to see a factor 3 instead of 6 about the dB spl. ;-)
Yes, correct. But a perceived doubling is, according to many people incl me, 10 dB (PCM).
Wytse, I really dig your channel but you need to be corrected here. 3db means doubled performance, 6db means doubled SPL and depending on the frequency content 10db means doubled perceived Loudness for the human ear (which isn't reliable as so many human factors come into play)
A couple of quick points. As dB is a ratio scale, -3dB on your DAW will always equal -3dB on your console (or any other equipment that that the same signal passes through). It doesn't matter which scale (dBV, dbFS etc) that the equipment uses for this, as it is just halving the volume of the signal, which all pieces of equipment will measure as -3dB.
...
Also 0 dbSPL is supposed to be the quietest thing that a human can hear, which is the source of that scale. This was set at 20 micro pascals by some audio boffins once upon a time. Talking of humans, we perceive +10dB as doubling the volume and =10dB as halving it. This means that you have to up the volume by around 10x to double the perceived volume and this is why you need LOTS of power for PA type sound. To double the loudness of a 200W system, you need 2000W!
Hope this helps someone. If not, meh.
hi! there are few things to lightly correct here. double loudness is not increase of 3dB. It is actually increase of +10db. you used the wording that if we want to hear something "twice as loud". so in your case it would mean 90db is "twice as loud" as 80dB. +3dB is double power as in watts (used with dBm measurement) and +6dB is double amplitude as in voltage (used in dBV and dBu). also, reference for dbu is 0.775 volts. dbSPLC is also used in measurements. its curve is largely flat as apposed to dbSPLA which mimics human ear.
As long as signal stays inside the computer it can go over to plus dB FS. There is no clipping inside most daws and generous headroom in many encoding formats.
j7ndominica0
I should have read this before commenting. It’s basically the same as my point with the caveat that you are recording in floating point. In finite/fixed point, 0 is the maximum, no exceptions.
Great introduction. Would have loved if you went into a bit more details tho!
Thanks a lot for the explanation.
How do I do I use compressors to get the best performance in a live outdoor setting? Better yet, how exactly should I connect it in my sound rack?
I have an analogue mixer, Amplifiers, equalizer for each amplifier(bass amp, low-mid amp, high-mid amp etc.) and a BEHRINGER ULTRAMIZER PRO. A video explaining all these would be nice. Don't mind an email either.
Kajene Dawkins If your mixer has insert points use it as an insert effect. If your mixer doesn't put it right after the external mic preamp. If you use the mic preamps of your mixer and it doesn't provide insert jacks you have no chance to use compression on each channel. The only way to use a compressor in this case is to use it as a master compressor right after the main outs and before the PA-amp.
Very cool man! I learnt something from that. What is that mixing desk to your side? Looks awesome!
dB(a) is often used in building engineering for measuring sound insulation in buildings. I.e. Building regulations state the sound insulation performance between two rooms (or houses) in dB(a). For appartment bulding the requirement is +5 dB. I didn’t know there where so many different dB scales, I only knew of the A weighted scale.
Thanks, clear and useful. Just one thing, for pro studio application you speak about mode A in SPL measure, but I knew for monitor calibration it is used C curve (an Slow time constant), isn't it ? :-)
impedence is basically frequency depending resistance
I'm still confused! I read that when dealing with power, measured typically in watts, then a 3db increase yields a doubling of power. But when dealing with voltage or SPL a 6db increase doubles voltage or sound pressure.
Yes, but what about floating point? You can have theoretical +1dbFS since the scale is relative, and... well... more complicated than is worth typing.
For floating point, non-clipped audio (up to 0 dBFS) is in the range from -1 to 1. 0 is silence. Adding 1 dB is multiplying by ~1.122. Subtracting 1 dB is multiplying by ~0.891.
Wait... isn't 6 dB difference roughly equivalent to twice the power? Why do you say it's 3 dB instead? What's going on?
it's a doubling per 6dB if we are talking amplitude, and 3dB if we're talking power.
If you combine two similar signals as in parallel compression, 6dB. Typically doubling power equates to 3dB.
Hello dude,
One of the first thing who surprise me when i was at audio school, it's the differents kinds of db... i was thinking that there is only one Bel meter. Lol. Thanks for this interresting point in vidéo. encouragement from France. If i well remenber, -18 dbfs digital = 0dbu Analog? huhu disturbing rules. Finaly the good master level of the software, and the master level of the console is? (i know it's depend of kind of music, and no real rule on it but...)
how are the differents types of dB useful in mixing and mastering, is it any difference between the two? cause i see lots of mastering plug ins use different types of measurements than in the mixing stage, any input about this topic?
Awesome Video!
soooo where dbRMS?
But then what are these Lufs?
kossa gubben thats general loudness of a track mesured over a period of time.
Similar to dBFS except with a frequency-weighting curve applied like in dBA. The exact curve is not the same, but the concept is.
My god I guess I'll chime in. I always hear 20 db was twice perceived loud??? am i wrong??????
Young Wytse