flat assembler
Message board for the users of flat assembler.
Index
> Heap > What is the Correct Sound Mixing Math? 
Author 

rocketsoft
An open question...
normally they are just added: but:1+1=2 volts gives 1+1=4 watts!(not correct!) yet everyone uses additive mixing! 

14 Aug 2017, 09:41 

revolution
Furs wrote: That's because amplitude has both positive and negative values. 

14 Aug 2017, 16:13 

Furs
Ehm, I wasn't speaking about math but an intuitive grasp of why it may seem "counterintuitive" for some people. After all, audio is mixed probably in software these days almost everywhere, where there's no volts, but the same mixing (with addition) happens.


14 Aug 2017, 17:11 

revolution
I don't see how it helps to explain that tripling the signal voltage gives nine times the power output.


14 Aug 2017, 17:16 

Furs
Because V includes a negative counterpart. Well rocketsoft is confused why "mixing" two audio sources results in something 4x the power output. Intuitiveness (?) is obviously subjective though, so idk how to explain it if others don't get it. (maybe I "picture" it wrong anyway but is enough intuitive for me)
I mean yeah you can say, well because the math says so, but he's confused why that is even correct in the first place. Personally, I think of it as due to it having negative counterpart. When we say 5V (amplitude) it means it oscillates +5/5 over time. So why isn't it 5V? Because it doesn't matter with squaring: x^2 = x^2. 

14 Aug 2017, 17:29 

revolution
rocketsoft wrote: An open question... output = sqrt(1/2) * (input0 + input1) This assumes you are adding voltages together and that the output is also a voltage level. 

14 Aug 2017, 17:58 

rocketsoft
sqrt(1/2)*(4 + 8)=0.707*12=8.484
(18 watts.... close but no sigar) 4 watts + 16 watts=20 watts 4 ohm! P=I*I*R I=sqrt(P/R) U=I*R U=sqrt((U*U)/(R*R))*R? U=sqrt(144/16=9)*4? 12=not correct! (always original U) restart! P1=(4v*4v)/4ohm=4Watt, P2=16watt P=P1+P2 P=I*I*R I=sqrt(P/R) U=I*R ... exact! For 1 Ohm: P1=U1*U1, P2=U2*U2, P=P1+P2,P=I*I I=sqrt(P) U=I P1=4*4=16,P2=8*8,P=16+64=80 U=sqrt(80)=8.9=correct conclusion:simple math for perfect mixing so... Output=sqrt(U1*U1+U2*U2+U3*U3+...) Otherwise very audible Distortion!!! I guess I answered my own question. Note...sqrt and squares are SignSensitive! sssqrt+ssmul... notify intel 2 new signsensitive fpu instruction "PowersMixer(TM)" I call this new mixer! Even better... its now possible to mix 1024 channels instead of max.32 with classic level(voltage) summers square root of sum of squares Contracts with companies are needed... all middle men(n) get 50%/n and I get 50% too... fair is fair! We will be bigger than Dolby(TM)!!!!! I Want 1% of shares from OS producers! 10% of all other companies that use mixers!!! ;) Last edited by rocketsoft on 19 Aug 2017, 21:10; edited 3 times in total 

14 Aug 2017, 20:43 

Furs
That hurts my eyes, looks like obfuscated math.


15 Aug 2017, 14:27 

rocketsoft
correct Revolution 1+1+1=9 watts!
lol and 1+1=4 watts... its not even linear! 

19 Aug 2017, 21:13 

revolution
rocketsoft wrote: correct Revolution 1+1+1=9 watts! (1 + 1 + 1) ^ 2 = 9 

20 Aug 2017, 04:36 

rocketsoft
spice:says linear mixing produces NO distortion.linear mixing is only correct mixing.Case closed!


18 Dec 2017, 23:26 

< Last Thread  Next Thread > 
Forum Rules:

Copyright © 19992020, Tomasz Grysztar. Also on YouTube, Twitter.
Website powered by rwasa.