Csound Csound-dev Csound-tekno Search About

Sr, kr and amplitude

Date2007-10-21 00:24
Fromad80a
SubjectSr, kr and amplitude
I don't understand why changing control rate and sampling rate I've
differences in timbre and amplitude values.
I'm working with an instrument that has white noise filtered by a bank of
resonant filters and a waveguide reverb with a lot of delay lines (with
delay opcode because I don't need to change the delay at kr).
Using kr 441 sr 44100 I obtain a sound I like (this is the strangest
thing!), but with kr equal to sr (44100/44100), I have some resonance peaks,
a slightly different timbre, and greater amplitude values...I don't
understand why! I DON'T use modulating delay lines! And I don't use any
krate variables! Why changing kr influences timbre and amplitude? Same thing
if i try with sr 96000, timbre changes (and it's natural because there are
more high frequencies and a cleaner timbre), but amplitudes are about TWICE
great!
Thanks for your help and sorry for this newbie question.
-- 
View this message in context: http://www.nabble.com/Sr%2C-kr-and-amplitude-tf4664296.html#a13324298
Sent from the Csound - General mailing list archive at Nabble.com.

Date2007-10-21 07:55
From"Chuckk Hubbard"
SubjectRe: Sr, kr and amplitude
AttachmentsNone  None  

Date2007-10-22 09:53
FromMark Van Peteghem
SubjectRe: Sr, kr and amplitude
ad80a wrote:
> I don't understand why changing control rate and sampling rate I've
> differences in timbre and amplitude values.
> I'm working with an instrument that has white noise filtered by a bank of
> resonant filters and a waveguide reverb with a lot of delay lines (with
> delay opcode because I don't need to change the delay at kr).
> Using kr 441 sr 44100 I obtain a sound I like (this is the strangest
> thing!), but with kr equal to sr (44100/44100), I have some resonance peaks,
> a slightly different timbre, and greater amplitude values...I don't
> understand why! I DON'T use modulating delay lines! And I don't use any
> krate variables! Why changing kr influences timbre and amplitude? Same thing
> if i try with sr 96000, timbre changes (and it's natural because there are
> more high frequencies and a cleaner timbre), but amplitudes are about TWICE
> great!
> Thanks for your help and sorry for this newbie question.
>   

I've seen some csound instruments that accumulate a signal, i.e. summing 
subsequent samples, to have a quick and dirty way to amplify low 
frequencies. The problem with this is that the result is dependent on sr 
or kr (whether it is an a-rate or k-rate signal). This is because when 
the rate is higher, the samples are closer so you add up more values per 
second, resulting in a higher amplitude.

The same thing happens when you use the delay1 opcode, to have a quick 
and dirty way to amplify high frequencies (by taking the difference of 
two subsequent samples). In this case when the samples are closer, the 
difference of two subsequent samples is smaller, resulting in lower 
amplitudes.

Are you sure that the instrument doesn't use delay1, and doesn't 
accumulate? There may be other opcodes with the same problem.

If this is the problem, I think it could be corrected by using the value 
kr or sr in the calculation, or by using the integ and diff opcodes.

-- 
  Mark
  _________________________________________
  When you get lemons, you make lemonade.
  When you get hardware, you make software.