Csound Csound-dev Csound-tekno Search About

[Csnd] Fft analysis of a 48000 Hz/24 bits sound file

Date2014-02-05 00:12
FromGpp
Subject[Csnd] Fft analysis of a 48000 Hz/24 bits sound file
Hallo,

I am wondering if it is possible to analyze a 48000 Hz - 24 bits sample and
create a .pvx file preserving its sample rate format.

Thank you.


Gpp



--
View this message in context: http://csound.1045644.n5.nabble.com/Fft-analysis-of-a-48000-Hz-24-bits-sound-file-tp5732264.html
Sent from the Csound - General mailing list archive at Nabble.com.

Date2014-02-05 00:44
FromRichard Dobson
SubjectRe: [Csnd] Fft analysis of a 48000 Hz/24 bits sound file
Yes. the file format does exactly that, preserving information on the 
source sample rate and sample type, as it contains the whole of a 
standard WAVEX format block as part of the extended header. Of course 
the analysis data itself has to be floating-point, but if you give the 
-3 flag to Csound it will create a 24bit output file. Anything less than 
32bit precision is not really adequate for the analysis frame data itself.

Richard Dobson


On 05/02/2014 00:12, Gpp wrote:
> Hallo,
>
> I am wondering if it is possible to analyze a 48000 Hz - 24 bits sample and
> create a .pvx file preserving its sample rate format.
>
> Thank you.
>
>
> Gpp
>
>


Date2014-02-05 11:32
FromGpp
Subject[Csnd] Re: Fft analysis of a 48000 Hz/24 bits sound file
Thank you.

Where do I need to add the flag -3? somewhere here?
pvanal -n8192 -w16 -c2 audiofile.aif audiofile.pvx
Does it work with .pv as well?



--
View this message in context: http://csound.1045644.n5.nabble.com/Fft-analysis-of-a-48000-Hz-24-bits-sound-file-tp5732264p5732286.html
Sent from the Csound - General mailing list archive at Nabble.com.

Date2014-02-05 11:45
FromRichard Dobson
SubjectRe: [Csnd] Re: Fft analysis of a 48000 Hz/24 bits sound file
No, that's a Csound command line flag, only meaningful when writing 
Csound output to a sound file. Internally, all Csound signals and 
processing is floating point (mostly double precision - 64bit - these 
days). If you are thinking of trying to create an analysis file with 
24bit integer values, sorry that is not possible, and would be 
technically inadequate, especially at higher sample rates and larger FFT 
sizes, and given the wild and wonderful processing people typically want 
to apply. Even a plain 24bit soundfile is automatically converted to 
floats (soundin, diskin etc) on loading into Csound. You will not lose 
any precision in the conversion; in the absence of any other processing 
the two-way conversion will be bit-exact. Csound is seriously HiFi!

Richard Dobson


On 05/02/2014 11:32, Gpp wrote:
> Thank you.
>
> Where do I need to add the flag -3? somewhere here?
> pvanal -n8192 -w16 -c2 audiofile.aif audiofile.pvx
> Does it work with .pv as well?
>
>