[flac-dev] Regain play analysis patches

[Earl Chew]

Erik,

It turns out bc(1) is too accurate, and a little slow, for this purpose.

I've switched to using awk(1) which uses floating point.

Do you feel I need to test for the presence of awk(1) ?

It is specified as one of the standard commands in the LSB :


http://refspecs.linuxfoundation.org/LSB_1.0.0/gLSB/command.html


Earl


    awk -- '
    BEGIN {
            samplerate = 8000;

            tone = 1000;
            duration = 1;
            bitspersample = 24;

            samplemidpoint = lshift(1, (bitspersample-1));
            samplerange = samplemidpoint - 1;

            pi = 4 * atan2(1,1);

            for (ix = 0; ix < duration * samplerate; ++ix) {
                    sample = sin(2 * pi * tone * ix / samplerate);
                    sample *= samplerange;
                    sample += samplemidpoint;
                    sample = int(sample);
                    for (bx = 0; bx < bitspersample/8; ++bx) {
                            byte[bx] = sample % 256;
                            sample /= 256;
                    }
                    while (bx--) {
                            printf("%c", byte[bx]);
                    }
            }

    }'  /dev/null




----- Original Message -----
From: Erik de Castro Lopo <mle+la at mega-nerd.com>
To: flac-dev at xiph.org
Cc: Earl Chew <earl_chew at yahoo.com>
Sent: Friday, February 17, 2012 3:30:00 PM
Subject: Re: [flac-dev] Regain play analysis patches

Earl Chew wrote:

> I'm a little reluctant to introduce another compiled program when there are
> so many other options that will work well enough out of the box.
> 
> Here are two ideas:
> 
> 1. Use bc(1) to compute the raw samples
> 2. Use perl(1) to compute the raw samples
> 
> To generate raw unsigned samples using bc(1) for example:
> 
> samplerate = 1000;
> duration = 2;
> bitspersample = 24;
> 
> samplerange = 2 ^ (bitspersample-1) - 1;
> samplemidpoint = 2 ^ (bitspersample-1);
> 
> pi = 4 * a(1);
> 
> scale = 18;
> obase = 16;
> 
> for (ix = 0; ix < duration * samplerate; ++ix) {
>   sample = samplemidpoint + samplerange * s(2 * pi * ix / samplerate);
>   s = scale;
>   scale = 0;
>   sample /= 1;
>   sample;
>   scale = s;
> }
> 
> 
> 
> Are you ok with bc(1) ?

Yes, bc looks quite nice.

As Richard Ash mentioned, the test should detect the presence of bc
and it its not available, print a nice big warning, and exit the 
script with an error code of 0. If bc is available, run the tests
and any failure should result in the script exiting with a non-zero
error code.

Cheers,
Erik
-- 
----------------------------------------------------------------------
Erik de Castro Lopo
http://www.mega-nerd.com/