Good.

I think is better to use the data-chunk header, cuz broken sound files are far less common than files with metadata.

And broken sound files can be fixed, (copy audio and paste to a new file).


Mensaje telepatico asistido por maquinas.


From: Pd-list <pd-list-bounces@lists.iem.at> on behalf of IOhannes m zmoelnig <zmoelnig@iem.at>
Sent: Thursday, February 23, 2017 10:09 AM
To: pd-list
Subject: Re: [PD] soundfiler features
 
(taking this back to pd-list)

On 2017-02-22 19:01, Lucas Cordiviola wrote:
> I seems that [soundfiler] correctly ignores metadata.
> [soundfile_info] counts it as length.
> At least for .wav files.

[soundfile_info] assumes that everything after the (1st) "data" chunk is
audio data.
this is obviously not true if you have other chunks (cue, ...) coming
after the audio.
i talked to thomas, and it seems that the behaviour was actually
intentional, meant to correctly handle truncated soundfile files (that
would be shorted than indicated in the data-chunk header)
as a side-effect, it would erroneously handle soundfiles that are longer
than indicated in the data-chunk header.

i fixed this in iemlib's git repository[1].
i'll try to talk thomas into doing a proper release via deken.

fmasdr
IOhannes

[1] https://git.iem.at/pd/iemlib