[Tremor] ov_pcm_tell > ov_pcm_total by design?

[Kwasi Mensah]

Clarified the behavior I'm seeing in ov_read.

On Wed, Dec 15, 2010 at 11:55 AM, Kwasi Mensah <kwasi.mensah at gmail.com>wrote:

> It's a attached. It's one second of A220 mono at 44100hz.
>
> Also attached is the source for an example that will show the problem. I
> had to run it on my iPhone so its a .mm. Also, you'll have to add A220.ogg
> to the top level of you're .app and then recompile it to make the code
> signature process happy.
>
> For the last ov_read which pushes ov_pcm_tell of ov_pcm_totoal ov_read
> calls _fetch_and_process_packet which will call vorbis_synthesis_blockin
> after making sure oldsamples == 0.  vorbis_synthesis_blockin updates
>  v->pcm_returned to 0 and v->pcm_current to 1024.
>
> Now,  vorbis_synthesis_pcmout will always return 1024.
>
> Thanks,
> Kwasi
>
>
> On Wed, Dec 15, 2010 at 5:53 AM, <xiphmont at xiph.org> wrote:
>
>> On Wed, Dec 15, 2010 at 3:20 AM, Kwasi Mensah <kwasi.mensah at gmail.com>
>> wrote:
>> > Hey guys,
>> >
>> > I'm using Tremor cl 17618. vorbis_synthesis_blockin is doing some type
>> of
>> > overlap adding that I think is pushing ov_pcm_tell to be greater than
>> > ov_pcm_total when reading the last bit of audio data from an ogg file.
>> This
>> > causes a later call I make to ov_pcm_seek fail. Is this by design?
>> Should I
>> > just fix my call to ov_pcm_seek to take this into account?
>>
>> That is not by design; tell and total should agree (and do on the test
>> cases I have here).  Could you perhaps send me the file causing
>> trouble?
>>
>> Monty
>>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.xiph.org/pipermail/tremor/attachments/20101215/b6f60583/attachment-0001.htm