Re: [libdv-dev] observations from dv_decode_full_audio()

[Dan D. <da...@de...>]

On Tue, Dec 2, 2008 at 4:58 PM, salsaman <sal...@gm...> wrote:
> Running dv_decode_full_audio() on a 1GB file, I get quite a few
> errors/warnings like:
>
> # audio block/sample failure for 0 blocks, 32 samples of 1280
> asf 00:00:00.00 2000-00-00 00:00:00 77 07 05 12 1/48
> asf 00:00:00.00 2000-00-00 00:00:00 77 07 08 12 1/48
> asf 00:00:00.00 2000-00-00 00:00:00 77 17 02 12 1/48
> asf 00:00:00.00 2000-00-00 00:00:00 77 17 05 12 1/48
> asf 00:00:00.00 2000-00-00 00:00:00 77 17 08 12 1/48
>
>
> Is this anything to be concerned about ? Is it possible to turn this

It depends upon your requirements and the frequency of these warnings.
If you need perfect audio, then it is a problem, but for most people
some of this is quite tolerable.

> debugging output off ?

you can direct to /dev/null using dv_set_error_log(). I do that in
Kino and dvgrab.

> Also, running through valgrind, at the end of audio processing I see this:
>
> ==8132== Use of uninitialised value of size 8
> ==8132==    at 0x1D126C26: (within /usr/lib/libdv.so.4.0.3)
> ==8132==    by 0x5: ???
> ==8132==    by 0x7FEFD36BF: ???
> ==8132==    by 0x1D11A8D8: dv_parse_ac_coeffs (in /usr/lib/libdv.so.4.0.3)
> ==8132==    by 0x1D1181DF: dv_decode_full_frame (in /usr/lib/libdv.so.4.0.3)

I use valgrind, but not heavily, and I always feel like such a noob
when doing so. I cross reference many messages with code and try to
figure out if there is really a problem or some silly complaining, or
I am just stupid? In this case, I have no comment. I try not to look
at its code much, but I still value it for its metadata and audio
functions.

-- 
+-DRD-+