[Ffmpeg-cvslog] r5515 - in trunk/tests: ffmpeg.regression.ref regression.sh rotozoom.regression.ref

[Justin Ruggles]

michael wrote:
> Author: michael
> Date: Sat Jun 24 13:40:21 2006
> New Revision: 5515
> 
> Modified:
>    trunk/tests/ffmpeg.regression.ref
>    trunk/tests/regression.sh
>    trunk/tests/rotozoom.regression.ref
> 
> Log:
> flac regression test
> 
> 
> Modified: trunk/tests/ffmpeg.regression.ref
> ==============================================================================
> --- trunk/tests/ffmpeg.regression.ref	(original)
> +++ trunk/tests/ffmpeg.regression.ref	Sat Jun 24 13:40:21 2006
> @@ -176,3 +176,7 @@
>  264236 ./data/a-adpcm_yam.wav
>  e92cec8c07913ffb91ad2b11f79cdc00 *./data/out.wav
>  stddev:18312.68 PSNR:11.06 bytes:1056768
> +7bcc7daf968fc489c9b4b7aca5ae380d *./data/a-flac.flac
> +799336 ./data/a-flac.flac
> +0116cdcefd0aeae3ab8e5140c19c725d *./data/out.wav
> +stddev: 51.59 PSNR:62.07 bytes:1032192
> 
> Modified: trunk/tests/regression.sh
> ==============================================================================
> --- trunk/tests/regression.sh	(original)
> +++ trunk/tests/regression.sh	Sat Jun 24 13:40:21 2006
> @@ -68,6 +68,7 @@
>      do_g726=y
>      do_adpcm_ima_wav=y
>      do_adpcm_ms=y
> +    do_flac=y
>      do_rc=y
>      do_mpeg4adv=y
>      do_mpeg4thread=y
> @@ -595,6 +596,16 @@
>  fi
>  
>  ###################################
> +if [ -n "$do_flac" ] ; then
> +# encoding
> +file=${outfile}flac.flac
> +do_ffmpeg $file -y -ab 128 -ac 2 -ar 44100 -f s16le -i $pcm_src -acodec flac $file
> +
> +# decoding
> +do_ffmpeg $pcm_dst -y -i $file -f wav $pcm_dst
> +fi
> +
> +###################################
>  # libav testing
>  ###################################
>  
> 
> Modified: trunk/tests/rotozoom.regression.ref
> ==============================================================================
> --- trunk/tests/rotozoom.regression.ref	(original)
> +++ trunk/tests/rotozoom.regression.ref	Sat Jun 24 13:40:21 2006
> @@ -176,3 +176,7 @@
>  264236 ./data/a-adpcm_yam.wav
>  e92cec8c07913ffb91ad2b11f79cdc00 *./data/out.wav
>  stddev:18312.68 PSNR:11.06 bytes:1056768
> +7bcc7daf968fc489c9b4b7aca5ae380d *./data/a-flac.flac
> +799336 ./data/a-flac.flac
> +0116cdcefd0aeae3ab8e5140c19c725d *./data/out.wav
> +stddev: 51.59 PSNR:62.07 bytes:1032192
> _______________________________________________
> ffmpeg-cvslog mailing list
> ffmpeg-cvslog at mplayerhq.hu
> http://lists.mplayerhq.hu/mailman/listinfo/ffmpeg-cvslog
> 

Hmm.  This is fine for now, but I really don't think I'll be able to add
useful LPC without using floating-point, which would not pass regression
tests on all machines.  I tried a fixed-point version once, but it was
not nearly accurate enough to give good compression results.  I'll
submit my floating-point version and if anyone has suggestions for how
to make it fixed-point and still give good results, I'll be all ears.
Maybe 32-bit fixed using 64-bit variables?

-Justin