I was transcoding a 128K MP3 speech podcast into 28K Opus and noticed that
part of musical opening sounded a little worse with the recently released
1.2 OpusTools compared to the earlier 1.1 build. I thought maybe you would
like to be made aware of a problem sample. I also thought there is a good
chance over-compressing a musical section of 128K MP3 to 28K Opus isn't
really an important use case, so no worries.
I uploaded a zip containing 7s clips of the original, the Opus 1.1, and
Opus 1.2 encodes here:
http://s000.tinyupload.com/index.php?file_id=06672937462298552381
The command line used: --quiet --bitrate 28 --vbr --ignorelength
Right around 1s is a pronounced guitar pluck that sounds a bit better in
the v1.2 encode. Thank you! Around 3s the narrator starts talking. Behind
the speaking at around 4.5s are four somewhat pronounced guitar notes that
sound more artifcaty in the 1.2 encode. Maybe that's just the VBR
trade-off to make the 1s chord sound better?
Thanks in advance!
-- Steve
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.xiph.org/pipermail/opus/attachments/20170620/70a5dca7/attachment-0001.html>
