[FFmpeg-devel] [PATCH] doc/encoders: Add libvo-aacenc doc

[Claudio Freire]

On Sat, Jun 8, 2013 at 3:07 PM, Hendrik Leppkes <h.leppkes at gmail.com> wrote:
> On Sat, Jun 8, 2013 at 7:07 PM, Lou Logan <lou at lrcd.com> wrote:
>> On Fri, Jun 7, 2013, at 06:54 PM, Timothy Gu wrote:
>>>
>>> +It is considered to be worse than the
>>> +native experimental FFmpeg AAC encoder.
>>
>> Has this been confirmed by you? I've seen this comparison:
>>
>> http://d.hatena.ne.jp/kamedo2/20120729/1343545890
>>
>> I don't doubt the results, but has anyone here actually performed blind
>> ABX tests on these two encoders and confirmed this notion (not that I
>> have)?
>
>
> You don't need to do ABX testing, i have some samples encoded with
> libvo-aacenc that have very obvious distortions, but sound fine with
> the experimental encoder.
> It may be something about the specific audio in that sample that it
> doesn't do well with, but its just so obvious.
>
> In fact, here it is:
> libvo-aacenc: https://dl.dropbox.com/u/59011278/MC.mp4
> experimental: https://dl.dropbox.com/u/59011278/ffmpeg%20aac.mp4

Do you have the sources?

I'm making changes to the experimental encoder and I'd like to make
sure there are no regressions. I'm already battling some VBR issues
and I wouldn't mind another set of known-hard samples.