[FFmpeg-devel] [PATCH] AAC: type puns for 16 bit floating point rounding

[Måns Rullgård]

"Guillaume POIRIER" <poirierg at gmail.com> writes:

> Hello,
>
> On Wed, Dec 10, 2008 at 12:04 AM, Alex Converse <alex.converse at gmail.com> wrote:
>> On Tue, Dec 9, 2008 at 5:58 PM, Robert Swain <robert.swain at gmail.com> wrote:
>>
>>> 2008/12/9 Guillaume POIRIER <poirierg at gmail.com>:
>>> > Hello,
>>> >
>>> > On Tue, Dec 9, 2008 at 7:27 PM, Robert Swain <robert.swain at gmail.com>
>>> wrote:
>>> >> 2008/12/9 Guillaume POIRIER <poirierg at gmail.com>:
>>> >> Grab:
>>> >>
>>> http://standards.iso.org/ittf/PubliclyAvailableStandards/ISO_IEC_14496-4_2004_Conformance_Testing/audio_conformance/mpeg4audio-conformance/compressedMp4/am01_48.mp4
>>> >>
>>> >> or some other am* file from there, though not am02* as there's a known
>>> >> bug with those at the moment.
>>> >
>>> > These samples are too short, so I'm afraid that I'm measuring ffmpeg's
>>> > startup time rather than AAC decoding (it takes about .2 seconds to
>>> > decode the file, from launch of FFmpeg to exit).
>>> >
>>> > Would it work with any iTunes-encoded AAC file? If not, maybe you
>>> > should direct me to a bigger sample from Mplayer's samples
>>> > collections...
>>>
>>> It needs to be an AAC Main profile file. iTunes files are LC AAC as
>>> are most other AAC streams in the wild. If I find a longer sample,
>>> I'll let you know. In the mean time, thanks for trying. :)
>>>
>>
>> A longer sample can be made with mp4box as follows.
>>
>> mp4box -raw 1 am00_48.mp4
>> cat am00_48_track1.aac am00_48_track1.aac (...) > out.aac
>> mp4box -add out.aac out.mp4
>
> Thanks for the tip; On my PPC970, decoding is a bit faster (around 1%).

Can you upload the file somewhere?  I'd rather not install mp4box just
for this.

-- 
M?ns Rullg?rd
mans at mansr.com