[FFmpeg-devel] AAC decoder round 5

[Robert Swain]

2008/8/7 Robert Swain <robert.swain at gmail.com>:
> 2008/8/7 Michael Niedermayer <michaelni at gmx.at>:
>> On Thu, Aug 07, 2008 at 04:04:44PM +0100, Robert Swain wrote:
>>> 2008/8/6 Michael Niedermayer <michaelni at gmx.at>:
>>> > On Wed, Aug 06, 2008 at 12:32:32AM +0100, Robert Swain wrote:
>>> > [...]
>>> >
>>> >> Index: libavcodec/aactab.c
>>> >> ===================================================================
>>> >> --- libavcodec/aactab.c       (revision 14625)
>>> >> +++ libavcodec/aactab.c       (working copy)
>>>
>>> [...]
>>>
>>> >> +DECLARE_ALIGNED(16, float,  ff_aac_kbd_long_1024[1024]);
>>> >> +DECLARE_ALIGNED(16, float,  ff_aac_kbd_short_128[128]);
>>> >> +DECLARE_ALIGNED(16, float, ff_aac_sine_long_1024[1024]);
>>> >> +DECLARE_ALIGNED(16, float, ff_aac_sine_short_128[128]);
>>> >> +
>>> >> +const uint8_t ff_aac_num_swb_1024[] = {
>>> >> +    41, 41, 47, 49, 49, 51, 47, 47, 43, 43, 43, 40
>>> >> +};
>>> >> +
>>> >> +const uint8_t ff_aac_num_swb_128[] = {
>>> >> +    12, 12, 12, 14, 14, 14, 15, 15, 15, 15, 15, 15
>>> >> +};
>>> >> +
>>> >>  const uint32_t ff_aac_scalefactor_code[121] = {
>>> >>      0x3ffe8, 0x3ffe6, 0x3ffe7, 0x3ffe5, 0x7fff5, 0x7fff1, 0x7ffed, 0x7fff6,
>>> >>      0x7ffee, 0x7ffef, 0x7fff0, 0x7fffc, 0x7fffd, 0x7ffff, 0x7fffe, 0x7fff7,
>>> >
>>> >
>>> >> @@ -795,4 +809,90 @@
>>> >>       4064.0312908,  4074.6805676,  4085.3368071,  4096.0000000,
>>> >>  };
>>> >>
>>> >> +/* [ 0, 255] scale factor decoding when using C dsp.float_to_int16
>>> >> + * [60, 315] scale factor decoding when using SIMD dsp.float_to_int16
>>> >> + * [45, 300] intensity stereo position decoding mapped in reverse order i.e. 0->300, 1->299, ..., 254->46, 255->45
>>> >> + */
>>> >
>>> > not doxygen compat, no description of what the table actually contains.
>>>
>>> Comment quoted above improved. Does your statement apply to all of the
>>> above or just the comment for pow2sf_tab[]?
>>
>> i just meant pow2sf_tab but the ff_aac_num_swb_128/1024 could also benefit
>> from a little doxy, the windows are obvious at least to me so i dont care
>> if they get a doxy or not.
>
> I'll put that swb/sfb documentation next to all the *swb* arrays plus
> a comment about what the offsets are and so on. I'll just have a look
> through and see what I think is non-obvious.

How about the attached?

Rob
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 20080808-1413-swb_sfb_comment.diff
Type: text/x-diff
Size: 1131 bytes
Desc: not available
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20080808/ac853831/attachment.diff>