On 11/08/2011 01:01 PM, Kostya Shishkov wrote:
> On Tue, Nov 08, 2011 at 12:33:56PM -0500, Justin Ruggles wrote:
>> On 11/07/2011 03:06 PM, Kostya Shishkov wrote:
>>
>>> On Mon, Nov 07, 2011 at 02:41:02PM -0500, Justin Ruggles wrote:
>>>> the user is not required to set bit_rate
>>>> ---
>>>> libavcodec/sipr.c | 13 +++++++++----
>>>> 1 files changed, 9 insertions(+), 4 deletions(-)
>>>>
>>>> diff --git a/libavcodec/sipr.c b/libavcodec/sipr.c
>>>> index 70227c3..4257496 100644
>>>> --- a/libavcodec/sipr.c
>>>> +++ b/libavcodec/sipr.c
>>>> @@ -480,10 +480,15 @@ static av_cold int sipr_decoder_init(AVCodecContext
>>>> * avctx)
>>>> SiprContext *ctx = avctx->priv_data;
>>>> int i;
>>>>
>>>> - if (avctx->bit_rate > 12200) ctx->mode = MODE_16k;
>>>> - else if (avctx->bit_rate > 7500 ) ctx->mode = MODE_8k5;
>>>> - else if (avctx->bit_rate > 5750 ) ctx->mode = MODE_6k5;
>>>> - else ctx->mode = MODE_5k0;
>>>> + switch (avctx->block_align) {
>>>> + case 20: ctx->mode = MODE_16k; break;
>>>> + case 19: ctx->mode = MODE_8k5; break;
>>>> + case 29: ctx->mode = MODE_6k5; break;
>>>> + case 37: ctx->mode = MODE_5k0; break;
>>>> + default:
>>>> + av_log(avctx, AV_LOG_ERROR, "Invalid block_align: %d\n",
>>>> avctx->block_align);
>>>> + return AVERROR(EINVAL);
>>>> + }
>>>>
>>>> av_log(avctx, AV_LOG_DEBUG, "Mode: %s\n", modes[ctx->mode].mode_name);
>>>>
>>>> --
>>>
>>> might be okay
>>
>>
>> what concerns do you have?
>
> I don't know whether bitrates and block sizes are related.
Actually, using block_align is less hackish in this case because the
mode is directly related to block_align.
from RM demuxer:
ff_sipr_subpk_size[4] = { 29, 19, 37, 20 };
...
st->codec->block_align = ff_sipr_subpk_size[flavor];
Also, from some google searching, this seems to match the blockalign
value from the wave header for sipr-in-wav/asf/avi. Although there's no
way to be sure whether or not all the search results were from
ffmpeg/libav-generated files...
-Justin
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel
