[FFmpeg-devel] [PATCH] add MTAF demuxer and decoder

[Paul B Mahol]

On 5/16/16, Paul B Mahol <onemda at gmail.com> wrote:
> On 5/16/16, James Almer <jamrial at gmail.com> wrote:
>> On 5/15/2016 6:33 PM, Paul B Mahol wrote:
>>> diff --git a/libavcodec/Makefile b/libavcodec/Makefile
>>> index 3f0ffd1..3a9dd95 100644
>>> --- a/libavcodec/Makefile
>>> +++ b/libavcodec/Makefile
>>> @@ -707,6 +707,7 @@ OBJS-$(CONFIG_ADPCM_IMA_WAV_ENCODER)      +=
>>> adpcmenc.o adpcm_data.o
>>>  OBJS-$(CONFIG_ADPCM_IMA_WS_DECODER)       += adpcm.o adpcm_data.o
>>>  OBJS-$(CONFIG_ADPCM_MS_DECODER)           += adpcm.o adpcm_data.o
>>>  OBJS-$(CONFIG_ADPCM_MS_ENCODER)           += adpcmenc.o adpcm_data.o
>>> +OBJS-$(CONFIG_ADPCM_MTAF_DECODER)         += adpcm.o adpcm_data.o
>>>  OBJS-$(CONFIG_ADPCM_PSX_DECODER)          += adpcm.o adpcm_data.o
>>>  OBJS-$(CONFIG_ADPCM_SBPRO_2_DECODER)      += adpcm.o adpcm_data.o
>>>  OBJS-$(CONFIG_ADPCM_SBPRO_3_DECODER)      += adpcm.o adpcm_data.o
>>> diff --git a/libavcodec/adpcm.c b/libavcodec/adpcm.c
>>> index 0b6b92e..e624b85 100644
>>> --- a/libavcodec/adpcm.c
>>> +++ b/libavcodec/adpcm.c
>>> @@ -107,6 +107,10 @@ static av_cold int adpcm_decode_init(AVCodecContext
>>> *
>>> avctx)
>>>      case AV_CODEC_ID_ADPCM_EA_XAS:
>>>          max_channels = 6;
>>>          break;
>>> +    case AV_CODEC_ID_ADPCM_MTAF:
>>> +        min_channels = 2;
>>> +        max_channels = 8;
>>> +        break;
>>>      case AV_CODEC_ID_ADPCM_PSX:
>>>          max_channels = 8;
>>>          break;
>>> @@ -159,6 +163,7 @@ static av_cold int adpcm_decode_init(AVCodecContext
>>> *
>>> avctx)
>>>          case AV_CODEC_ID_ADPCM_AFC:
>>>          case AV_CODEC_ID_ADPCM_DTK:
>>>          case AV_CODEC_ID_ADPCM_PSX:
>>> +        case AV_CODEC_ID_ADPCM_MTAF:
>>>              avctx->sample_fmt = AV_SAMPLE_FMT_S16P;
>>>              break;
>>>          case AV_CODEC_ID_ADPCM_IMA_WS:
>>> @@ -342,6 +347,15 @@ static inline int16_t
>>> adpcm_yamaha_expand_nibble(ADPCMChannelStatus *c, uint8_t
>>>      return c->predictor;
>>>  }
>>>
>>> +static inline int16_t adpcm_mtaf_expand_nibble(ADPCMChannelStatus *c,
>>> uint8_t nibble)
>>> +{
>>
>> You should use local variables here, like in other expand_nibbble()
>> functions.
>
> I don't see point of that.
>
>>
>>> +    c->predictor += ff_adpcm_mtaf_stepsize[c->step][nibble];
>>> +    c->predictor = av_clip_int16(c->predictor);
>>> +    c->step += ff_adpcm_index_table[nibble];
>>> +    c->step = av_clip(c->step, 0, 31);
>>
>> av_clip_uintp2(step, 5);
>
> Done.
>

Will apply soon if nobody objects.