/// <summary>
/// Convert to audio frame to <paramref name="dstFotmat"/>
/// <para><see cref="DepthType"/> to <see cref="AVSampleFormat"/> mapping table.
/// if <see cref="Mat.NumberOfChannels"/> > 1, use packet format, otherwise planar</para>
/// <list type="table" >
/// <item>
/// <term><see cref="DepthType.Cv8U"/></term>
/// <description1><see cref="AVSampleFormat.AV_SAMPLE_FMT_U8"/>/<see cref="AVSampleFormat.AV_SAMPLE_FMT_U8P"/></description1>
/// </item>
/// <item>
/// <term><see cref="DepthType.Cv16S"/></term>
/// <description1><see cref="AVSampleFormat.AV_SAMPLE_FMT_S16"/>/<see cref="AVSampleFormat.AV_SAMPLE_FMT_S16P"/></description1>
/// </item>
/// <item>
/// <term><see cref="DepthType.Cv32S"/></term>
/// <description1><see cref="AVSampleFormat.AV_SAMPLE_FMT_S32"/>/<see cref="AVSampleFormat.AV_SAMPLE_FMT_S32P"/></description1>
/// </item>
/// <item>
/// <term><see cref="DepthType.Cv32F"/></term>
/// <description1><see cref="AVSampleFormat.AV_SAMPLE_FMT_FLT"/>/<see cref="AVSampleFormat.AV_SAMPLE_FMT_FLTP"/></description1>
/// </item>
/// <item>
/// <term><see cref="DepthType.Cv64F"/></term>
/// <description1><see cref="AVSampleFormat.AV_SAMPLE_FMT_DBL"/>/<see cref="AVSampleFormat.AV_SAMPLE_FMT_DBLP"/></description1>
/// </item>
/// <item>
/// <term><see cref="DepthType.Cv64F"/></term>
/// <description1><see cref="AVSampleFormat.AV_SAMPLE_FMT_S64"/>/<see cref="AVSampleFormat.AV_SAMPLE_FMT_S64P"/></description1>
/// </item>
/// <item>NOTE: Emgucv not supported int64, mapping Cv64F to int64,
/// so set Mat with int64 if <paramref name="dstFotmat"/> is <see cref="AVSampleFormat.AV_SAMPLE_FMT_S64"/> or <see cref="AVSampleFormat.AV_SAMPLE_FMT_S64P"/>
/// </item>
/// </list>
/// </summary>
/// <param name="mat"></param>
/// <param name="dstFotmat">Default is auto format by <see cref="Mat.Depth"/> and <see cref="Mat.NumberOfChannels"/> use mapping table</param>
/// <param name="dstSampleRate">Mat not have sample rate, set value here or later</param>
/// <returns></returns>
public static AudioFrame ToAudioFrame(this Mat mat, AVSampleFormat dstFotmat = AVSampleFormat.AV_SAMPLE_FMT_NONE, int dstSampleRate = 0)
{
AVSampleFormat srcformat;
switch (mat.Depth)
{
case DepthType.Default:
case DepthType.Cv8U:
case DepthType.Cv8S:
srcformat = mat.NumberOfChannels > 1 ? AVSampleFormat.AV_SAMPLE_FMT_U8 : AVSampleFormat.AV_SAMPLE_FMT_U8P;
break;
case DepthType.Cv16U:
case DepthType.Cv16S:
srcformat = mat.NumberOfChannels > 1 ? AVSampleFormat.AV_SAMPLE_FMT_S16 : AVSampleFormat.AV_SAMPLE_FMT_S16P;
break;
case DepthType.Cv32S:
srcformat = mat.NumberOfChannels > 1 ? AVSampleFormat.AV_SAMPLE_FMT_S32 : AVSampleFormat.AV_SAMPLE_FMT_S32P;
break;
case DepthType.Cv32F:
srcformat = mat.NumberOfChannels > 1 ? AVSampleFormat.AV_SAMPLE_FMT_FLT : AVSampleFormat.AV_SAMPLE_FMT_FLTP;
break;
case DepthType.Cv64F:
srcformat = mat.NumberOfChannels > 1 ? AVSampleFormat.AV_SAMPLE_FMT_DBL : AVSampleFormat.AV_SAMPLE_FMT_DBLP;
break;
default:
throw new FFmpegException(FFmpegException.NotSupportFormat);
}
if (dstFotmat != AVSampleFormat.AV_SAMPLE_FMT_NONE && dstFotmat != srcformat)
{
// converter must need set sample rate
using (SampleConverter converter = new SampleConverter(dstFotmat, mat.NumberOfChannels > 1 ? mat.NumberOfChannels : mat.Height, mat.Width, Math.Min(1, dstSampleRate)))
{
AudioFrame frame = converter.ConvertFrame(MatToAudioFrame(mat, srcformat, Math.Min(1, dstSampleRate)), out int a, out int b);
unsafe
{
// set real sample rate after convert
((AVFrame *)frame)->sample_rate = dstSampleRate;
}
}
}
return(MatToAudioFrame(mat, srcformat, dstSampleRate));
}
/// <summary>
/// convert current frame to packet frame.
/// if current frame is planer return new packet frame
/// else return current frame.
/// </summary>
/// <returns></returns>
public AudioFrame ToPacket()
{
unsafe
{
if (ffmpeg.av_sample_fmt_is_planar((AVSampleFormat)pFrame->format) <= 0)
{
return(this);
}
AVSampleFormat outFormat = (AVSampleFormat)pFrame->format;
if (outFormat == AVSampleFormat.AV_SAMPLE_FMT_NB || outFormat == AVSampleFormat.AV_SAMPLE_FMT_NONE)
{
throw new FFmpegException(FFmpegException.NotSupportFormat);
}
switch ((AVSampleFormat)pFrame->format)
{
case AVSampleFormat.AV_SAMPLE_FMT_U8P:
outFormat = AVSampleFormat.AV_SAMPLE_FMT_U8;
break;
case AVSampleFormat.AV_SAMPLE_FMT_S16P:
outFormat = AVSampleFormat.AV_SAMPLE_FMT_S16;
break;
case AVSampleFormat.AV_SAMPLE_FMT_S32P:
outFormat = AVSampleFormat.AV_SAMPLE_FMT_S32;
break;
case AVSampleFormat.AV_SAMPLE_FMT_FLTP:
outFormat = AVSampleFormat.AV_SAMPLE_FMT_FLT;
break;
case AVSampleFormat.AV_SAMPLE_FMT_DBLP:
outFormat = AVSampleFormat.AV_SAMPLE_FMT_DBL;
break;
case AVSampleFormat.AV_SAMPLE_FMT_S64P:
outFormat = AVSampleFormat.AV_SAMPLE_FMT_S64;
break;
}
AudioFrame outFrame = new AudioFrame(pFrame->channels, pFrame->nb_samples, outFormat, pFrame->sample_rate);
outFrame.pFrame->channel_layout = pFrame->channel_layout;
using (SampleConverter converter = new SampleConverter(outFrame))
{
return(converter.ConvertFrame(this, out int _, out int __));
}
}
}