FFMPEG: adding support for ffmpeg planar formats also to writeffmpeg.c.

1 files changed, 82 insertions, 42 deletions

diff --git a/source/blender/blenkernel/intern/writeffmpeg.c b/source/blender/blenkernel/intern/writeffmpeg.c
index a88a9e4954b..f671a3cfc05 100644
--- a/source/blender/blenkernel/intern/writeffmpeg.c
+++ b/source/blender/blenkernel/intern/writeffmpeg.c
@@ -88,10 +88,11 @@ static uint8_t *video_buffer = 0;
 static int video_buffersize = 0;
 
 static uint8_t *audio_input_buffer = 0;
+static uint8_t *audio_deinterleave_buffer = 0;
 static int audio_input_samples = 0;
-static uint8_t *audio_output_buffer = 0;
-static int audio_outbuf_size = 0;
 static double audio_time = 0.0f;
+static bool audio_deinterleave = false;
+static int audio_sample_size = 0;
 
 #ifdef WITH_AUDASPACE
 static AUD_Device *audio_mixdown_device = 0;
@@ -122,37 +123,67 @@ static int write_audio_frame(void)
 {
 	AVCodecContext *c = NULL;
 	AVPacket pkt;
+	AVFrame* frame;
+	int got_output = 0;
 
 	c = audio_stream->codec;
 
 	av_init_packet(&pkt);
 	pkt.size = 0;
+	pkt.data = NULL;
+
+	frame = avcodec_alloc_frame();
+	frame->nb_samples = audio_input_samples;
+	frame->format = c->sample_fmt;
+	frame->channel_layout = c->channel_layout;
 
 	AUD_readDevice(audio_mixdown_device, audio_input_buffer, audio_input_samples);
 	audio_time += (double) audio_input_samples / (double) c->sample_rate;
 
-	pkt.size = avcodec_encode_audio(c, audio_output_buffer, audio_outbuf_size, (short *) audio_input_buffer);
+	if(audio_deinterleave) {
+		int channel, i;
+		uint8_t* temp;
+
+		for(channel = 0; channel < c->channels; channel++) {
+			for(i = 0; i < frame->nb_samples; i++) {
+				memcpy(audio_deinterleave_buffer + (i + channel * frame->nb_samples) * audio_sample_size,
+					   audio_input_buffer + (c->channels * i + channel) * audio_sample_size, audio_sample_size);
+			}
+		}
+
+		temp = audio_deinterleave_buffer;
+		audio_deinterleave_buffer = audio_input_buffer;
+		audio_input_buffer = temp;
+	}
+
+	avcodec_fill_audio_frame(frame, c->channels, c->sample_fmt, audio_input_buffer,
+							 audio_input_samples * c->channels * audio_sample_size, 0);
 
-	if (pkt.size < 0) {
+	if(avcodec_encode_audio2(c, &pkt, frame, &got_output) < 0) {
 		// XXX error("Error writing audio packet");
 		return -1;
 	}
 
-	pkt.data = audio_output_buffer;
+	if(got_output) {
+		if (c->coded_frame && c->coded_frame->pts != AV_NOPTS_VALUE) {
+			pkt.pts = av_rescale_q(c->coded_frame->pts, c->time_base, audio_stream->time_base);
+			PRINT("Audio Frame PTS: %d\n", (int) pkt.pts);
+		}
 
-	if (c->coded_frame && c->coded_frame->pts != AV_NOPTS_VALUE) {
-		pkt.pts = av_rescale_q(c->coded_frame->pts, c->time_base, audio_stream->time_base);
-		PRINT("Audio Frame PTS: %d\n", (int) pkt.pts);
-	}
+		pkt.stream_index = audio_stream->index;
 
-	pkt.stream_index = audio_stream->index;
+		pkt.flags |= AV_PKT_FLAG_KEY;
 
-	pkt.flags |= AV_PKT_FLAG_KEY;
+		if (av_interleaved_write_frame(outfile, &pkt) != 0) {
+			fprintf(stderr, "Error writing audio packet!\n");
+			return -1;
+		}
 
-	if (av_interleaved_write_frame(outfile, &pkt) != 0) {
-		fprintf(stderr, "Error writing audio packet!\n");
-		return -1;
+		av_free_packet(&pkt);
 	}
+
+	avcodec_free_frame(&frame);
+
 	return 0;
 }
 #endif // #ifdef WITH_AUDASPACE
@@ -608,8 +639,6 @@ static AVStream *alloc_video_stream(RenderData *rd, int codec_id, AVFormatContex
 	return st;
 }
 
-/* Prepare an audio stream for the output file */
-
 static AVStream *alloc_audio_stream(RenderData *rd, int codec_id, AVFormatContext *of, char *error, int error_size)
 {
 	AVStream *st;
@@ -659,11 +688,6 @@ static AVStream *alloc_audio_stream(RenderData *rd, int codec_id, AVFormatContex
 		}
 	}
 
-	if (c->sample_fmt == AV_SAMPLE_FMT_FLTP) {
-		BLI_strncpy(error, "Requested audio codec requires planar float sample format, which is not supported yet", error_size);
-		return NULL;
-	}
-
 	if (codec->supported_samplerates) {
 		const int *p = codec->supported_samplerates;
 		int best = 0;
@@ -692,24 +716,23 @@ static AVStream *alloc_audio_stream(RenderData *rd, int codec_id, AVFormatContex
 	st->codec->time_base.num = 1;
 	st->codec->time_base.den = st->codec->sample_rate;
 
-	audio_outbuf_size = FF_MIN_BUFFER_SIZE;
-
-	if ((c->codec_id >= CODEC_ID_PCM_S16LE) && (c->codec_id <= CODEC_ID_PCM_DVD))
-		audio_input_samples = audio_outbuf_size * 8 / c->bits_per_coded_sample / c->channels;
+	if (c->frame_size == 0)
+		// used to be if((c->codec_id >= CODEC_ID_PCM_S16LE) && (c->codec_id <= CODEC_ID_PCM_DVD))
+		// not sure if that is needed anymore, so let's try out if there are any
+		// complaints regarding some ffmpeg versions users might have
+		audio_input_samples = FF_MIN_BUFFER_SIZE * 8 / c->bits_per_coded_sample / c->channels;
 	else {
 		audio_input_samples = c->frame_size;
-		if (c->frame_size * c->channels * sizeof(int16_t) * 4 > audio_outbuf_size)
-			audio_outbuf_size = c->frame_size * c->channels * sizeof(int16_t) * 4;
 	}
 
-	audio_output_buffer = (uint8_t *) av_malloc(audio_outbuf_size);
+	audio_deinterleave = av_sample_fmt_is_planar(c->sample_fmt);
 
-	if (c->sample_fmt == AV_SAMPLE_FMT_FLT) {
-		audio_input_buffer = (uint8_t *) av_malloc(audio_input_samples * c->channels * sizeof(float));
-	}
-	else {
-		audio_input_buffer = (uint8_t *) av_malloc(audio_input_samples * c->channels * sizeof(int16_t));
-	}
+	audio_sample_size = av_get_bytes_per_sample(c->sample_fmt);
+
+	audio_input_buffer = (uint8_t *) av_malloc(audio_input_samples * c->channels * audio_sample_size);
+
+	if(audio_deinterleave)
+		audio_deinterleave_buffer = (uint8_t *) av_malloc(audio_input_samples * c->channels * audio_sample_size);
 
 	audio_time = 0.0f;
 
@@ -1010,12 +1033,28 @@ int BKE_ffmpeg_start(struct Scene *scene, RenderData *rd, int rectx, int recty,
 		AVCodecContext *c = audio_stream->codec;
 		AUD_DeviceSpecs specs;
 		specs.channels = c->channels;
-		if (c->sample_fmt == AV_SAMPLE_FMT_FLT) {
-			specs.format = AUD_FORMAT_FLOAT32;
-		}
-		else {
+
+		switch(av_get_packed_sample_fmt(c->sample_fmt))
+		{
+		case AV_SAMPLE_FMT_U8:
+			specs.format = AUD_FORMAT_U8;
+			break;
+		case AV_SAMPLE_FMT_S16:
 			specs.format = AUD_FORMAT_S16;
+			break;
+		case AV_SAMPLE_FMT_S32:
+			specs.format = AUD_FORMAT_S32;
+			break;
+		case AV_SAMPLE_FMT_FLT:
+			specs.format = AUD_FORMAT_FLOAT32;
+			break;
+		case AV_SAMPLE_FMT_DBL:
+			specs.format = AUD_FORMAT_FLOAT64;
+			break;
+		default:
+			return -31415;
 		}
+
 		specs.rate = rd->ffcodecdata.audio_mixrate;
 		audio_mixdown_device = sound_mixdown(scene, specs, rd->sfra, rd->ffcodecdata.audio_volume);
 #ifdef FFMPEG_CODEC_TIME_BASE
@@ -1138,15 +1177,16 @@ static void end_ffmpeg_impl(int is_autosplit)
 		MEM_freeN(video_buffer);
 		video_buffer = 0;
 	}
-	if (audio_output_buffer) {
-		av_free(audio_output_buffer);
-		audio_output_buffer = 0;
-	}
 	if (audio_input_buffer) {
 		av_free(audio_input_buffer);
 		audio_input_buffer = 0;
 	}
 
+	if (audio_deinterleave_buffer) {
+		av_free(audio_deinterleave_buffer);
+		audio_deinterleave_buffer = 0;
+	}
+
 	if (img_convert_ctx) {
 		sws_freeContext(img_convert_ctx);
 		img_convert_ctx = 0;