Fix no resize transcode of 10bits items with hwaccel (#793)

transcoder/src/hwaccel.go

@@ -12,7 +12,7 @@ func DetectHardwareAccel() HwAccelT {
 	}
 	log.Printf("Using hardware acceleration: %s", name)
 
-	// superfast or ultrafast would produce a file extremly big so we prever to ignore them. Fast is available on all hw accel modes
+	// superfast or ultrafast would produce a file extremely big so we prefer to ignore them. Fast is available on all hwaccel modes
 	// so we use that by default.
 	// vaapi does not have any presets so this flag is unused for vaapi hwaccel.
 	preset := GetEnvOr("GOCODER_PRESET", "fast")
@@ -25,9 +25,9 @@ func DetectHardwareAccel() HwAccelT {
 			EncodeFlags: []string{
 				"-c:v", "libx264",
 				"-preset", preset,
-				// sc_threshold is a scene detection mechanisum used to create a keyframe when the scene changes
+				// sc_threshold is a scene detection mechanism used to create a keyframe when the scene changes
 				// this is on by default and inserts keyframes where we don't want to (it also breaks force_key_frames)
-				// we disable it to prevents whole scenes from behing removed due to the -f segment failing to find the corresonding keyframe
+				// we disable it to prevents whole scenes from being removed due to the -f segment failing to find the corresponding keyframe
 				"-sc_threshold", "0",
 				// force 8bits output (by default it keeps the same as the source but 10bits is not playable on some devices)
 				"-pix_fmt", "yuv420p",
@@ -53,16 +53,23 @@ func DetectHardwareAccel() HwAccelT {
 			// see https://trac.ffmpeg.org/wiki/Hardware/VAAPI#Encoding for more info
 			// we also need to force the format to be nv12 since 10bits is not supported via hwaccel.
 			// this filter is equivalent to this pseudocode:
-			// if (vaapi) {
+			//
-			//   hwupload, passthrough, keep vaapi as is
+			// if (format != nv12 && format != vaapi) {
-			//   convert whatever to nv12 on GPU
+			//   convert_to_nv12() (in cpu space, the format is invalid only if hw-decode failed)
+			//   hwupload (send to gpu space)
+			//   scale_to(width, height, output_format="nv12") (output format doesn't change since we are already in nv12)
 			// } else {
-			//   convert whatever to nv12 on CPU
+			//   _format_has_nothing_to_do (do nothing since the format is already valid)
-			//   hwupload to vaapi(nv12)
+			//   _hwupload (the format is already a gpu space format so this is also a no-op)
-			//   convert whatever to nv12 on GPU // scale_vaapi doesn't support passthrough option, so it has to make a copy
+			//   scale_to(width, height, output_format="nv12") (if we are already in nv12 it's only a scale, otherwise it converts to it)
+			//     note: nv12 is 8bit format while vaapi can be 10bits or anything else
 			// }
 			// See https://www.reddit.com/r/ffmpeg/comments/1bqn60w/hardware_accelerated_decoding_without_hwdownload/ for more info
 			ScaleFilter: "format=nv12|vaapi,hwupload,scale_vaapi=%d:%d:format=nv12",
+			// Exactly the same as above but without the scaling.
+			// Yes scale_vaapi is used to convert the format (the `format` filter will pick a format for
+			// the next filter so it doesn't work if there is no next filter).
+			NoResizeFilter: "format=nv12|vaapi,hwupload,scale_vaapi=format=nv12",
 		}
 	case "qsv", "intel":
 		return HwAccelT{
@@ -77,7 +84,8 @@ func DetectHardwareAccel() HwAccelT {
 				"-preset", preset,
 			},
 			// see note on ScaleFilter of the vaapi HwAccel, this is the same filter but adapted to qsv
-			ScaleFilter: "format=nv12|qsv,hwupload,scale_qsv=%d:%d:format=nv12",
+			ScaleFilter:    "format=nv12|qsv,hwupload,scale_qsv=%d:%d:format=nv12",
+			NoResizeFilter: "format=nv12|qsv,hwupload,scale_qsv=format=nv12",
 		}
 	case "nvidia":
 		return HwAccelT{
@@ -91,11 +99,12 @@ func DetectHardwareAccel() HwAccelT {
 			EncodeFlags: []string{
 				"-c:v", "h264_nvenc",
 				"-preset", preset,
-				// the exivalent of -sc_threshold on nvidia.
+				// the equivalent of -sc_threshold on nvidia.
 				"-no-scenecut", "1",
 			},
 			// see note on ScaleFilter of the vaapi HwAccel, this is the same filter but adapted to cuda
-			ScaleFilter: "format=nv12|cuda,hwupload,scale_cuda=%d:%d:format=nv12",
+			ScaleFilter:    "format=nv12|cuda,hwupload,scale_cuda=%d:%d:format=nv12",
+			NoResizeFilter: "format=nv12|cuda,hwupload,scale_cuda=format=nv12",
 		}
 	default:
 		log.Printf("No hardware accelerator named: %s", name)

transcoder/src/settings.go

@@ -22,10 +22,11 @@ type SettingsT struct {
 }
 
 type HwAccelT struct {
-	Name        string
+	Name           string
-	DecodeFlags []string
+	DecodeFlags    []string
-	EncodeFlags []string
+	EncodeFlags    []string
-	ScaleFilter string
+	NoResizeFilter string
+	ScaleFilter    string
 }
 
 var Settings = SettingsT{

transcoder/src/videostream.go

@@ -81,6 +81,8 @@ func (vs *VideoStream) getTranscodeArgs(segments string) []string {
 			"-vf", fmt.Sprintf(Settings.HwAccel.ScaleFilter, width, vs.quality.Height()),
 		)
 	} else {
+		args = append(args, "-vf", Settings.HwAccel.NoResizeFilter)
+
 		// NoResize doesn't have bitrate info, fallback to a know quality higher or equal.
 		for _, q := range Qualities {
 			if q.Height() >= vs.video.Height {
@@ -90,7 +92,7 @@ func (vs *VideoStream) getTranscodeArgs(segments string) []string {
 		}
 	}
 	args = append(args,
-		// Even less sure but bufsize are 5x the avergae bitrate since the average bitrate is only
+		// Even less sure but bufsize are 5x the average bitrate since the average bitrate is only
 		// useful for hls segments.
 		"-bufsize", fmt.Sprint(quality.MaxBitrate()*5),
 		"-b:v", fmt.Sprint(quality.AverageBitrate()),
@@ -100,7 +102,7 @@ func (vs *VideoStream) getTranscodeArgs(segments string) []string {
 		// without this option, some hardware encoders uses others i-frames and the -f segment can't cut at them.
 		"-forced-idr", "1",
 		"-force_key_frames", segments,
-		// make ffmpeg globaly less buggy
+		// make ffmpeg globally less buggy
 		"-strict", "-2",
 	)
 	return args