FFmpeg 参数一览
FFmpeg VP8 Encoding Options
See http://ffmpeg.org/ffmpeg.html#libvpx for additional parameter information.
See http://ffmpeg.org/trac/ffmpeg/wiki/vpxEncodingGuide for a discussion of bitrate parameters.
| Options | Description | Values |
-vb |
target bitrate | |
-deadline |
Encoder speed/quality and cpu use | best, good, realtime |
-cpu-used |
CPU Used | (-16..16) |
-rc_init_occupancy |
number of bits which should be loaded into the rc buffer before decoding starts | |
-bufsize |
ratecontrol buffer size (in bits) | |
-pass |
Pass to execute (1/2) | |
-passlogfile |
First pass statistics file name | |
-vframes |
Stop encoding after n input frames | |
-minrate |
Minimal bitrate allowed in vbr/2pass mode | |
-maxrate |
Maxium bitrate allowed in vbr/2pass mode | |
-bt |
bitrate tolerance in VBR/2pass mode | |
-qmin |
(valid values 0-63, recommended value 0-4) | |
-qmax |
(valid values --min-q to 63, recommended value 50-63) |
|
-qcomp |
float 0 - 1, 0 = CBR, 1 = VBR, (default 0.5) | |
-g |
Maximum keyframe interval (frames) | |
-keyint_min |
Minimum keyframe interval (frames) | |
-auto-alt-ref |
Enable use of alternate reference frames (2-pass only) | (0 disabled, 1 enabled <default 0>) |
-lag-in-frames |
number of frames to look ahead for when encoding, 0 no limit | 0-25 |
-arnr-maxframes |
max frames used in creating alt. ref. | |
-arnr-strength |
Strength of the temporal filter used in altref | 0-6 |
-arnr-type |
altref noise reduction filter type | backward (Backward Blur), forward (Forward Blur), centered (Center Blur) <default> |
-slices |
directs the encoder to split the coefficient encoding across multiple data partitions that can be encoded independently. At the moment this parameter is interpreted as follows (1 = 1 coefficient partition, 2 = 2 partitions, 4 = 4 partitions, 8 = 8 partitions) | 1,2,4,8: recommended 1 for small images, 4 or 8 for HD |
-threads |
number of threads to use for encoding, can't be 0 [auto] with VP8 | (recommended value : number of real cores - 1) |
-skip_threshold |
Temporal resampling threshold (buf %) | (0=disabled to 100) |
-rc_resize_allowed |
Spatial resampling enabled (bool) | (0 disabled, 1 enabled) |
-rc_resize_down |
Spatial resampling down watermark, percentage of target data buffer. | (0-100) |
-rc_resize_up |
Spatial resampling up watermark, percentage of target data buffer. | (--resize-down-100) |
-error-resilient |
Enable error resiliency features | (0 disabled, 1 enabled <default 0>) |
-mb_threshold |
The static threshold imposes a change threshold on blocks below which they will be skipped by the encoder. This can be used to suppress signal noise and enhance the encode speed in situations where there are low levels of real movement. Values of above 1000 are not recommended and any non zero value runs the risk of introducing artifacts caused by regions of the image not being updated. In most scenarios this value should be set to 0. | |
-profile |
Bitstream profile number to use; reduce complexity if > 0 | (0-3: default and recommended value = 0) |
-sharpness |
Anything above 0 weakens the deblocking effect of the loop filter. | (0-7 : default and recommended value = 0) |
-nr |
crude temporal noise filter. | (0-6: default and recommended value = 0) |
-rc_buf_aggressivity |
undershoot target bitrate for each frame by x percent | (valid float 0.1-1) |
-crf |
enables constant quality mode and sets quality | (0-63) lower values are better quality |
