spdl.io.VideoFrames¶

class VideoFrames¶

Video frames.

See Packets and Frames for information about the Frames base concept.

Attributes

`clone`	Clone the frames, so that data can be converted to buffer multiple times.
`get_pts`	Get the PTS (Presentation Time Stamp) in timebase unit.
`get_timestamps`	Get the timestamp of frames.
`height`	The height of video.
`num_frames`	The number of video frames.
`num_planes`	The number of planes in the each frame.
`pix_fmt`	The name of the pixel format.
`time_base`	Get the time base of PTS.
`width`	The width of video.

__getitem__¶

Slice frame by key.

Parameters:: key – If the key is int type, a single frame is returned as ImageFrames. If the key is slice type, a new VideoFrames object pointing the corresponding frames are returned.
Returns:: The sliced frame.

clone¶

Clone the frames, so that data can be converted to buffer multiple times.

property num_frames¶: The number of video frames. Same as __len__ method.

property num_planes¶: The number of planes in the each frame.

Note

This corresponds to the number of color components, however it does not always match with the number of color channels when the frame is converted to buffer/array object.

For example, if a video file is YUV format (which is one of the most common formats, and comprised of different plane sizes), and color space conversion is disabled during the decoding, then the resulting frames are converted to buffer as single channel frame where all the Y, U, V components are packed.

SPDL by default converts the color space to RGB, so this is usually not an issue.

property time_base¶

Get the time base of PTS.

The time base is expressed as (Numerator, denominator).

PTS (in seconds) == PTS (in timebase unit) * Numerator / Denominator