Streams

Stream collections

class av.container.streams.StreamContainer

Bases: object

A tuple-like container of Stream.

# There are a few ways to pulling out streams.
first = container.streams[0]
video = container.streams.video[0]
audio = container.streams.get(audio=(0, 1))

Dynamic Slicing

StreamContainer.get(streams=None, video=None, audio=None, subtitles=None, data=None)

Get a selection of Stream as a list.

Positional arguments may be int (which is an index into the streams), or list or tuple of those:

# Get the first channel.
streams.get(0)

# Get the first two audio channels.
streams.get(audio=(0, 1))

Keyword arguments (or dicts as positional arguments) as interpreted as (stream_type, index_value_or_set) pairs:

# Get the first video channel.
streams.get(video=0)
# or
streams.get({'video': 0})

Stream objects are passed through untouched.

If nothing is selected, then all streams are returned.

StreamContainer.best(type: Literal['video', 'audio', 'subtitle', 'attachment', 'data'], /, related: Stream | None)

Finds the “best” stream in the file. Wraps av_find_best_stream. Example:

stream = container.streams.best("video")

Parameters:

• type – The type of stream to find

• related – A related stream to use as a reference (optional)

Returns:

The best stream of the specified type

Return type:

Stream | None

Typed Collections

These attributes are preferred for readability if you don’t need the dynamic capabilities of get():

StreamContainer.video

A tuple of VideoStream.

StreamContainer.audio

A tuple of AudioStream.

StreamContainer.subtitles

A tuple of SubtitleStream.

StreamContainer.attachments

A tuple of AttachmentStream.

StreamContainer.data

A tuple of DataStream.

StreamContainer.other

A tuple of Stream

Streams

class av.stream.Stream

Bases: object

A single stream of audio, video or subtitles within a Container.

>>> fh = av.open(video_path)
>>> stream = fh.streams.video[0]
>>> stream
<av.VideoStream #0 h264, yuv420p 1280x720 at 0x...>

This encapsulates a CodecContext, located at Stream.codec_context. Attribute access is passed through to that context when attributes are missing on the stream itself. E.g. stream.options will be the options on the context.

Basics

Stream.type

The type of the stream.

Type:

Literal[“audio”, “video”, “subtitle”, “data”, “attachment”]

Stream.codec_context

Stream.id

The format-specific ID of this stream.

Type:

int

Stream.index

The index of this stream in its Container.

Type:

int

Timing

See also

Time for a discussion of time in general.

Stream.time_base

The unit of time (in fractional seconds) in which timestamps are expressed.

Type:

fractions.Fraction | None

Stream.start_time

The presentation timestamp in time_base units of the first frame in this stream.

Type:

int | None

Stream.duration

The duration of this stream in time_base units.

Type:

int | None

Stream.frames

The number of frames this stream contains.

Returns 0 if it is not known.

Type:

int

Others

Stream.profile

The profile of this stream.

Type:

str

Stream.language

The language of the stream.

Type:

str | None

Stream.discard

Controls which packets of this stream are discarded by the demuxer.

Set this to e.g. Discard.all on streams you don’t need so that Container.demux() and Container.seek() skip them, avoiding the cost of synchronizing streams you never read.

Type:

Discard