video¶

this module provides the video codec interface. Encoders, Decoders and their support code.

class pyrana.video.Decoder(input_codec, params=None)¶

Decodes video Packets into video Frames.

decode(packets)¶: Decode data from a logical stream of packets, and returns when the first next frame is available. The input stream can be - a materialized sequence of packets (list, tuple...) - a generator (e.g. Demuxer.stream()).

decode_packet(packet)¶: Generator method. Decode a single packet (as in returned by a Demuxer) and extracts all the frames encoded into it. An encoded packet can legally contain more than one frame, altough this is not so common. This method deals with the [one packet -> many frames] scenario. The internal underlying decoder does its own buffer, so you can freely dispose the packet(s) fed into this method after it exited. raises ProcessingError if decoding fails; raises NeedFeedError if decoding partially succeeds, but more data is needed to reconstruct a full frame.

extra_data¶: bytearray-like, read-write

flush()¶: emits all frames that can be recostructed by the data buffered into the Decoder, and empties such buffers. Call it last, do not intermix with decode*() calls. caution: more than one frame can be buffered. Raises NeedFeedError if all the internal buffers are empty.

classmethod from_cdata(ctx)¶: builds a pyrana Video Decoder from (around) a (cffi-wrapped) libav* (video)decoder object. The libav object must be already initialized and ready to go. WARNING: raw access. Use with care.

media_type¶: the codec media type.

open(ffh=None)¶: opens the codec into the codec context.

params¶: the codec parameters.

ready¶: is the codec readu to go?

setup()¶: Dispach the given parameters to the internal (FFmpeg) data structures.

static wire(dec)¶: wire up the Decoder. See codec.wire_decoder

class pyrana.video.Encoder(output_codec, params)¶

Encode video Frames into Packets.

encode(frame)¶: Encode a logical frame in one or possibly more)packets, and return an iterable which will yield all the packets produced.

extra_data¶: bytearray-like, read-write

flush()¶: emits all packets which may have been buffered by the Encoder and empties such buffers. Call it last, do not intermix with encode*() calls. caution: more than one encoded frame (thus many packets) can be buffered. Raises NeedFeedError if all the internal buffers are empty.

classmethod from_cdata(ctx, params, codec=None)¶: builds a pyrana video Encoder from (around) a (cffi-wrapped) liabv* (audio) context. WARNING: raw access. Use with care.

media_type¶: the codec media type.

open(ffh=None)¶: opens the codec into the codec context.

params¶: the codec parameters.

ready¶: is the codec readu to go?

setup()¶: Dispach the given parameters to the internal (FFmpeg) data structures.

static wire(enc)¶: wire up the Encoder. See codec.wire_encoder

class pyrana.video.Frame(width, height, pixfmt)¶

A Video frame.

asr¶: The sample aspect ratio of the frame.

cdata¶: Direct access to the internal C AVFrame object.

coded_pict_number¶: Picture number in bitstream order.

display_pict_number¶: Picture number in display order.

classmethod from_cdata(ppframe)¶: builds a pyrana generic Base Frame from (around) a (cffi-wrapped) libav* AVFrame object. The libav object must be already initialized and ready to go. WARNING: raw access. Use with care.

image(pixfmt=None)¶: Returns a new Image object which provides access to the Picture (thus the pixel as bytes()) data.

is_interlaced¶: Is the content of the picture interlaced?

is_key¶: Is this a key frame?

pict_type¶: Picture type of the frame, see AVPictureType.

pts¶: The Presentation TimeStamp of this Frame.

top_field_first¶: If is_interlaced(), is top field displayed first?

class pyrana.video.Image¶

Represents the Picture data inside a Frame.

blob()¶: returns the bytes() dump of the object.

convert(pixfmt)¶: convert the Image data in a new PixelFormat. returns a brand new, independent Image.

classmethod from_cdata(ppframe, sws=None, parent=None)¶: builds a pyrana Image from a (cffi-wrapped) libav* Frame object. The Picture data itself will still be hold in the Frame object. The libav object must be already initialized and ready to go. WARNING: raw access. Use with care.

height¶: Frame height. Expected to be always equal to the stream height.

is_shared¶: Is the underlying C-Frame shared with the parent py-Frame?

pixel_format¶: Frame pixel format. Expected to be always equal to the stream pixel format.

plane(idx)¶: Read-only byte access to a single plane of the Image.

planes¶: Return the number of planes in the Picture data. e.g. RGB: 1; YUV420: 3

width¶: Frame width. Expected to be always equal to the stream width.

class pyrana.video.SWSMode¶: SWS operational flags. This wasn’t a proper enum, rather a collection of #defines, and that’s the reason why it is defined here.

pyrana.video.fill_yuv420p(frame, i)¶: fill a video frame with a test pattern.