fpsdk::common::video::VideoFrameDecoder Class Reference

Helper for decoding video frames. More...

#include <video.hpp>

Collaboration diagram for fpsdk::common::video::VideoFrameDecoder:

Public Member Functions
	VideoFrameDecoder (const VideoDecoderParams &params)
	Constructor.
virtual	~VideoFrameDecoder ()
	Destructor.
virtual std::optional< ImageData >	DecodeFrame (const uint8_t *data, const std::size_t size)=0
	Decode one frame.
std::optional< ImageData >	DecodeFrame (const std::vector< uint8_t > &data)
	Decode video data.
virtual bool	IsOkay () const =0
	Check if decoder is in error state.

Protected Attributes
VideoDecoderParams	params_
	Params.

Detailed Description

Helper for decoding video frames.

Note

This is only available if compiled with FFmpeg, see Dependencies.

Definition at line 197 of file video.hpp.

Constructor & Destructor Documentation

VideoFrameDecoder()

fpsdk::common::video::VideoFrameDecoder::VideoFrameDecoder ( const VideoDecoderParams & params )

explicit

Constructor.

Don't use, use the CreateVideoFrameDecoder() factory function.

Parameters

[in]

params

The parameters

Member Function Documentation

DecodeFrame() [1/2]

virtual std::optional< ImageData > fpsdk::common::video::VideoFrameDecoder::DecodeFrame ( const uint8_t * data, const std::size_t size )

pure virtual

Decode one frame.

• This works for data that is available frame by frame. That is, it does not work for regular video streams where the data is chunked arbitrarily. Ideally, start by supplying an I frame incl. all the necessary NAL units (NALUs) to decode a first frame. After that, the following P frames should decode fine.
• For example, for HEVC (H.265) a full set of data for a I frame should contain:
  • NALU type 32 (VPS, Video Parameter Set)
  • NALU type 33 (SPS, Sequence Parameter Set)
  • NALU type 34 (PPS, Picture Parameter Set)
  • NALU type 19 (IDR, Instantaneous Decoding Refresh, and RADL, Random Access Decodable Leading) (possibly other types 16-23 (IRAP, Intra Random Access Point) NALUs might work)
• The video encoder should be configured accordingly (to repeat the VPS/SPS/PPS for each I-frame, and possibly to produce all/only I frames.
• Separate instances of this should be used to process different video streams

Parameters

[in]	data	Data for one (not more, not less) video frame
[in]	size	Size of data

Returns

the decoded, converted and scaled image, nullptr otherwise (not enough data, bad data, ...)

DecodeFrame() [2/2]

std::optional< ImageData > fpsdk::common::video::VideoFrameDecoder::DecodeFrame

(

const std::vector< uint8_t > &

data

)

Decode video data.

Parameters

[in]

data

Data for one (recommended) or more video frame(s)

Returns

the decoded, converted and scaled image, nullptr otherwise (not enough data, bad data, ...)

IsOkay()

virtual bool fpsdk::common::video::VideoFrameDecoder::IsOkay ( ) const

pure virtual

Check if decoder is in error state.

This is useful to check when DecodeFrame() didn't return a frame. If the decoder is in error state, it cannot be used anymore and must be discarded.

Returns

true if the decoder is okay, false if it is in error state

Member Data Documentation

params_

VideoDecoderParams fpsdk::common::video::VideoFrameDecoder::params_

protected

Params.

Definition at line 257 of file video.hpp.

---

The documentation for this class was generated from the following file:

• fpsdk_common/include/fpsdk_common/video.hpp