Function: vdec_send_frame

Applicability

Product	Supported
Atlas 350 Accelerator Card	x
Atlas A3 training product / Atlas A3 inference product	√
Atlas A2 training product / Atlas A2 inference product	√
Atlas training product	√
Atlas inference product	√
Atlas 200I/500 A2 inference product	√

Description

Feeds a frame (the input buffer, as well as the output buffer) to the decoder for decoding. This API is asynchronous.

Prototype

C Prototype

        
             aclError aclvdecSendFrame(aclvdecChannelDesc *channelDesc, acldvppStreamDesc *input, acldvppPicDesc *output, aclvdecFrameConfig *config, void* userData)

Python Function

        
             ret = acl.media.vdec_send_frame(vdec_channel_desc, dvpp_stream_desc, dvpp_pic_desc, vdec_frame_config, user_data)

Parameters

Parameter	Description
vdec_channel_desc	Int, pointer address of the channel description. The value must be the same as that of dvpp_channel_desc specified when the acl.media.dvpp_create_channel API is called to create a stream. Call the acl.media.vdec_set_channel_desc APIs to set attributes of the channel description, including the decoding channel ID, thread, callback function, and video encoding protocol. NOTE: For the Atlas training product , due to software restrictions, the acl.media.vdec_set_channel_desc_ref_frame_num API cannot be called to set the attributes of the channel description. If this API is called, an error is returned.
dvpp_stream_desc	Int, pointer address of the input stream description. You need to allocate the input buffer in advance. Call acl.media.dvpp_create_stream_desc to create an image description. Call the acl.media.dvpp_set_stream_desc APIs to set the attributes (such as the stream format) of video streams. The input stream format is subject to the format set in the channel description. The acl.media.dvpp_set_stream_desc_format call to set the input stream format is invalid. For details about requirements on the resolution and format of the input video stream, see Functions and Restrictions.
dvpp_pic_desc	Int, pointer address of the output image description. You need to allocate the output buffer in advance. When dvpp_pic_desc is an input, the user needs to: Call the acl.media.dvpp_create_pic_desc API to create the image description. Call the acl.media.dvpp_set_pic_desc APIs to set the data buffer, buffer size, width, height, and format of the output image. The output width and height correspond to four parameters: width, height, width_stride, and height_stride. When the width is less than 16, width_stride must be aligned to 32 instead of 16. If none of the four parameters are set, the input width and height are used. If only some of the four parameters are set, the API returns an error. If the output image format is set in this API call, the setting is applied. Otherwise, the setting in the channel description is applied. For details about the requirements on the output image format, see Functions and Restrictions.
vdec_frame_config	Int, pointer address of the decoding configuration. This parameter is reserved and can be set to 0 currently.
user_data	Python object, pointer address of the user-defined data. If you want to obtain the sequence number of the decoded frame, define the user_data parameter. Then, the sequence number can be passed to the VDEC callback function to determine the frame to be processed.

Return Value

Return Value	Description
ret	Int, error code. 0 on success; else, failure.

Restrictions

This API is asynchronous. The API call delivers a task rather than executes a task. After this API is called, call the synchronization API (for example, acl.rt.synchronize_stream) to ensure that the task is complete.
Before sending data, ensure that a channel is created. Otherwise, an error code is returned.
Streams must be sent in the unit of frames. Only one complete frame of streams is sent at a time.
It is not allowed to send an empty stream packet with eos = 0 (stream length is 0 or stream address is empty).
After all streams are sent, an empty stream packet with EOS = 1 can be sent, indicating the end of the current stream file. After the EOS frame is sent, this API call returns the decode result only after all received frames are decoded and the user-defined callback function handling is complete.
Due to stream exception or decoding timeout, the acl.media.vdec_send_frame API may fail to send frames or the EOS frame. You are advised to obtain the return code of this API when writing code. If this API fails to be called, handle the exception.
acl.media.vdec_send_frame encapsulates the acl.rt.launch_callback API to add a callback function to be executed to the stream task queue. When implementing the VDEC function, you do not need to call the acl.rt.launch_callback API separately.

Reference

For details about the API call sequence and examples, see VDEC.

Parent topic: VDEC