Media (Intel® Video Processing Library, Intel Media SDK)
Access community support with transcoding, decoding, and encoding in applications using media tools like Intel® oneAPI Video Processing Library and Intel® Media SDK
Announcements
The Intel Media SDK project is no longer active. For continued support and access to new features, Intel Media SDK users are encouraged to read the transition guide on upgrading from Intel® Media SDK to Intel® Video Processing Library (VPL), and to move to VPL as soon as possible.
For more information, see the VPL website.
3075 Discussions

IMSDK encoder behaviour when input timestamps are not monotonically incrementing

Anita_C_
Beginner
440 Views

Hi,

We want to use the encoder input surface "Timestamp" field to pass some other information(like pointers) instead of frame timestamps. Please note that the encoder input surface "Timestamp" field value (which are memory addresses) will not be incrementing by frame duration in this case. We want to understand the encoder behaviour in this case. Below are our queries :

1. Is it possible that the encoder will not work as expected or return some error in this case since input surface "Timestamp" field value is not monotonically incrementing ?

2. Does encoder uses timestamp for rate control ? 

3. As per our understanding the IMSDK encoder internally uses input surface "Timestamp" field value to calculate "DecodeTimeStamp". We expect that "DecodeTimeStamp" for encoded bitstream will have random/garbage values in this case. Is that correct ?

Note : We are feeding the input to decoder frame by frame by setting MFX_BITSTREAM_COMPLETE _FRAME.

Regards, Anita

0 Kudos
1 Reply
Surbhi_M_Intel
Employee
440 Views

Hi Anita, 

1. Is it possible that the encoder will not work as expected or return some error in this case since input surface "Timestamp" field value is not monotonically incrementing ?
>> Yes it should not return error for the encoder. Let us know if you see any issue. 

2. Does encoder uses timestamp for rate control ?
>> No
 

3. As per our understanding the IMSDK encoder internally uses input surface "Timestamp" field value to calculate "DecodeTimeStamp". We expect that "DecodeTimeStamp" for encoded bitstream will have random/garbage values in this case. Is that correct ?
>> Might have to check, will get back on this. 

Thanks,
Surbhi

0 Kudos
Reply