Welcome to Doom9's Forum, THE in-place to be for everyone interested in DVD conversion.

Before you start posting please read the forum rules. By posting to this forum you agree to abide by the rules.

 

Go Back   Doom9's Forum > Video Encoding > High Efficiency Video Coding (HEVC)

Reply
 
Thread Tools Search this Thread Display Modes
Old 22nd January 2016, 11:32   #1  |  Link
serhannn
Registered User
 
Join Date: Sep 2015
Posts: 11
HEVC Decoding process

This is probably explained in great detail in the standard but I need to see the "big picture" on the following topic: Lets assume that I have a compressed HEVC bitstream and I do not want to fully re-construct the pixels; or in other words, I don't want to go through the full-decoding but I need only some certain information such as motion vectors (MVs) and DCT coefficients from the compressed bitstream for further processing with this information. To achieve this, entropy decoding must be done in any case (de-CABAC). However, I'm not sure how much computational effort (relative to the full decoding) is needed in order to be able to extract MVs and DCT coefficients from the bitstream after entropy decoding. Regarding that I have two questions:

1. Is that possible at all in HEVC to do such partial decoding?
2. If yes, what is the approximate percentage of computational time spent on obtaining the DCT coefficients compared to the full decoding (total reconstruction of pixels)?

Of course, this can be highly dependent on the respective decoder implementations, I guess, but I just want to have an insight into the subject to see if it is really feasible to do such partial decoding and "compressed domain" processing in HEVC.

Thanks a lot.
serhannn is offline   Reply With Quote
Old 22nd January 2016, 12:21   #2  |  Link
birdie
Artem S. Tashkinov
 
birdie's Avatar
 
Join Date: Dec 2006
Posts: 337
I'm not an expert in this topic, but take a look at ffmpeg sources - perhaps what you need is already there.
birdie is offline   Reply With Quote
Old 23rd January 2016, 06:32   #3  |  Link
foxyshadis
ангел смерти
 
foxyshadis's Avatar
 
Join Date: Nov 2004
Location: Lost
Posts: 9,558
The bitstream parsing is incredibly fast; the CABAC is actually faster than AVC's, and with wavefront, it's perfectly parallelizable up to a few cores. (There are optimized CABAC implementations so you don't have to re-implement that wheel.) Nearly all of the time is spent in qpel expansion and reduction, copying pixels around, and applying sao/deblocking, plus odds and ends like IDCT. I can do parsing in python, easily, almost as fast as it can read from the disk.
foxyshadis is offline   Reply With Quote
Old 25th January 2016, 17:44   #4  |  Link
serhannn
Registered User
 
Join Date: Sep 2015
Posts: 11
Very interesing. Are you aware of any benchmark study which quantitatively measures which part of the decoding process is the most computationally intensive ? If entropy decoding can be done so fast, it shouldn't be a problem to access information such as motion vectors and DCT coefficients very quickly (relative to the full-decoding time).
serhannn is offline   Reply With Quote
Reply

Tags
bitstream, compressed domain, decoder, hevc

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +1. The time now is 10:26.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.