[opendtv] Re: Integer vs. Floating Point
- From: Ron Economos <k6mpg@xxxxxxxxxxx>
- To: opendtv@xxxxxxxxxxxxx
- Date: Thu, 09 Jun 2005 16:24:02 -0700
H.264 makes extensive use of prediction, since even the intra
coding modes rely upon spatial prediction. As a result, H.264 is
very sensitive to prediction drift. In prior standards, prediction
drift accumulation can occur once per P-frame. In contrast, in
H.264 prediction drift can occur much more frequently. As an
illustration, in an I-frame, 4x4 blocks can be predicted from
their neighbors. At each stage prediction drift can accumulate.
For a CIF image, which has a width of 88 4x4 blocks, prediction
drift can accumulate 88 times in decoding one row of an
I-frame. Thus, it is clear that as a result of the extensive use of
prediction with H.264, the residual coding must be drift-free.
http://www.ee.bgu.ac.il/~marcuse/Project/New%20stuff%20from%20Shlomo/H264_Low-complexity_transform_and_quantization.pdf
Ron
dgrimes@xxxxxxxxxxxxxxxxx wrote:
>"Sidebar: use of the integer cosine transform was not
>done to increase accuracy compared with the DCT, but
>rather to reduce computational complexity." -Bert
>
>The way I remember it, DSPs all process in floating point. If the
>processing is done on an ALU, I can see how working with integers is
>beneficial. Is it as beneficial when processing on a DSP? So what
>decoders are using DSPs and which are only using the microprocessor's ALU?
>It has been a long time since I have developed anything based on processing
>with DSP hardware and software, so perhaps I am getting something confused.
>Anyone care to shed some light on the subject for me?
>
>Dan Grimes
>
>
----------------------------------------------------------------------
You can UNSUBSCRIBE from the OpenDTV list in two ways:
- Using the UNSUBSCRIBE command in your user configuration settings at
FreeLists.org
- By sending a message to: opendtv-request@xxxxxxxxxxxxx with the word
unsubscribe in the subject line.
Other related posts:
