Please disregard the first of the previous git patches as it was incorrect.
I'll stop spamming the list with patches, so please instead pull from the
tt-errata branch of
git://repo.or.cz/dirac-spec-errata.git
to see my proposed changes.
FWIW, many but not all issues are corrected already in SMPTE 2042.
I have more changes waiting, but there are a couple of issues that I'm
unsure of:
- Most importantly, 12.3.4 block_vector() uses RMODE as a map or an array,
but
block_ref_mode() and block_global clearly uses it as a scalar
value (which is either INTRA, REF1ONLY, REF2ONLY, or REF1AND2.
Reading the dirac-research codec, it appears that it really means:
"if ((state[BLOCK_DATA][y][x][RMODE] & ref) != 0):"
See also 12.3.6.5. Not sure what "...[RMODE][ref] == True" is supposed
to mean.
I'm really not sure how this should be corrected. Probably being
explicit about [RMODE] being
an integer.
- 9.3 parse_sequence() as well as 14 decode_sequence() start by clearing the
state (state = {}), but that
wipes out state[NEXT_BIT] and state[CURRENT_BYTE], so either those have to
be preserved or
the stream must be initialized after this point.
- 10.3 Page 32, Table 10.1: HD 1080I-60 claims 60 Field/s. Shouldn't it be
59.94 Field/s?
- 12.3.6.4 majority(), however trivial, isn't defined.
- 15.8.2 clip() isn't defined (trivial though)
Thanks
Tommy
------------------------------------------------------------------------------
Download Intel® Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Schrodinger-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/schrodinger-devel
