Reverted some of Timothy's changes to the range decoding that were causing decode random decode problems at the end of the stream (but only once in a while).

git-svn-id: http://svn.xiph.org/trunk/ghost@14413 0101bb08-14d6-0310-b084-bc0e0c8e3800

encode function. Also, trying to properly flush the bit packer
(still some problems left).

tterribe authored 17 years ago and Jean-Marc Valin committed 17 years ago

Update ec_dec_bits64() to correspond to the new ec_enc_bits64(). Not sure why this wasn't committed before.

git-svn-id: http://svn.xiph.org/trunk/ghost@14409 0101bb08-14d6-0310-b084-bc0e0c8e3800

use it now.

tterribe authored 17 years ago and Jean-Marc Valin committed 17 years ago

Add ec_dec_tell() to report the same information as ec_enc_tell() decode side. Remove ec_enc_tell() and replace it with ec_enc_tellf(): just request 0 fractional bits if you want an integer value.

git-svn-id: http://svn.xiph.org/trunk/ghost@14394 0101bb08-14d6-0310-b084-bc0e0c8e3800

tterribe authored 17 years ago and Jean-Marc Valin committed 17 years ago

git-svn-id: http://svn.xiph.org/trunk/ghost@14393 0101bb08-14d6-0310-b084-bc0e0c8e3800

tterribe authored 17 years ago and Jean-Marc Valin committed 17 years ago

Remove the special case for 0 in EC_ILOG, as we never pass it 0 and this
 generates slightly better code.

Update ec_enc_bits64 to split the encoded values along word boundaries...
 this should generate slightly better code, as well as fix a subtle bug (the
 proper bits were not being masked out of the low part).
However, this will render previous streams that used this function undecodable
 (to my knowledge, no one is actually using it yet).

git-svn-id: http://svn.xiph.org/trunk/ghost@14391 0101bb08-14d6-0310-b084-bc0e0c8e3800

time)

latency as well.

to multiple MDCTs per frame. Not complete yet.

Timothy B. Terriberry authored 17 years ago and Jean-Marc Valin committed 17 years ago

http://people.xiph.org/~tterribe/notes/cwrs.html
Removed dead code.

pulse spreading (and to work at all).

frame.