It probably would suffer some, but it would still be better than raw. The MSBs should compress well, just as if it were reduced to 16-bit; the frequencies should not change. The LSBs would be pretty erratic though; like trying to compress noise. But you don't know unless you test.
Can wav files store 32b float? I have never worked with 32b audio data, but I have worked with 32b sensor data many times. If so, this may be a trivial thing to test and it should involve very few code changes. --- Paul Winkler <[EMAIL PROTECTED]> wrote: > Could do that... I wonder if the compression ratio would suffer, > though. Just a hunch. > > -PW > > On Wed, Jul 25, 2007 at 01:37:18PM -0700, Gregory Alan Hildstrom wrote: > > The JasPer library documentation states that a single image component can > > be from 1 to 16 bits > per > > sample, so not directly. 24-bit color images are actually treated as 3 > > 8-bit components: like > one > > red component, one green component, and one blue component. > > > > For lossless compression, we could easily place the most significant 16 > > bits in the first > image > > component and place the least significant 16 bits in a second image > > component of each frame. > It > > should also be possible to code something up that does the same sort of > > 16-bit splitting into > 2 > > flac files. -Greg > > > > --- Paul Winkler <[EMAIL PROTECTED]> wrote: > > > > > Pretty cool. Does JPEG 2000 handle float data? > > > > > > I'd love to be able to write an archive script for Ardour sessions > > > that compresses the audio data losslessly, but FLAC won't do it. > > > > > > > > > -PW > > > > > > On Wed, Jul 25, 2007 at 12:47:16PM -0700, Gregory Alan Hildstrom wrote: > > > > Hello. I just finished writing and documenting a new lossless and lossy > > > > audio codec called > > > > j2kaudio. It uses the wavelet compression in JPEG 2000 to transcode > > > > audio wav data to and > from > > > j2a > > > > files. I have not written any sort of media player plugin yet; more > > > > testing and file > format > > > > refinement are necessary first. > > > > > > > > I tested with 44100Hz, 16-bit, stereo CD-quality wav files, but it also > > > > supports other > sample > > > > rates, precisions, and channel numbers. > > > > > > > > I tested against flac, lpac, shorten, and monkey's audio for lossless > > > > encoding. j2kaudio > beat > > > the > > > > nearest competition by over 10%. > > > > > > > > If anyone is motivated to test some additional songs, provide me with > > > > additional test > data, or > > > > suggest another untested codec, please let me know. > > > > > > > > http://geocities.com/hildstrom/projects/j2kaudio/index.html > > > > > > > > Let me know what you think. Thanks. -Greg > > > > > > > > > > > > Gregory Alan Hildstrom > > > > Secure Systems Engineer - Trusted Computer Solutions > > > > trustedcs.com > > > > [EMAIL PROTECTED] > > > > Software Engineer - Hildstrom Engineering > > > > geocities.com/hildstrom > > > > [EMAIL PROTECTED] > > > > Mobile:(210)413-6082 > > > > Fax:(210)599-0469 > > > > Office:(210)340-3151x117 > > > > _______________________________________________ > > > > Linux-audio-dev mailing list > > > > [email protected] > > > > http://lists.linuxaudio.org/mailman/listinfo.cgi/linux-audio-dev > > > > > > -- > > > > > > Paul Winkler > > > http://www.slinkp.com > > > _______________________________________________ > > > Linux-audio-dev mailing list > > > [email protected] > > > http://lists.linuxaudio.org/mailman/listinfo.cgi/linux-audio-dev > > > > > > > _______________________________________________ > > Linux-audio-dev mailing list > > [email protected] > > http://lists.linuxaudio.org/mailman/listinfo.cgi/linux-audio-dev > > -- > > Paul Winkler > http://www.slinkp.com > _______________________________________________ > Linux-audio-dev mailing list > [email protected] > http://lists.linuxaudio.org/mailman/listinfo.cgi/linux-audio-dev > _______________________________________________ Linux-audio-dev mailing list [email protected] http://lists.linuxaudio.org/mailman/listinfo.cgi/linux-audio-dev
