On Thu, Sep 4, 2014 at 5:22 PM, Joshua Judson Rosen <roz...@geekspace.com> wrote: > Hrm. Is that actually an appropriate use of floats?
I think so. With modern hardware, even efficient time-wise. But chews up RAM a little faster. > I imagine > people working with JPEGs as source material basically can't care > about whatever precision is being lost; is the lost precision > `down in the noise' for RAW, too? yeah, JPEGs are hardly precise or accurate or anything else. One reason to access the RAW ... RAWs *are* precise, like a 12, 14, or 16 bit per channel GIF/PNG/TIFF; the only compression if any is RLE. If i save a RAW out again as TIFF16, I shouldn't have lost any original bits in the diversion into Floats with 'merely' 24 bits of fraction, but autoscaling avoids trouble in exposure compensations etc. (File-on-disk hash difference should be dominated by metadata changing when re-writing.) ( A little noise down in the noise may actual help avoid Mach banding during processing, so 'useless' extra precision of 24 bits may actually be helpful too.) -- Bill Ricker bill.n1...@gmail.com https://www.linkedin.com/in/n1vux _______________________________________________ gnhlug-discuss mailing list gnhlug-discuss@mail.gnhlug.org http://mail.gnhlug.org/mailman/listinfo/gnhlug-discuss/