To elaborate on the glitch in EMU2WAV, consider the 8-bit source sample (before it gets expanded back to 12 bits) ranging from -128 to +127; 0-127 pass undistorted, but -1 to -127 get translate to -2 to -128. That makes half the waveform distorted and a small discontinuity around zero, but the really bad partthe one that gets hit converting ZD707is that -128 is translated to -1, making a huge spike for any sample that taps or clips at the most negative value. Mixed Choir does that a lot, apparently. --- In emax@yahoogroups.com, "codehead" <codehead_1@...> wrote: > > I haven't had my emax out in years. It was a plain one that I upgraded to SE when that came out, and factory upgraded to HD when that came out. I sure was surprised at how incredibly loud the hard drive was when it returned, and I recall that being a factor thereafter in using it. When I last fired it up a few years ago, I had HD errors on a couple of banks...if I resurrect it, it will certainly be with that HD replaced with flash or something... > > Anyway, I came across the online sample sample floppy images, and found this group. To start, I wanted to convert some of those old samples to wav. I found EMU2WAV here... > > 1. If there a source of the disk image format? EMU2WAV just converts the entire disk to audio, data sections and all (so, huge burst of garbage, then the samples in order with short glitches between them). I took a quick look in a file editor, and I could figure out what I need to given time, but it's probably not worth it to me. I don't know if the modern source to emxp (is it even available?) has that info. > > 2. EMU2WAV has a serious flaw in the code. For the 8-bit samples it reads and decodes, the negative half of the waveforms are off by 1 (causing a discontinuity and clipping). I'm surprised that for most disks, the distortion isn't noticeable in casual listening, but on ZD707 (the mixed choir disk), the sounds is incredibly distorted, with continuous artifact throughout the sound (it sounds like the scratchiest record you ever hearda continuous stream of clicks and pops). After fixing the code error, ZD707 translated into its old familiar self. I can upload the fixed and modernized source code, but I might improve it more first if I can find details on the disk format. >
Message
Re: emax disk header format
2012-03-25 by codehead
Attachments
- No local attachments were found for this message.