iie `c2t -d` causes `ERRERR`, most `c2t-96h` files cause `CHKSUM ERROR`

[fredsa]

I tested different formats using my Apple IIe (enhanced). Results below.

Currently, c2t and c2t-96h produce very different looking files. From a quick inspection, it seems that c2t-96h uses square waves and consistent amplitude, which after reading #27 makes me think that's why my IIe is uphappy with these files. The c2t formats work much better, with the exception of -d. (As an aside, I also have issues with some of the disk images on asciiexpress.net, which upon inspection also use square waves and consistent amplitude.)

I'd be happy to troubleshoot different formats on my end. I have a II plus and a IIe, and access to an oscilloscope.

Tests were run on Windows 10, using the current executables (https://github.com/datajerk/c2t/tree/master/bin at 76c7a64c388c850db8bc9705be6ef52f04b2abd8). Based on a few repeat tests, successes & failures are consistent with the same file, so I think I can rule out audio (cable) flakiness.

Various c2t speeds work, except for -d:

# OK
bin/c2t.exe -b tests/images/moon.patrol,801 moon.patrol-c2t-b.aif

# OK
bin/c2t.exe -b8 tests/images/moon.patrol,801 moon.patrol-c2t-b8.aif

# ERRERR
bin/c2t.exe -bd tests/images/moon.patrol,801 moon.patrol-c2t-bd.aif

# OK
bin/c2t.exe -bf tests/images/moon.patrol,801 moon.patrol-c2t-bf.aif

However, c2t-96h.exe doesn't (for me) appear to be in a good shape. Perhaps it would be best to (for now) remove the statement in the README that suggests it's the better version :).

# OK
bin/c2t-96h.exe -b tests/images/moon.patrol,801 moon.patrol-96h-b.aif

# CHKSUM ERROR
bin/c2t-96h.exe -b8 tests/images/moon.patrol,801 moon.patrol-96hb8.aif

# CHKSUM ERROR
bin/c2t-96h.exe -bd tests/images/moon.patrol,801 moon.patrol-96h-bd.aif

# CHKSUM ERROR
bin/c2t-96h.exe -bf tests/images/moon.patrol,801 moon.patrol-96h-bf.aif