TomCh

(My post to Slack on 21 Jan) >Looks like a track sync issue: start at T$00 (at a precise position), seek to T$23, read 256 nibbles and find a sequence...

Thanks Matthew - I'll take another look at this.

Here's the cross-track sync protection check code: Summary: - Seek to T$00 - Read Address Field for S$0A - Seek to T$35 - Read $100 nibbles to $DC00 - Search...

The original .woz from @medasaro had `optimalBitTiming=31` (on Slack he said he was trying multiple bit timings to see if it make a difference), but this is wrong, and messes...

Here's the T$23 data, roughly where the head must be after seeking from T$00: ``` A600: FF FF FF FF FF FF FF FF FF FF FF FF FF FF...

Details on the AppleWin fix: The WOZ 2.1 spec says to use this formula: ``` new_position = current_position * new_track_length / current_track_length ``` Using integer maths, this will round down...

The improvement in 5888a2b gives better fidelity to the "+1 nibble" fix above, as it now operates on the bit offset (instead of nibble offset). So using a "+7 bit"...

4am attached another 'Balance of Power' woz to Slack on 22nd Aug (`applesauce` channel). This has optimalBitTiming=32, and works in both AppleWin (with the above fix) and MAME 0.236. ```...

@medasaro - fixed in this new [AppleWin 1.30.12.0](https://github.com/AppleWin/AppleWin/releases/tag/v1.30.12.0).

Hmm, seems similar to #716, especially my last comment: https://github.com/AppleWin/AppleWin/issues/716#issuecomment-557900181. I made this change at 1.29.7.0: > For first-time installs of AppleWin, soundtype now defaults to WAVE (was NONE). Seems...