Sharp MZ

Let me PM you

I'm interested in.

Yep, those are external QDDs for MZ-2000/2200 with an MZ-1E18 interface.

HL is already saved in stack and that may be the only thing which needs to be preserved when calling USR - that is, no need to save/restore in the code called by USR.

Yes, I believe the version with keyboard buffer is better in the way that it doesn't "corrupt" the MZ header comment. The extra length and the extra loading loading time is too small to be a concern.

Not mine, it is just a "random" picture to illustrate what should be the /CSYNC but I totally agree with you, it is a lovely scope output. I have no idea about the model. And I have only a Hantek 6022BL for analogous outputs and mostly use a LE1010 for logical outputs.

Using decompression through keyboard buffer? There is an option to choose between decompressor inside the MZF header or through the keybaord buffer. For 3 minutes version, it was because each block were compressed. Compressing one huge block appears to give more compression ratio.

En WAV - sur cassette - ça prend combien de temps en chargement ? j'avais fait un changement pour que ça prenne moins de 3 minutes au lieu des 12 minutes mais je n'avais pas fait l'effort de rassembler tout ça en un seul MZF.

Now a space invader with, say, 112x100 pixels :p.

Yeah, without looking at the code, I'm pretty sure you use 112 pixels per line, but instead displaying a rectangle, you shift the X character start position X at each pixel line Y. ;P