Frank

I agree with @blazoncek - I see lots of styling, but still no clue what your problem is about. > I can't show my full codebase, but these should be...

> Maybe there should be an option to configure what actually is outputted to DMX, instead of just all registered strips? @Stijn-Jacobs i've checked the DMX serial output code, and...

@DedeHai and @netmindz one thing you might want to try - currently chained hub75 panels are not compatible with the wled 2D layout settings. The reason is that HUB75 is...

> . I need to check that but if it is true hardware scaling, i think it is "kind of" hardware scaling. the hub75 driver will PWM-modulate the "enable" signal,...

> This means that drivers no longer need to maintain pixel data and any "double" buffering can be removed as it is done in WS2812FX class. Hi @blazoncek, hope you...

> One thing that is not clear to me from the MM or Netmindz's code is if it is allowed to have Hub matrix with dimensions of 16, 48, or...

> Conclusion: > Dirty tracking helps. The fastest is with dirty + draw loop. > So basically exactly like now, just without CRGB. > CRGB uses less RAM, and it’s...

> _all_ pixels are updated during `show()`. @blazoncek I see your point. Actually, the dirty bits logic can _possibly_ still help, at least when the framebuffer on lowest level does...

> @softhack007 you seem most knowledgeable about Hub75 implementation in WLED. > Looking at the code I can see that if `_ledBuffer` allocation fails the `BusHub75Matrix` class can work without...

> I'll dump this here. @blazoncek settings page looks good :-) all the options I can imagine are there, only two panel sizes still missing: 16 (high or wide), 128...