jwt27

I'd like to push these changes somewhere so you can evaluate them, and we can discuss what needs to be changed. I'm all for constructive criticism, that's what PRs are...

I was not familiar with KiwiSDR, so I just looked up a public one. And yes... I see what you mean. There also seems to be no consistent visual style...

I just like a high-resolution, high-contrast display. My server can easily handle it, and adpcm compression keeps the bandwidth down to reasonable levels. I imagine it's most useful for wideband...

Thanks for the pointer, I worked around it by slicing the waterfall data into fixed size chunks and computing min/max in two passes. The slice size is set to 100k...

Also, I think it's better if the maximum zoom level and number of zoom steps were calculated from the configured FFT size (and perhaps the profile bandwidth), instead of hardcoding...

Some feedback on your recent changes: I pulled in the latest commits and rebased my local changes on top of that. I found that I had to revert fa08f1e since...

Yep, that doesn't look so good. I think it's better to leave the old code in for now, there is very little (if any) visual difference but the overhead is...

Oh nice, that's a lot better. I rebased this PR on top of your latest changes now. Quite surprising that it's faster now without the alpha channel. If this were...

The FFT data still goes straight to the canvas, there's no additional buffering involved as far as I'm aware. As I said before I haven't noticed any difference in performance...

> Hi @jwt27 ! > > > Native-format packages may not contain hyphens in their version numbers. I > > changed it to a plus sign. > > I don't...