Search options (Page 16 of 21)

Thought to make a note on the row transition noise. There is a way to minimize it. The idea is to avoid row transitions as is. Rather than storing song as a set of rows, with constant skipping of empty fields, store delays between any actual changes that is taking place. This should keep the sound loop tight, and longer gaps when actual changes happens will be masked by the fact that these changes actually change/restart a note, or trigger a drum, i.e. introduce a major change in the sound anyway.

I.e., rather than:

C-1 ...
... ...
... ...
... ...
... C-2
... ...
C-1 E-2

have:

C-1 ... > run sound loop for four rows without parsing anything
... C-2 > run sound loop for two rows
C-1 E-2

v0.26 is up. Much faster Z80 assembler (in AngelScript), still not as fast as any native would be, but work gets more comfortable. Many minor fixes as well.

v0.25 is out of WIP. Now powered with SDL2, which supposedly solves some issues. For one, it now switches to the fullscreen and back blazing fast under Windows, unlike the old version. It also brings separate files for fonts (which is now 1-bit) and color schemes, and a pile of the font and color scheme files to look pretty and provide flexibility regarding screen sizes. It seemignly compiles fine under Ubuntu without editing the code, just by 'make' command (given SDL2 installed, not sure about AngelScrpit). I accidentally Pytha engine, too.

No, the idea of the changes in file structure and makefile was to compile and link AngelScript as part of the program, directly through object files, not as a lib.

Migrating to SDL2 is actually a good idea, I had a few SDL2 projects done, so will put it to plans. It just happened that the program does not really needed things that SDL2 has to offer, so I kept it plain SDL so far.

The display issue is really weird, as I don't have it neither under Windows nor under Linux. The construct is for normal view, it aligns the scrolling for blocks properly. My only guess is that it is some unitialized variable or getting out of bounds, althogh considering proper results in two very different environments, I'd expect the issue to be more exposed. Also, row < 24 means it is somehwere in the scrolling logic, as it is half of the pattern view height, the position to keep cursor line once scrolling starts. I'll think more about it.

The first = is a bug, I normally don't use such unsafe constructions (apart from a piece related to dirent code).

Yes, I think it going to work if the line will be disabled in the chopped view. I'm trying to compile it under Linux at the moment, to look into directory navigation issue. Not sure if I'll be able to fix that part, but will put an updated version soon anyway, with the display fix at least.

1tracker got an update. No new engines this time, sorry. Instead, I tried another improvement to the 'single pattern' concept, that is called 'chopped view'. When it is enabled through the new alternative menu (hold Ctrl to see it), only current block is displayed starting from the top of the screen, like a separate pattern. I hope that obstructing visibility of other blocks should help to focus attention on current block, and make it more comfortable to work in case of short blocks (less scrolling). If you split your song into blocks of 16 or 32 lines, it will work similar to the traditional pattern/order list view in a way.

Linux users, please let me know what changes this version needs in order to compile it. Hopefully less changes will be needed this time. I also just noticed that dirent.h had a wrong include path, docs say it should be there, so file dialogs should work properly.

Now that's some black magic is going on here.

You would need to spend most of the CPU time on both machines to run the MIDI protocol, as Spectrum does not have hardware serial I/O.

Minor update to the 1bitstudio. All plugins now support Program Change event to switch between presets, also support preset name edits. phasesynth got extra oscillator mode, pin pulse in addition to the square/pulse wave.

Thinking on the 1-bit SuperSaw on ZX and messing up with assembly code a bit, I tried this thing. The idea behind it is the early electronic orgrans that used frequency dividers to form sound registers. So only one real counter here. Can't remember if we ever used something like this?

    ld hl,0
    ld de,400
    
loop
    
    add hl,de
    ld a,h
    cp #80    ;duty 1
    sbc a,a
    out (#fe),a

    push hl
    
    add hl,hl
    ld a,h
    cp #80   ;duty 2
    sbc a,a
    out (#fe),a
    
    add hl,hl
    ld a,h
    cp #80   ;duty 3
    sbc a,a
    out (#fe),a
    
    pop hl
    jp loop

This is one channel with 16-bit adder, the usual idea with duty cycle control like in Tritone. The trick is to generate a number of octave doubled derivatives (phase synced, though) by multiplying the adder by 2 (adding the adder to itself) a few times, two in this example. It adds octave doubles to the sound. Interesting thing is that the three duty cycles kind of have a limited control on the strength of the octave harmonics.

Another thing is that just applying a few duty checks to single unchanged adder also affects the timbre in a way:

    add hl,de
    ld a,h
    cp #80
    sbc a,a
    out (#fe),a

    ld a,h
    cp #08
    sbc a,a
    out (#fe),a

Well, basically that's what Mister Beep always doing in his 8-channel songs for bass, to increase its strength - a few voices of slightly detuned copies of the same bassline. Just pushed way further, like all 8-16 channels set up to do this thing automatically. So a basic version can be done on the Octode (narrow pin wave) or Squeaker (square wave) base, but for more interesting results more channels (CPU power) is needed.