Probably the second most important element of a good Tracker, second only to the music engine of course, is a streamlined and effective pattern editor. Many people are drawn to the Tracker approach by the interface’s ability to afford rapid and effective data entry and editing. This is made possible in no small part by the simple and effective pattern editor interface. Tuned over many years on very low end hardware (although not necessarily considered low end at the time) such as the Amiga and DOS PC, this interface is an exercise in streamlined and unobtrusive user interaction.
After some initial experiments in pure React, I got very concerned that the performance of the pattern editor was going to be an issue, so I decided to run some separate exploration to see what could be achieved in isolation, to see if React or other UI elements in use at the time were hampering the potential to achive the low latency, low demand interface I needed.
I created a simple, single page, proof of concept that exercised the basic building blocks:
- Frozen first column and header row.
- Fast navigation.
- Row based scrolling, rather than pixel based.
- Cycled scrolling (when it reaches the bottom, it returns to the top).
- Edit cursor.
- Ideally, all in standard HTML with CSS styling.
This had to be able to render as fast as the browser could go in order to support playback at higher rates, such as 200+ beats per minute (bpm) with 16 or even 32 lines per beat (lpb).
The result was encouraging, click on the screenshot below to load the experiment and test it for yourself. Only basic interaction is included, mouse wheel to scroll, but you can see that even with frantic scrolling, the display keeps up, running smoothly at 60fps even on low end hardware.
Next up, I need to translate this into the React framework that is being developed. I’m still seeing significant slowdown, even using this same technique, so there’s work to do, but at least the theory is proven.