So while this isn’t a game, I can still talk about the decisions, processes, and learnings when designing this.

One of my decisions I made when making this that was not covered in the other devlogs was the decision to preserve compatibility as much as possible, and I’m not just referring to the file size and copying over parameters/etc. No, I’m talking about preserving Micasddsa (a co-author)’s patch ordering that at first glance may seem incorrect given the order of General MIDI patches. HOWEVER, based on many rounds of listening, this ordering seems to be intentional. Given this, I elected to preserve his ordering.

Another decision I made was to keep everything below 1GiB so it would work on the most players, and I did manage that by the skin of its teeth, a story I detailed in a Tech Discussion devlog for this. I also had successfully mapped all the non-GM patches I was given to Roland SC-8850 GS equivalents, as well as remain compatible with OpenMPT and libre. 3 calendar years of thought and choices went into making this.

As for my process in making this, I used Polyphone, and I had originally gotten lucky and a lot of the process since was also luck. Having said that, parts of it were tedious, such as looking through the Roland SC-8850 manual searching for patches that would fit the non-GM ones. It ultimately worked out in the long run. Boosting the drumkit was a tedious process where I cross-checked everything I integrated, and I made sure to enter the right values where they were supposed to be. In the end, it all worked out. SoundFont making involves a lot of steps that would be the worst nightmare for anyone with dyslexia, something I have, so a lot of the extreme amounts of checking were to make sure that didn’t muck things up, as well as satisfy my OCD (a condition I ALSO have). A lot of my process involves manual checks that take a while. It was a very long and intensive process over 3 calendar years, but I did pull it off well.

Now for learnings, I would say that an important lesson I learned in this adventure would be that if you want to do something, you should do it right, and do it well, as well as make the best of the source material you are given, even if it seems like an uphill battle. I almost wrote the whole idea off as a loss in November 2020, but by December 2020 I decided to just go for it, and it WORKED! Also, I mustered up the courage to get in contact with my co-author, and it paid off in the long run, because that chat lead to JummBox support rather than just BeepBox. I also would give the lesson that it is important to keep your content compatible, and that sometimes, things one thinks are mistakes may be just be intentional for a reason one isn’t aware of at the moment.