วีดีโอ

Dusk Forest is one of the many dungeons that has a second half with different music, in this case Deep Dusk Forest. Aside from coming up with a plainer and more "expository" ending for the main theme of Dusk Forest (0:21 and 2:33), I added Deep Dusk Forest's first theme in counterpoint with it (0:57 and 2:20), and also added the final section of Deep Dusk Forest (1:54) to make the development section last longer. Ultimately there were only 4 bars of Deep Dusk Forest that I didn't use in some way.

Thanks! Taking two similar tracks and merging them together is always a fun compositional puzzle, even if I don't do it often.

lol no

'Will be?' You know something we don't know? But seriously tho, dude. People had the same reaction to Stravinsky's own compositons. How is this not playing something on a different instrument?

Famitracker for the music (used 0cc-Famitracker for this, now I use the latest branch, Dn-Famitracker) Beforehand, Musescore 3 for writing a full-length reduction of the ballet score to study it, and also for exporting some of the DPCM samples I used in Famitracker. Obs to screen-record both Famitracker and NSFPlay. Blender's VSE for stitching segments together, mirroring the NSFPlay output, and generally fix the timing. Blender's Nodes Editor with videos as textures for the outline effect and brightness changes. On my very old computer the final rendering took like 4 consecutive days with aggressive throttling (using Battle Encoder Shirasé) to avoid overstressing my CPU.

I have technically shared some modules on my neocities site (first of my channel links for now), in the "player" page. Not in "raw" form, the "player" page converts them from a text format back to modules and makes them downloadable. The biggest insight I think I can share is "articulation". Before writing a phrase, I need to imagine how I would play/sing/conduct it, and try to match it in FamiTracker with tools and techniques I know how to use. And by "match it" I mean continuously tweaking it until it feels right, by paying attention to legato & portato (how lax or quick the transition between notes needs to be) vs separato & staccato (how early and quickly I can release the note), volume swells on melodic peaks, starting and stopping vibrato on longer notes, tweaking tempo and feeling it out until a ritenuto feels right; even adding slight delays to desynchronize and humanize some attacks a bit. Usually I don't have to do that for literally every phrase: most of the time I can copy and change a similar phrase that I got right the first time, or mask a less refined low octave under more articulated upper sections. I find that thinking about articulation intensely helps me automatically deal with the other 2 important aspects of arrangement: On one hand, thinking about articulation implies an analysis, either in terms of "what type of instrument can I assign to this phrase" (for piano pieces, where I take more liberties) or "which parts of the score are a highlight, and which parts I can group together in one channel or squeeze in the background of other channels" (for orchestral scores). On the other hand, writing and continuously tweaking the same passage (trying vibrato/volume changes in different spots, trying out different pulse widths or different channels) makes me aware of how the NES sounds interact with each other. For example I learned that: smaller pulse widths are quieter but give more "punch" and roughness to low notes; arps are also quieter, and are better left in the upper register or in the background, but are great for adding a bit of roughness to a ff accent; octaves are better off slightly detuned, and jump out to my ear a lot; vibrato feels like a slight volume increase. Basically a lot of trial and error that comes from combining the basic features in FamiTracker.