I also wanna mention that all of these vocal charts have been made to run independently of the tempo map for actual instruments. (I only tempo mapped Ultra Soul so far because I did some guitar work there) I thought that would be easier for future implementation. Everything runs fine in the tests I’ve run so it seems okay to do it this way. (If anyone more experienced has any insight on this causing later problems let me know)

The tempo map is the foundation of your song. Without a good tempo map, you shouldn’t even start vocals.

What I have essentially done with these customs so far is assign a certain type of vocal-specific tempo that should be quite close to the later instrumental tempo map, but of course is not as exact (but it is consistent, which makes for a smooth scroll). This is not achieved by a typical tempo map or envelope or anything, but from a gameplay perspective, everything seems to be completely fine and work well.

 

The reasoning behind this method I’ll describe here is to make things easier, make things a little more consistent and free of user-error (aka me ) and much less time intensive on my part. To begin with, I start with an accurate MIDI that accounts for all of the notes in the song. This MIDI is developed through a direct soundboard recording from the vocal stem of a MIDI engine (which I can’t rip directly). Once this WAV is in hand, I take it to Audacity with the original song audio. There, I sync them up and individually make sure each note is playing over the syllables in the song that it should be (very simple, takes less than 10-15 minutes per song at most).

 

After that, I export the synced MIDI recording in WAV form and use Melodyne to clean up any big artifacts and then export it as a MIDI. I then take this MIDI to Reaper along with the WAV recording of it. I sync these up at the same start point as the song by setting the MIDI track to run on “Time” as its timebase and doing a minor timestretch of the MIDI to make sure that this MIDI is an exact match for the corresponding WAV. This creates a “faux tempo” with Reaper’s delay function.

If you know how to do all that other massaging, you’ve got enough experience with REAPER to just do it right. My first tempo map took hours, and every other one has taken 30 minutes tops. Make a real tempo map, hopefully one marker per measure, and you won’t have to fuss.

Me doing it by hand with the tempo map of the song itself versus using the method described above, I’ve not seen a difference in the gameplay experience that would warrant investing the many extra hours in adjusting and nudging the chart around just to meet that exact tempo when a mostly-accurate/suitable one can be generated with that method

There are many players who rely far, far too much on their *eyes* when playing a chart. That population is generally oblivious to variances of 100 ms (or worse). The rest of us are very grateful to play charts with accurate tempo maps.

I’ve not seen a difference in the gameplay experience that would warrant investing the many extra hours in adjusting and nudging the chart around just to meet that exact tempo when a mostly-accurate/suitable one can be generated with that method…

 

If you only do vocals you can ignore the tempo map and essentially have a playable chart, that’s true. That’s why I only commented when you suggested your charts can work on any tempo map: if you do vocals only customs, it’s fine, but they are close to useless if somebody else wants to build a song around them or add them to their songs, because the tempo map will be different and considering vocals work on a1/64th grid (at least) all notes would need realignment.

Also, the animations will look really weird in-game and it will actually be harder to get the chart completely in sync with the music than if you work with a correct tempo map, so really there should always be a correct tempo map in all charts no matter the parts in it.