607 wrote: ↑Tue Mar 02, 2021 8:45 pm
This isn't really a question, but I thought it'd be interesting to share here. I'm working on a cover of a pop song that has two voices for most of it. Of course, I could simply have both voices in separate channels, but that seems like a waste, doesn't it?
Depends on how important those voices are in the composition. If they're an important feature, maybe it's best to have those 2 channels reserved for the harmonies. Don't save channels just for saving channels. I don't know the song nor how you hear it, maybe you should write down the priorities, and implement what's most important, maybe there are instruments that don't add much to the composition, and that you can just drop?
There is guitar too, which would already leave me with just the DMA channels for all other instruments I'd like to include, including percussion.
Maybe you can pre-mix some samples in external software?
My usual solution is working with arpeggios, but I thought it would be nice to hear both voices separately. So a while ago, I had got the idea of enabling both the square and the wave/SID generator on an instrument, and enabling arpeggio on only one of them... this would allow me to create two separate voices playing at the same time. However, it doesn't work too well!
Are you using arp sequences with only 1 value, so that you actually don't have an arp? I achieve this result by using the fixed frequency commands instead
. Love that technique but I find it to be more useful for stab-like instruments than for harmonised melodies, maybe try having one melody at a different octave and see how it affects the interferences between the two melodies? The result here is significantly less good than the 2 channels version imo. The resulting timbres with interfering waves are great though, there's just not as much focus on the melodies so it depends what you're aiming at.
I thought it might be interesting to compare, so here is a recording of the two voices on separate channels (with the same custom wave instrument), a recording of the two voices on one channel using the technique I invented myself (although I'm probably not the first to try it
), and a recording of the two voices on one channel using arps in the conventional way, which is probably what I'll have to go with.
https://drive.google.com/drive/folders/ ... sp=sharing
Something interesting seems to happen in the square+wave experiment in the third measure. It sounds like the lower voice actually moves up a bit, even though it is programmed at the same pitch, as you can hear in the other versions. I guess there might be some interference going on?
Let me know if anyone recognises the song, by the way. ;P
To me the arpeggiated version is the least appealing, unless you really want an old school feel indeed. I'd say if you can find a way to keep the two channels for the melodies, do it, as I guess the harmonised voices are an important feature of the song. Also, I don't know the song but maybe some parts would benefit from one technique, and other parts from another technique? Maybe it's best to emphasize the harmonies in certain parts by using 2 channels, and having them on 1 channel when they're supposed to be in the background (if that ever happens)?