And my other cover placed 15th, so I'm pretty satisfied with that too. :)
I'm not a bad spott, but I dont think the Beethoven cover should've places 1st, if only because it was submitted in an unfinished state. But in fairness the first movement is so awesome it almost makes up for it.
I really liked LIQUID BOOTY EXPERIMENT. I am furious it ended up so low.
Also lol @ my wrong guesses. I totally thought mokimaF was FoD. I also totally thought Scaling The Dragon Fortress was Diad. And there were quite a few others I got wrong. At least I submitted half of my comments anonymously!
Well I was 50% right when guessing that diad made cheetahmen...
Also I didn't expect a 18th place, really happy about that! I wasn't even that satisfied with the end result, because I ran out of time and didn't have time to polish it, but it seems that it didn't matter
I'm slightly disappointed in 39th in covers, but I guess it's to be expected on something that was honestly rushed. And I can't really top virt with Hey Ocean, no matter how many Streets of Rage callbacks and sweeping arps I throw in.
Allot of the placings this year are head scratchers...
Anyway, some info as someone from Japan was kind enough to do a hardware recording and pack of split chip channels. Copypasta...
---
Okay so listening to the split audio pack turned up some helpful information for other people I believe... I used nezplug++ primarily for audio and wavetable testing (and still rocking notsofatso to sort out timing issues.) The test this year was a ghetto fix for the 8 channel hissing issue that bit me last year (and nearly everyone who used N163 in the compo this year) by turning on only 7 channels.
In nez++, weird artifacts still occur that I assume are from sample rate or aliasing, but the banding is at lower frequencies. Pardon me as I'm prolly not using the right terms. However the hiss was gone so I went ahead with it. Problem is, every single note for that chip must be retuned which is a pain but doable, involved making a big chart that hits a certain range of needed note detunes with one SA value. Nez++ ended up sounding very close to hardware with 7 channels with a couple issues - the wavetables seem to have slightly more bite to them, or just slightly more harsh (again, I guess the aliasing?) and the 8 channel hiss still actually remains, but MUCH quieter than all the 8 channel compositions, and plenty listenable! But it's still there and care must be taken to deal with the issue (if you care to, or just treat the N163 like piston collage or milkytracker because who cares =P) Ultimately Nez++ realmode still sounds closest to hardware and I'm happy I used it for testing, ended up sounding fine, however I believe 2a03 channels may need to be written softer as they were much stronger than regular settings. oops. Also super embarrassed to find an octave mistake on one channel for a short chunk of the song. heh.
----
Sadly the banding seemed to screw a bit with the harmony effects as they sound weaker. But that might be volume, and I prolly should've dialed the 2a03 channels even further back as they came out much louder than expected. The triangle ended up overpowering the doubled bass channel and kinda killed the Koshiro bass effect. All in all sounded reasonably good across all useful players and hardware.
Another note from earlier in the month; when paring channels together for instruments on previous projects, there was always a slight phasing issue just from how the chip itself works. Manually retuning every single different tone may have also ended up being a blessing in disguise as it seemed to cure this issue, but it may have been also because less channels being activated for the tune (with ppmck, if you call too many commands at the exact same time with all 8 channels going, the chip stalls pretty hard for a few moments and tries to quickly catch up back to place.)
So there you go. Tuning with Famitracker may be easier than mml or impossible. I don't know!
