Chat

Author Topic: [SPLIT] Questions about samples in Defle  (Read 5417 times)

Offline marcb0t

  • Member
  • Posts: 125
[SPLIT] Questions about samples in Defle
« Reply #15 on: April 30, 2016, 04:53:13 am »
Sure, but I don't see why the goal shouldn't be to emulate at least one of the hardware versions, or the "on average" levels between them all. DM is now, what, 15% lower than an exported VGM sounds like in winamp, and lower than that compared to Regen IIRC. It's not a huge problem if I made it sound like that, just something that can be better.
Delek's goal was to make a tracker based on a particular cell phone FM chip from the late 90's/early 2000's. VGM's, emulators, and Mega Drives were not a concern at that point. Since then, it has evolved into so much more. But with sympathy for the programmer, it would probably require rebuilding the core programming which might not be worth all the extra effort. Especially if it is still possible to get accurate hardware sound through trial and error tweaking. Delek is working hard on working out lots of little bugs with the sound already, and there are probably lots of issues that we are not aware of from a programmer's perspective.

Keeping samples from changing pitch from DM to real hardware seems like a bigger concern, if I'm correct about that one?
They have a calculation as to exactly how many semitones in Deflemask you need to adjust your samples to, before hardware playback. Check out this thread: http://gendev.spritesmind.net/forum/viewtopic.php?f=8&t=1702&hilit=ocean+melody

Also, you'll find a hardware recording of "Ocean Melody" from a Model 2 PAL Mega Drive on one of those posts! ;)

I asked about this because someone recently wanted to make recordings of my MD covers on real hardware, a modded Model 2 I think it was, and I had to tell them how it's kind of a hassle to get it right.
Hehe, I look forward to hearing that, actually. Ask him to send you a test recording of your tracks, then adjust accordingly, and see how that goes. Trial and error for right now, bud.

Hey, didn't you use having very high TL's to achieve a saw sound in the lunar light zone track? ;) Besides that, I think it's a matter of preference or what you're trying to achieve. I already mentioned fuzzy noise artifacts, which for example Time Trax uses well for its bass sounds.
In terms of preference, you are partially correct, and I did make mention of that.

As for that saw sound in Lunar Light Past, Operator 4 is completely silent. Operators 2 and 3 are at 0, but Operator 1 is at 6. The instrument is not even playing at 3/4 of the FM amplitude. In this case it is safe to go above 10TL without FM clipping, even in Op 1. There is a method to this madness.

Here's the tricky part. Op 1 is the only Operator that is ever directly effected by FB (Feedback control). So having that one over 10 can make  or break an instrument. But as I said, it all depends on what kind of sound you are looking for. As a general rule, keep them lower than 10TL. But if you want grungy metal sounds, hi-hats, harpsichords, thick and sharp leads, or other distorted effects; that's when you want to throw that rule out the window! ;) Like with my funky saw lead!

Also, it was set to Algorithm 7 which ONLY Op 1 is effected by FB. It would be much harsher if I used Algorithms "0" because Operator 1 feeds into all the other 3 operators, and all 4 operators are effected by Feedback control of Operator 1.

Gosh, I hope I did not just confuse the living daylights out of everyone! :P