Synthmaker bundle

[nix]

Oh-it's pretty easy to cut the 4 files marked sq-80_waveform

They are just multiples of 2. So the entire files are 32678 samples. Chop this into multiples of 2 from the start.
So you can use 64,128,256,512,1024 or 2048 if you use the wt osc. U can ofc resample to whatever.
I think this works for us.

Will look at the wavhi/lo now

edit: in truth sq-80_waveform s don't necessarily have to be cut- we can index to them in that simple multiple of 2 way
-- but let me know what length we need, if you'd like me to cut them

[nix]

wave hi/lo need to be cut at 512 samples imo

...so we should be back in shape here to build a synth from these files

edit-The single cycles do not appear to be different lengths actually, so sorry for the prior mistake
edit2-So these 6 forms are great! so much thanks to Mr. Buchty and u 2

[tulamide]

Hey nix,

great effort, thanks!

Yes, as soon as we have the single cycles, we can resample them and change the bit depth to anything we like. What I still don't understand is

The single cycles do not appear to be different lengths actually, so sorry for the prior mistake

Do you mean that all single cycles are of an exact length of 512, and only the number of single cycles differ per waveform? In other words, if I cut esq1lo and hi in chunks of 512, I get a total of 128 single cycles?

Regarding zero-crossing: I had a look at a few waveforms from the aiff collection. They definitely are not zero crossings, but loop at any value.

[nix]

no worries!8D-so glad to finally help u Tula
sweet it means we don't have to manually part them I think.
yes, I think that the lo/hi wavs should be cut into 128 single cycles at 512 samples,
but it's not written in stone that we must.
The other ones may be easier to work with if they are a different(most likely shorter) power of 2((depending on their nature)if they seem to be a different period).
But you could pick 512 for everything?hmm, they seem shorter sometimes in the 4 sq-80waveforms

mmm-as far as looping at whatever- I don't recommend it-0 cross is safe.
Perhaps at wt-type periods they are so fast that it doesn't really cause a discontinuity.
It is a design decision you can take. A really fast fade is safe. U could make 1 sample fades if like AFAIK

edit-mmm, the fades are a little difficult, we would have to experiment. Maybe 8 samples is a rough figure?
I'm pretty sure that cutting in powers of 2 is what's needed here though. They seem to be cycles and sets

[nix]

here's the first hi file to look at linked.
something weird happened and I have resampled, but upwards, no data is lost. The single cycles(sc s?) are 1024 now.
I can share the chopper/cutter, but it is bad practice somehow. If u accidentally leave it on, u can fill ur HD. It's ugly and stuff(functionally- can try fix it too I guess.
It worked like this->

example.zip

(53.98 KiB) Downloaded 996 times

this is the hi bank, it has 32 forms, not 128
These are fine to work in SM, np

[tulamide]

here's the first hi file to look at linked.
something weird happened and I have resampled, but upwards, no data is lost. The single cycles(sc s?) are 1024 now.
I can share the chopper/cutter, but it is bad practice somehow. If u accidentally leave it on, u can fill ur HD. It's ugly and stuff(functionally- can try fix it too I guess.
It worked like this->

example.zip

this is the hi bank, it has 32 forms, not 128
These are fine to work in SM, np

Nix, you made the same mistake I made. That's why I asked Rainer Buchty for help. Please refer to the translation I gave from his reply. These are the facts:
1) The waveforms are of different lengths. If we come up (like we both did) with Waveforms of same lengths, we've done it wrong.
2) The only way to get to the start address and length of a waveform is through the DOC's WSR-value. We need someone who is trained in electronics and can read circuits and interpret the WSR-value. I tried, but failed miserably.
3) All waveforms are already perfectly looped by ensonic. If we need to crossfade or any other technique to run them, we've done it wrong.
4) The first few waveforms of esq1lo schould be variants of noise, which makes it impossible to get them right just by looking at the waveform display, since noise, by nature, is a bunch of random sample values. (I know, because I tried)

Isn't it ironic that we already fail at the assumed easiest part, ripping the original waveforms from the eprom?

[nix]

oh-ok
the saw series looked like a standard form, I assumed it is a saw
hmm- will try and comprehend

the hi file certainly has a 512 sample periodicity of some kind raw

[nix]

maybe we need to sample the VSTi?-- it's one way maybe

These cycles are an osc of some kind. There is some order to the cut.

Maybe I should hear how this existing VST sounds

[Halon]

Thank you so much for these Kevin! Having some serious fun with these. Seriously one of my favorite synths.

[k brown]

Very nice -

Which is one of your favorites - the EZQ-1?