Most efficent way to "emulate" my Ensoniq SQ1?

[Parduz]

Hello from Italy.
In order to have my questions make senses, i have to write a fairly long premise..... sory for the long post.

I'm a programmer and an hobbyst musician. It was already about 6 years ago when i started to try to have my good ol' SQ1 sounds available on my PC, as a soundfonts or using some VST instrument.
I was able to dump the samples ROMs, but only this year i managed to get from the SQ1 firmware all the samples data (start, end, loop points, finetuning, etc).

With the advent of SFZ format i think i may have found the right way to finally get back my SQ1 sounds. I think that the TX16Wx could be my choice.
What i have
What i have now is a 2MB wav which is the whole SQ1 Sample ROM, a table of samples name with start, end and loop sample, root key and fine tune.
SQ1 sound synthesys
Each "patch" in SQ1 is made from 1 up to 3 "voices", each using one "wave" in the OSC.
This is how a single voice is made:

Some of these "waves" are in reality multisampled sounds, so (back to my table and wav) they are various samples assigned to different keyranges.... we could call it "composed waves" using "group of splits"...right?
The question
Now, it's time for my question:
How would you organize things in order to have a bank of the 80 SQ1 ROM patches?
This means a bank of "up to 3 voices" patches, where each of these voices will use the same "groups of splits"?
it is more efficent to keep my single, big WAV and define the various "zones" by loop points or is better to split the big wav in its various pieces?
Any other suggestion?
Or even a "no go" 'cause i'm missing some keypoint that it could'nt be done using the TX16Wx?

Thanks a lot

[elcallio]

Well, the sound shaping diagram you attached should be easy enough to emulate in TX programs.
You will need to split the ROM samples into separate WAV:s though, the TX does not have support for using wave regions in the way you suggest (at least not with looping preserved).

I'm not sure if ripping the SQ-1 roms are 100% legal (copyright etc), but if it is would you consider sharing the samples (and programs as well). Being an old-school guy myself I love the idea of using these kinds of tight, small samples to build instruments. So obviously I love what you are doing here.

[rockstar_not]

I seem to recall that the SQ1 sounded pretty close to the ESQ-80. And there's a VST for that already: http://www.kvraudio.com/product/sq8l_sq ... nn/details

[Parduz]

Well, the sound shaping diagram you attached should be easy enough to emulate in TX programs.
You will need to split the ROM samples into separate WAV:s though, the TX does not have support for using wave regions in the way you suggest (at least not with looping preserved).

Thanks. So my approach building SFZ files like this:

Code: Select all

	<region>	offset=318965	loop_start=334150	end=349177		loop_mode=loop_continuous	pitch_keycenter=G3				hikey=54	// STRING ENSEMBLE 1
	<region>	offset=349179	loop_start=359466	end=370970		loop_mode=loop_continuous	pitch_keycenter=C#5	lokey=55	hikey=77	// STRING ENSEMBLE 2
	<region>	offset=370972	loop_start=391139	end=409476		loop_mode=loop_continuous	pitch_keycenter=F6	lokey=78				// STRING ENSEMBLE 3

is deemed to failure. Good to know

Your kind answer also mean that i will bother you with question about the various parameters and how to emulate the voice structure.
The plan (once i have the individual samples) is to write some fast source-code that will convert the sysex with the sounds patches to TX prog files. The SQ1 manual was fairly accurate and provided the time or freq table for the display values.
This will not create nothing accurate (i guess the filters will be the most hard part to tweak) but at least i will have a starting point.

I'm not sure if ripping the SQ-1 roms are 100% legal (copyright etc), but if it is would you consider sharing the samples (and programs as well).

Frankly, i'd love to know if it is legal as well. But the same samples (and firmware dump) are now running into the MESS emulator (a MAME project: the guys there are the ones that provided me the samples loop table 'cause they know the Motorola 68000 that runs the SQ1), so i trust they judgment about the freedom to use it, as long as there's no money involved.

So: i still have some "black hole" in understanding some of the samples data, but once things will comes alive you'll know

I seem to recall that the SQ1 sounded pretty close to the ESQ-80. And there's a VST for that already: http://www.kvraudio.com/product/sq8l_sq ... nn/details

Not really. the SQ80 sounds much more "older" and "electronic" (sorry, my english is not good enough to describe sounds).
The SQ1 is (to my eyes -or ears-) the US answer (in term of timbre) to the Korg M1. I never liked so much the "Japan flavour" of Korg M1 and Roland keyboards, and being the SQ1 in the same price range i choosed him. Sure it's cheesy but the organs are fantastic, and the multieffect section makes everything sound great. I still love it from the bottom of my heart (not saying Korg is bad: i also have a Wavestation A/D which i truly love as well).

[rootberg]

I really like the idea behind your project Parduz, keep us updated if you will!

One of the most annoying developments in software instruments, in my opinion, is all the 10Gb+ of slow loading, ram-chocking, bloated... etc sample libraries. I loved to play my Motif Rack ES, it was extremely musical and it had a capacity of 175MB of wav roms and the rest was synthesized. Same with my Roland D50... and as you mention the Wavestation.

I'm all for characterful, fun, big sounds in small sizes!

[Parduz]

I really like the idea behind your project Parduz, keep us updated if you will!

One of the most annoying developments in software instruments, in my opinion, is all the 10Gb+ of slow loading, ram-chocking, bloated... etc sample libraries. I loved to play my Motif Rack ES, it was extremely musical and it had a capacity of 175MB of wav roms and the rest was synthesized. Same with my Roland D50... and as you mention the Wavestation.

I'm all for characterful, fun, big sounds in small sizes!

Eheh, i ear you
I admit that at the times i wanted to have more samples to work with.... but i agree that 10Gb to find a couple of nice moving pads are a bit too much

I guess it also depends on the genre.
I'm slowly converting my old music from "real instruments" (the SQ1, the WS A/D, an FB01 coupled with a multieffect for distortion guitar emulation) and cassette tapes to "all virtual" and MP3.
If you're courious, here you can listen to my pieces: these are just the simpler and faster to convert, so maybe they're also a bit "dull". I'm working hard on some complex composition but i'd really need my emulated SQ1 and i have to work hard to get a decent guitar simulation.

[sraymar]

I've got an SQ1+ and a few other older Romplers but I'm not ambitious enough to clone them. However I have sampled some patches from them then modified them in Cool Edit and VAZ modular. I prefer to use TX16Wx, as far as my own sampling goes, for exotic samples and layers.

[theUnTall]

AUTO-ADMIN: Non-MP3, WAV, OGG, SoundCloud, YouTube, Vimeo, Twitter and Facebook links in this post have been protected automatically. Once the member reaches 5 posts the links will function as normal.

Dear Parduz,

was your project successful ? If not, you might try Crystal instead of the TX : http://www.greenoak.com/crystal/about.html (http://www.greenoak.com/crystal/about.html)

It is free, provides 3 OSCs with many filter and envelope options AND you can use own waveforms (stored in soundfount format).
Looks ideal to me for your project.

[MaxKosmos]

Hi @Parduz,
I'm old user and lover of SQ1, and I'm trying to follow your steps.
Can you tell me how did you dump waves in rom out of SQ1? Or better share them?
Thank you so much.
Ps. I shared time ago all SQ1 card I owned. Hope you already have or love them.
I also created a patchbay naming for Reaper.

[Parduz]

HEY!

As i work in an electronic lab, i used our EEprom programmer to read the chips. It required an inversion on one pin, i can't remember, it was more than 10 years ago

I'll PM you this weekend.
Meanwhile, pls make a zip of all your files/patches/whatever .... i think i may have missed them

[fercsi]

I have an SQR+32 Voice, which is a slightly improved version of SQ1 (in terms of sound creation). I checked out several things inside it. So, first of all: the ROM content which they have contain really low quality wave signals. The magic happens in different hardwares on those signals: MC68000 CPU, ES5505 (Ensoniq OTIS R2), ES5510 (Ensoniq ESP R6).

The CPU controls the operation of both others with a 1kHz frequency refresh rate (with tons of continuously varying features, depending on the key hit level, the key itself, LFO, controls, etc.). Even the envelope is more complicated than as usual (e.g. amp env attack is not linear). This part is really complicated in this family.

The 5505 reads the ROM with the desired pitch and aplies a 4 pole filter. This can be simulated quite well (however it's control from the CPU not!).

But the signal processor is really a cool device. And the guys went over the top programming it. So this is also something which cannot be simply simulated.

So, I can't imagine how some could create a tool without knowing the source of those sound modules and spending a huge amount of time recreating it.

But if anyone knows such a tool, I am also interested in that .