DSP Robotics Support

The Quilcom Q-50 is a synthesiser inspired by the Roland D-50.

It’s not an emulation and the presets and embedded wave files are not a copy of the originals. However the architecture is very similar and I’ve kept the original’s nomenclature.

For those who just want to get on and try the presets here’s the Dropbox link (it’s 71MB):

https://www.dropbox.com/s/5b33p023h0zf4 ... m.zip?dl=0

EDIT 14.4.16 This is version 1.06 which now includes Martin Vicanek's excellent optimised reverb. No changes to presets or sound quality.

The original D-50 replaced the Yamaha DX7 in terms of popularity back in 1987. It used a method now referred to as sample and synthesis and was easier for users to understand, and use, than FM synthesis and yet was still all completely digital.

The principle of the synth was to use tried and tested familiar analogue-type metaphors and routing, but combine these with acoustic and other samples for the attack phase only. Our brains apparently recognise a sound with greater emphasis on the initial sound than the Sustain or Decay phases but this Attack phase is the most difficult to synthesise properly. Since memory was very expensive back then it simply wasn’t economical to create a full blown sampler so Roland supplied a built-in ROM with a 'whopping' 8MB capacity pre-loaded with 100 8 bit attack-phase and other PCM samples.

The basic architecture for the D-50 and Q-50 is as follows: A patch (preset) is made from 2 Tones called Upper and Lower. Each of the 2 Tones is made up of 2 Partials. The 2 Partials can each be either a Wave or a Synth. This means you can be using any 4 from 8 generators available.

Each of the 2 Tones has a static filter, a chorus/delay and a Ring Modulator. The final outputs from the 2 Tones are fed into a reverb.

Each Wave Partial has a choice of sample used, pitched or static, single shot or looped and can be tuned in pitched mode. The wave is then fed into an ADSR envelope stage.

Each Synth Partial comprises a tuner, waveform selector followed by a filter with its own ADSR and an overall amplitude ADSR.

Both types of Partial have, in addition, LFOs available for various functions and a Pitch envelope generator.

Of course there are differences in the Q-50 and I see these as improvements:
-Each Partial has 4 LFOs rather than 3. This means each one has a specific assignment.
-Stereo is possible since the Upper Tone is routed left and the Lower right. Width can be adjusted.
-The samples are 16 bit 44.1Khz so sound much cleaner
-You can load any single User sample instead of using the ROM
-There are more waveforms available for the Synth Partials and the LFOs
-There are more filter options available for the Synth Partials
-The Tone mixers can have levels set for Partials 1 and 2 AND a Ring Modulator; you don’t have to choose either-or.
-The parametric EQ is far more advanced than the D-50’s

A point of interest to sound designers is that you can make this your own instrument! In Flowstone the schematic’s top level allows you to easily load a different ROM for each Wave Partial, should you want to experiment. As supplied, all 4 ROMs (wave arrays) are loaded with the same WAV files. I have provided detailed instructions in the schematic to show how you do this.

The first 50 presets all use at least one of the samples. The Q-50 should really be thought of firstly as a synthesiser but which can also be enhanced with samples, and I made preset 51 (Fatima) to show just how phat this can sound without samples at all.

The zip contains the FSM, presets, the WAV files loaded and a PDF about the original D-50 which may explain better what’s going on. The exported VSTi will be found on http://www.flowstoners.com .

I should maybe add that all the samples I used were in the public domain and came with no usage restrictions.

I hope you enjoy!
------------------------------------------------------------------------------------------------
A bit more background on the D-50…
https://en.wikipedia.org/wiki/Roland_D-50
http://www.soundonsound.com/sos/1997_ar ... ndd50.html

Cheers

Spogg

Wow, congradulations, Spogg! Those are beautiful sounds! These ones struck me most:

Florence - nice traverse flute. I love the sound of a traverse flute, my attempts to synthesize it all failed miserably.
Betty - upright bass! yeah!
Desdemona - This layered sound is so expressive! Wow! That's definitely my favorite.
Tinkerbell - interesting rythmic pattern, reminds me of the "zapateo" of a Flamenco dancer. Pity that the speed changes acccording to pitch.
Mary - convincing marimba
Xena - authentic xylophone

Really really nice job, you certainly have a great imagination, and you know how to put it in practice.
Keep on!

...ohh wait,
remembering this keyboard thingy in the corner of my booth...

...just a second...


Ahh, here it is:


15 years on tour with that little baby

Another ghost from the past. The rack version D-550 was part of the synth rack in my boss' studio who allowed me to use it how often I wished to. I admit I used this synth just as a "rompler", the library of presets from my boss was awesome.

Here's what a song made solely with the D-550 can provide still today (found this after searching for images of the rack version): https://youtu.be/5T5WFGv9ZEM


On topic: If only you would optimize your synths so that they wouldn't use up to 50% of my PC's cpu. I mean, even the most demanding synth currently existing, the Diversion, doesn't use nearly as much
https://youtu.be/Y0aGeOPhj1E

Once again, some really cool sounds in here! I've never had the pleasure of playing a D-50, so I can't really compare, but I definitely had fun playing with it. Thanks for another cool toy!

Hey thanks guys for the nice comments!

@tulamide
I would really welcome advice on how I could make this synth use less CPU.

All the oscillators are by Martin (thanks Martin!) so are already optimised. He made the optimised exponential ADSRs too (thanks Martin!). The LFOs are Martin's (thanks Martin!). The Wave playback stuff uses Prims so maybe that could be improved?

It was important to me to replicate the features of the original and that had a lot going on like 2 chorus/delays, 4 active voice generators and 2 parametric eqs and so on and all hardware-based. There was a single hardware DSP that was multiplexed by a controlloing CPU for the polyphony.

I've made the schematic as neat as I could so if someone could take a look inside and make suggestions or offer alternative modules with the same function I would be very grateful and many others could benefit too.

I'm currently developing what I hope to be a low CPU additive synth so any suggestions might help me to better achieve that goal.

I have noticed that the VSTi plugin seems to use less CPU than when running the instrument in FS but that may be down to how it's measured. Dunno.

Cheers

Spogg

I hope you don't misunderstand me. The motivation for my moaning is that I really like the sound output of your synths, esp. the QX7, and it frustrates me that I can't use them in production.

I would have been accepting that my pc might just be to weak, but from all the VSTi I have, either free, commercial or demo, yours are the only ones that use so much cpu power. I took Diversion as an example because this monster has so much going on under the hood, more than your plugins. Still, it only uses up to 25%, which is half as much as the QX7.

I just assumed optimization from this experience. It might also be that it simply is more than Flowstone can handle. But then again I see realtime FFT used for zero delay convolution, and that is probably the most demanding task currently.

Also, while you get postive feedback, nobody seems to try using the synths in production. Maybe I'm the only one who actually wants to work with the synths?

Just a few thoughts:

You might compare the CPU usage of the wave oscs with the MV ones... I know Martin's oscs are more efficient than the stock ones, but often a sampled osc is cheaper than a real-time-calculated one (not sure if that's necessarily the case here). If the samples are cheaper, you might replace the synth oscs with sampled waveforms.

This next one comes down to how "smart" the selector is, but you might try changing your filter setup so that any filters that aren't being used are switched out of the signal path entirely. It's possible that your switching setup for the SVF/Ladder filters is sending audio through both filters when it doesn't have to. I could be wrong, someone please correct me if so.

That's all I've got...

PS: can't say I've used any Quilcom stuff in a track yet, but the next time I want FM, I'll definitely be reaching for QX-7.

Hello Spogg,

Wow, you really have a talent for creating some great sounding synth's. Some of the presets are beautiful! I really think retirement is working for you. You must get a lot of free time, no honey do's I guess, ha ha. One of these days if I every have lots of free time I should try a complete synth. Keep them coming. Why girls names for all the presets?

Anyway I thick Noisenerd maybe right, to many streams going at once. I have a similar problem in one of my delays.

It would really be great if you could get it where tulamide could use it for studio work.

Take care, BobF.....

Spogg, a few alternative modules that you might want to consider. The wave readers could also use some optimization but I don't have off-the-shelf equivalents ready. Same for chorus, and of course the Moog Ladder.