The VP-330 project
  • OK, so this is a very long project I started months ago. I had this in mind since the first version of Audulus, I however had to wait for the modulated delay lines to begin the design of this project.

    My goal is to produce effects and instruments based on the Roland VP-330 and its brothers, the RS-505 and the RS-09. They have in common to be string machines with 4 delay lines ensemble units, which are my favourites.

    Today I release a chorus effect based on the VP-330 design. I do not have the privilege to own a VP-330, so I used the service manuals as references to fine tune the timings and LFO amounts, as well as samples coming from different sources.

    I have not modeled the BBD audio artifacts at all, so don't expect a perfect emulation. I however added some features compared to the original:
    - The original was mono -> stereo, this effect is stereo -> stereo
    - You can control the rate and depth of the chorus
    - Inspired by the RS-505 design, I've added a feedback loop that turns this chorus effect into a flanger. You can control the flange amount.

    A few sound examples:

    Have fun!
    VP-330 Ensemble v1.0.audulus
  • Wow! That's super cool! I'm going to use your patch for some ambient guitar stuff :-)

    (Btw, I just upgraded the forum so your soundcloud tracks appear right in the thread :-)

    - Taylor
  • Thanks! Yes I think it works extremely well with guitars, not sure the originals were used this way very often though. The fact that you couldn't adjust depth and rate was a problem I guess.

    I must add that at low rate settings (and to some extent lower depth) the effect becomes more subtle and is closer to a spacializer. The design is actually close to the roland Dimension D.
  • A little audio test of my upcoming effect for Audulus :

  • wow, this VP-330 human voice sounds realistic! I'm interested in the filters you used, because I've always found it a bit difficult to create that human voice timbre, be it with hardware or software.

    Besides, XilsLab recently announced the XILS V, a very good VP-330 emulation.
  • Yes, very nice formant filtering going on there!
  • Thanks!

    Well Phil999, I used the exact filter parameters derived from the VP-330 service manual, ie frequency, resonance (Q), and gain. And I could compare the results with a set of samples I already own. The effect is almost ready but I have a few things to add, plus I have to find a way to implement both human voice and ensemble in the same effect on an Ipad 2... I'm CPU limited.

    I know about Xils V+, that´s really cool. But I discovered its existence very recently, and that´s partly why I wanted to post audio tests of my human voice effect before its release. I did all this research myself, it took a lot of time and I didn't want to be suspected of using Xils work as a source.

    In the end my plan is to release a full instrument. I have some nice ideas and Audulus seems almost ready feature wise (except that I need 6 db filters).
  • ça c'est tres bien. I also got to know XILS V recently, and took part of the recognising contest between hardware and software. Like always, I'm sure I'm able to hear the differences, especially as a VP-330 owner. Unfortunately I only know the sound in my ear/memory, the device itself 'passed away' ten years ago. But your human voice sample made it immediately clear you're on the right track. You use the right frequencies, it's audible.

    Glad to meet a VP-330 admirer, it is such an extraordinary instrument. One day I will have my VP repaired and can use it again like in the old days. Meanwhile, software will give some support or replacement (Eiosis for vocoder, G-Media Stringmachine for the very important carrier).
  • That's great to read such positive comments from a VP-330 owner. Once the human voice effect is released, I will post technical details if you wish. Yes, I love that sound, and thanks to internet (and Audulus), I can construct sounds that are close, it's fantastic!
    I hope that your VP will be repaired soon!
  • Ok, time to update this thread with my new effect, the VP-330 Human Voice! Before talking to much, an audio example:

    Done with Sunrizer + Audulus, recorded in Beatmaker 2 with audiobus. No additionnal effect.

    This effect features: all VP330 voices (upper Female 4' and Male 8', lower Male 4' and 8'), vibrato with rate and depth control (original settings), voice amount, VP330 ensemble model with Dry/Wet, rate and depth control. The human voice effect is obtained by using 7 banks of 2 BP filters, each voice type being assigned to 4 banks with varying gains. In audulus, I used 12db/oct HP and LP filters instead. Though I reproduced the original upper and lower settings (corresponding to notes upper and lower than C3), and the possibility to layer them, it's currently not possible to get the original instrument layering features due to the nature of this patch (an effect affecting the incoming sound, with no keyboard control). However, my initial intention was to use pitchshifters to have at least the possibility to layer the upper or lower voices. It was unfortunately impossible on my Ipad 2. You can try that if you wish, otherwise you will have to do overdubs like in the example above.

    Now time to thank a few people: I extensively used afta8 sub-patches in this patch, so thanks to him. I could understand and calculate the right gains, filters frequencies and resonances thanks to Yves Usson in this thread ( , in french) and in his website ( , deliyannis filter).

    Comments appreciated!

    VP-330 Human Voice v1.1.audulus
  • nicely done. Thank you very much for sharing, and the associated info.

    Edit: I've been testing this patch with various virtual string synths, on iPad and Mac, and the result is awesome. To finally have this human voice filter arrangement is a revealing experience.
  • Wow. Thanks, I'm glad that you like it!
  • can hardly express what it means to me. I simply didn't find a replacement for that sound. The samples one can find on the internet have all the ensemble effect in them, I couldn't use them for chords because then the ensemble effect gets duplicated and everything starts to sound wrong.

    Here is a quick Lemurtemplate for the patch. Nothing important, it's just for easier testing. And in the attachment the premapped patch.
    VP-330 Human Voice v1.1c.audulus
  • Hi folks,

    I tried to use the patch on Audulus for iPad, but it doesn't seem to work. I tried all the versions. The patch loads successfully but playing it I get no sound and a strange noise.
    Did someone else try it on an iPad?
  • @aantetomaso, trying it out now :)
  • The Ensemble effect is great !
    However I can´t figure out how the Human Voice works. I tried to insert the patch as a AU Instrument in Logic but it seems to be no MIDI input ? How do you play it...sorry for the noob question :D
  • try as an insert effect.
  • Thanks Phil999 , no I can´t get it to do you do to play it with MIDI ?
  • I've used it on the iPad. Inside Audiobus as effect, similar to insert effects in Logic.

    You run audio through it, not MIDI. Maybe that's your problem?
  • Ahh sorry , yes I thought the Choir was played with MIDI. So you run audio and the choir sound will follow the audio , kind of audio to CV... ?
  • Now I have tried to run simple sinus waves to trig the choir sound and something is starting to happen ! It would be nice to have a MIDI to Choir too ;)
    Ahh , it sounds best sending a clean SAW tooth.
  • yes, sawtooth waves work well. These are used in classic sting synthesizers like the VP-330.

    MIDI is only note message (and control message). It's not an audio signal. Check out
  • Yes thank you. Watch my vintage synth demo videos on Youtube. I own quite a few poly synths , however not a VP-330 , yet ;)
  • I found out the ultimate sound source for the Human Choir is the String 3 preset on my CS-80 will post an example soon.
  • I did a short video demo when playing the CS-80 through the Vp-330 patch.
  • @Firechild, that sounds great!
  • Firechild, now I understand the misunderstanding. Nice sounds you got out of your CS-80.

    One can use any source with these brilliant VP-330 filters built in Audulus.
  • I've been looking for an iPad ensemble effect for a couple of years, and bought Audulus specifically for this patch. It sounds good, but there's a pronounced drop in volume. I didn't notice this on the demos -- am I doing something wrong?
  • I am using Audulus on OS X and definitely can´t hear any drop in volume with this patch. iPad users should chime in, I don´t use that version.
  • I think I figured it out this morning: the patch is distributed with the crossfader split between voices and ensemble. If I turn the crossfader all the way to ensemble, there's no volume drop.
  • Hey guys, it's been a while! I should come back with a new project, hopefully in Audulus 3.0. Glad some of you liked this one, and thanks Firechild for the CS80 demo, amazing seeing my patch associated to this beast.

    Feel free to ask questions or post suggestions.
  • I have nothing to say or contribute. I just want to say thank you again for your work with this filter bank.

    Indeed the CS-80 demo is very impressing. I want to thank Firechild as well.
  • I want to say thanks for this patch too!! I bought Audulus (and upgraded it) because I love this ensemble patch so much!! (Just like the user said in a post above)!! Great Stuff!!
  • Will someone do me a huge favor and make this 3.0 compliant, i.e., make it look neater? If you do that, I'll put it in a module and add it to the module library - I just have other things to do, but I think this would be great, one of the better Audulus patches.
  • Yes I'm going to do a v3 version of these patches, this will be my way to learn how to use some of the very promising new toys.
  • Thank you so much! The idea is you can design with modules rather than just nodes, and pack those modules in an UI. Believe me, it's infinitely faster and more creative than working with nodes alone. It's very easy to quickly build out a section and then tear it back down - all while the individual modules are clearly labeled in the signal flow, which gives this side benefit of making other people's designs more comprehensible. If everyone makes things from nodes, it's not always clear what's going on in a patch where. As someone who's learned a lot from other Audulus users's designs, I know that if people start designing this way, it will be much easier to "read" a design without needing a lengthy explanation. It also makes it much easier to clip out an element of a patch and put it into another design.

    There's no extra CPU time, it just makes the file size a little larger since there's more data - you know, lines of code exist for where things are placed on a UI, there are redundant controls in some places.

    BUT, if you want to super optimize your patch, you can go about designing with the modules and then once your design is set, you can work piece by piece, copying the guts of each module out in to the layer above and then deleting the original module. For the most part, this merely makes the file size smaller - it has no affect on the CPU time, because static elements like inputs and outputs and lights and text aren't continuously evaluated like points in a signal chain are.

    If this is confusing, don't worry - I'm going to make a video tomorrow about how to do this.

    And again, thank you so much for this. This is truly a great effect you've made/cloned - the human choir is jaw-droppingly good. I would buy it if it were a standalone plug. Will definitely try it out even in just mixing stuff to add a little dimension on a buss.
  • Hey biminiroad, thanks for your comments and kind words. Well my plan was to rebuild the subpatches and using existing modules whenever possible, and of course build a proper UI. Will definetly check your video out to better understand your point.
  • OK it's time to post the updates of these patches. As of now the audiobus fix isn't available yet, but I could do some tests using IAA and everything seems to work fine. The 3.0 Audulus upgrade did mess up the original patches so it was again a lot of work (along with adding a few things and a new design).

    The VP-330 ensemble update just provides a proper UI. Just remember that it is not meant solely to reproduce that string synth sound. With all the available adjustable parameters it's excellent with guitars, bass...

    With the human voice patch update I introduced the original delay time adjustment on the voice vibrato. I've added a threshold control over the gate that will trigger the vibrato. Adjust this according to what you're playing, best results are obtained with sounds featuring low release times and low velocity sensitivity.
    Upper corresponds to everything above C3 (261 Hz), lower below. Remember that this effect is designed to process external audio, so choose the settings accordingly to what you're playing. In a next patch I will adapt this effect to a use inside Audulus.

    Thanks to the Audulus team for providing all these new toys and features, this is overall much more usable than before.

    Feel free to post comments and music you've done using these patches.

    VP-330 Ensemble v3.0.audulus
    VP-330 Human Voice v3.0.audulus
  • thank you very much.
  • Small update:

    - Now you can switch 2 delay lines off to produce a more traditional chorus effect. Set the dry/wet knob to 50 % for a stronger effect.
    VP-330 Ensemble v3.1.audulus
    VP-330 Human Voice v3.1.audulus
  • @SansNom- Great work man! I love the module. In the first version there was some drift applied by a random S&H. Is that gone now or am I just missing it is in the patch? I thought it was a cool idea to give it some "analog" drift. Also I noticed a lot of math in the "round" sub-patch, so I threw in a quicker round expression. I hope you don't mind. The performance boost is minimal, but every little bit helps.
    VP-330 Ensemble v3.1.audulus
  • @drewyeah - Thank you. Yes it's definetly better with a proper use of the math node, I still have to learn all these logic expression.
    As for the drift, you're right it's gone. Actually it's gone since the release of the original HV 330, for several reasons:
    - the patch has been done on an ipad 2, and I needed less load on CPU.
    - this drift effect was practically inaudible, it was more psychological than anything else.
    - the drift settings, and the fact that only 2 LFOs were drifting, were arbitrary unlike the rest that was closely modeled from the original. I have no clue if and how the LFO are drifting. Actually, based on samples I've studied I'm pretty sure they are quite stable. Even without drifting the ensemble is alive because of the right frequencies and depths of the LFOs.

    If you think that one can hear the difference, I can definetly put the drifting back. However, in the priority list I would focus more on the BBD emulation, though I'm convinced that the BBD influence on the sound is very subtle. I did some research back in the day and from what I recall, there is only a few % distorsion for some given frequencies. What all these analog components introduce massively though is noise, especially on the human voice section.
  • This is more valuable to me than my TC and MXR chorus pedals, for both guitar and analog synths.
  • Do I have to buy the Audulus app again ? No upgrade path to V3 from Audulus v2 ?
  • Ok , You have to be on sad.
  • Hey Firechild! Unfortunately it seems that Apple doesn't allow upgrades...
  • Sans Nom - VP-330.audulus
    Screen Shot 2016-01-06 at 9.05.20 PM.png
    1920 x 1080 - 291K
  • @Mark - OK no problem, I understand the library constraints and users can still use my version anyway. A few remarks though:
    - ideally Audulus would support presets and there would be a preset called "VP-330 default" where the rate, depth and dry/wet knobs would be set to 1. But it's not the case unfortunately, this is why I had to write "Default values: 1". In your module there's nothing. So It should be in the commentary I guess and the module saved with knobs set to 1.
    - why no VU-meter on the output? This effect can produce a huge drop in volume, I wanted to user to fine tune this setting and compare the 2 VU-meters directly.
    - I prefer the more conventional dry/wet terminology (than "effect").

  • @SansNom - I changed it from Dry/Wet initially to save space - I pushed the knobs closer together initislly - but theres room and have no problem adding it back in :)

    There is a new preset function thats coming soon in 3 - I saw the default values and am going to save it like that they're still in the design just inside rather than on the front panel. I just grabbed a screencap while I was playing around with it.

    Both the input and ouput VU meters are on the left side - I thought they would look better collected and sort of stacked pseudo 3d on the side rather than at both ends - is lets that light element I expanded be more of its own thing visually.

    If you could write a metadata description for it, that would great. Or, I can write something for you - just describing the effect and how it works and what each parameter does. Maybe a little history of the real effect you're modelling too? Will upload my version of the Vox Humana one later - same thing just condensed it.