The PPG Wave in your Virus TI

  • A couple of comments sparked my curiosity about the sounds and characteristics of the PPG Wave, and how it could be replicated on the Virus TI. And since the Virus TI also offers Wavetable oscillators it didn't seem that far out of reach. At first. Then I spend a couple of hours watching YT videos and reading the original PPG Wave 2.2 Manual, and the documentation for the VST recreations offered by Waldorf. From all that, I understood that the Wave 2.2 (8 Bit) and Wave 2.3 (12 Bit) are actually pretty simple and pretty restricted regarding their VCO, VCA and VCF structure.


    Here's a compilation of the basic concepts:

    • There are 8 main oscillators, used to create the 8 voice polyphony. These oscillators can be split into two groups. Each group has their own settings for its oscillators. In general, this means that either all voices are used to create one sound with 8 voice polyphony or some of them are used to create a different sound, e.g. 2 sounds with a 4 voice polyphony. The groups can either be used for a keyboard split or as a layer. A main oscillator can be set to use any of the 31 available wavetables, or an additional fix wavetable called "Upper Waves" which is always loaded in memory.
    • Each of the main oscillators has a so-called "suboscillator". These, in contrast to common understanding, do not create a sub octave, but are more like a twin or derivate of the main oscillator which can be tweaked. The suboscillator always uses the same wavetable as the main oscillator, or the "Upper Waves".
    • All wavetables offer the basic waveforms Triangle, Pulse, Square, Saw as the last four waves (60 - 63).
    • There are 2 ADSR envelopes and one AD envelope available: Envelope 1 is used for the Filter cutoff and Wavetable position, Envelope 2 for the oscillator volume. Envelope 3 controls the pitch of the main and/or the suboscillator, or the Wavetable position of the suboscillator.
    • Each voice has their own analog 24db Low pass filter available. These 8 filters are digitally calibrated to match each other, but in reality have some differences and may drift.
    • The oscillators where digital with a 8 Bit (v2.2) or 12 Bit (v2.3) resolution. The audio signal was then sent to analog amplification and filter stages, which smoothed some of the digital graininess, but not all.
    • A very characteristic feature is the stereo panning of polyphonic voices. Each voice sounds at a different location in the stereo panorama, thus creating a wide spread when used with chords.
    • There is one LFO available per voice, which can generate Triangle, falling Saw, rising Saw and Square waves. It is also calculated digitally and therefore rather grainy to be less heavy on the CPU. The LFO can be used to modulate the pitch of oscillator and/or suboscialltor, the wavetable index of oscillator and/or suboscillator, the filter cutoff and the volume. The mod wheel always controls the depth of the LFO, the current setting of the mod wheel is saved with a patch.
    • Collecting all the available sources and destinations reveals a quite elaborate modulation matrix:
    Control Filter Wavetable Volume Pitch
    Key X X X
    Velocity X X
    Mod Wheel X X X
    Pitch Bend X X both or only SUB
    Aftertouch X X X
    Env ADSR 1 X OSC and/or SUB
    Env ADSR 2 X
    Env AD SUB OSC and/or SUB
    LFO X OSC and/or SUB X OSC and/or SUB



    Collecting all these informations, it may not seem too difficult to set up the Virus TI in a similar way:

    • Set OSC 1 to Wavetable
    • Set OSC 2 to:

      • Wavetable using the same wavetable as OSC 1 to use it as "suboscillator". Set the balance either to -64 ("suboscillator" off) or 0 ("suboscillator" on)
      • Wavetable using a different wavetable than OSC 1 to use it as main oscillator for Group B, use Unison set to "Twin", to create "suboscillator" doubling for both. Make sure that the Unison LFO Phase is set to 0.
      • Classic and use Triangle, Saw or Square to replicate using these basic wave forms which are available in every PPG wavetable. Set the balance either to -64 ("suboscillator" off) or 0 ("suboscillator" on).
    • Set the balance either to -64 ("suboscillator" off) or 0 ("suboscillator" on)
    • Use only LFO 1 or LFO 2 set to Triangle or Square and use the Mod Matrix to modulate additional destinations.
    • Use a Mod Matrix slot with Random source to modulate the Panorama to approximate the original stereo spread of the 8 voices.
    • Use a Mod Matrix slot with Filter Envelope source to modulate the Shape/Index of OSC 1 & OSC 2 independently.
    • Set Filter Saturation to "Bit Reducer" to achieve some of the grainy 8 Bit character. On the PPG, the digital oscillators go into an analog 4 pole filter, so some of the graininess is smoothed out. Unfortunately, we can not change Filter 2 to 4 pole. Therefore the closest we can do is either to run Filter 1 and 2 with cutoff and filter link in Serial 4 mode thus applying the Bit Reducer between the filters. Or use only Filter 1 in Seres 6 mode apply the Bit Reducer with a less severe setting.
    • Since the PPG Wave included a sequencer capable to record modulation parameter changes, you can use a Mod Matrix slot with the Arp Input source and destinations like Shape/Index and Filter cutoff.

    However, fact is that the two main ingredients for the PPG Wave sound are missing: The original wave tables (!!!), and the characteristic SSM 2044 24dB filter. Since some highly regarded sound designers claim that not even the Waldorf Blofeld, which offers all the original PPG wavetables and an emulation of the PPG filter, is able to properly replicate the PPG wave sound, this whole exercise is more about how to create "PPG style" sounds than to replicate any of its factory patches or well known sounds from world famous songs.



    Anyway, just for fun, I tried to do that typical "PPG Choir" using the "Pingvox" wavetable and ended up with a couple of variations which sound acceptable to me.


    Enjoy and share your thoughts and experiences.

    Dateien

    Bass Player and Synthesist.
    Virus TI2 Darkstar | Virus TI2 Desktop | Sub 37 | Voyager RME | Machinedrum | Analog Four | Digitone | MPC Live | NI Maschine+
    Mac OS 13.5.1 (Ventura) | Cubase Pro 11.0 | Ableton Live 9.6 | Logic 10.4 | MainStage 3.4 | NI Komplete Ultimate 13 | RME Fireface UFX+

    5 Mal editiert, zuletzt von ozon () aus folgendem Grund: Update Patches. Fix table.

  • Oli, this is incredible! You have such a great talent for analyzing a subject (in this case the PPG Wave) and disseminating an objective, lucid, and thorough overview, and then prescribing suitable workarounds in the context of our application abilities (the TI2). Much appreciated!
    To me, the most interesting workaround presented here is using Bit Reducer Filter Saturation to achieve some of the PPG's grainy character. I might also suggest trying the wavetable oscillator Interpolation parameter to make wave scanning a little more coarse to possibly help emulate that PPG sound. Too bad the Blofeld doesn't have that.
    Also, of note, the original PPG 2.2 manual states that the wavetable suffixed waves are, "tri, pulse, rectangular, and saw". When I investigated this on the PPG and Waldorf wavetables as presented on the Blofeld, to my ears they were only Tri, interpolating to Square. interpolating to Saw.

    Two questions:
    "Wavetable using a different wavetable than OSC 1 to use it as main oscillator for Group B, use Unison set to "Twin", to create "suboscillator" doubling for both. Make sure that the Unison LFO Phase is set to 0."
    Is your suggestion of using Unison to create suboscillator doubling for two groups (Group A and Group B) actually more than what the PPG could originally do?
    And, did you mean the PPG is capable of LFO modulation of the wavetable index? (You just mention "wavetable").


    And, also of note from the PPG manual, the Upper Waves wavetable is actually a best of collection of single waves taken form the other wavetables. I'm not sure how it works but I understand it was supposed to be used as an "extension" of a regular wavetable.


    Anyway, once again, a big "Thank You!" for your very insightful research! I can't wait to try some of this stuff out on my TI2! Actually, PPG choir sounds which we've discussed before are what got me into synthesis in the first place. I am very curious to try what effect these suggestions will have on other PPG style patches as well! :)

  • Thank you very much for your kind words.


    I might also suggest trying the wavetable oscillator Interpolation parameter to make wave scanning a little more coarse to possibly help emulate that PPG sound.


    This is a very interesing and valid suggestion, indeed!


    It's not clear to me to which extend the PPG Wave already did interpolation. Because according to some compiled documents, about half of the wavetable waves were not actual samples but interpolations. Plus it already had fully algorithmic wavetables, specifically the table which emulated a pulse width modulation from square to narrow pulse.


    Also, of note, the original PPG 2.2 manual states that the wavetable suffixed waves are, "tri, pulse, rectangular, and saw". When I investigated this on the PPG and Waldorf wavetables as presented on the Blofeld, to my ears they were only Tri, interpolating to Square. interpolating to Saw.


    Interesting. I have to check that with an oscilloscope.


    Is your suggestion of using Unison to create suboscillator doubling for two groups (Group A and Group B) actually more than what the PPG could originally do?


    No. In fact the PPG could do up to four different waves at once. Which is one of the reasons why none of the Waldorf hardware synths are able to fully replicate the PPG sounds.


    Thanks to the additional four classic waves in every wavetable, and the Upper Waves which were always available, the otherwise slightly limited "suboscillators" became much more than just twins of the main oscillator. True, a "suboscillator" always used the same wavetable as the main oscillator, but it could be set to a different position, and therefore provide tri, pulse, square, saw or any of the other 60 Upper Waves.


    To make that more clear, using suboscillators and Group A and B you can do the following oscillator combinations:


    Group A





    Group B









    Main-Osc




    Sub-Osc




    Main-Osc




    Sub-Osc




    WT1(+UW)/4-Waves
    WT1(+UW)/4-Waves WT1(+UW)/4-Waves
    WT1(+UW)/4-Waves WT2(+UW)/4-Waves
    WT1(+UW)/4-Waves WT1(+UW)/4 Waves WT2(+UW)/4-Waves
    WT1(+UW)/4-Waves WT2(+UW)/4-Waves WT2(+UW)/4-Waves
    WT1(+UW)/4-Waves WT1(+UW)/4 Waves WT2(+UW)/4-Waves WT2(+UW)/4-Waves




    So in some sense, it could do more than the Virus. But you can easily compensate on the Virus and just combine two Parts to create one sound.


    And, did you mean the PPG is capable of LFO modulation of the wavetable index? (You just mention "wavetable").


    Yes, the PPG LFO can modulate the wavetable index. Original post edited for clarification.


    And, also of note from the PPG manual, the Upper Waves wavetable is actually a best of collection of single waves taken form the other wavetables. I'm not sure how it works but I understand it was supposed to be used as an "extension" of a regular wavetable.


    Sort of. If a modulation would end up with a wavetable index higher than 64, it would select waves from the Upper Waves. Although that seems to be a cool idea at first, I doubt that it does sound good, because the modulation would first select some of the suffix waves and then select some totally unrelated waveforms from the Upper Waves. I think the best way to think of the Upper Waves is like the Classic Oscillator on the Virus.


    What this all comes down to is, that you can use pretty much any combination of Classic or Wavetable Oscillators on the Virus (heck, even a Hypersaw with Density 4.0 or less), just don't go overboard with the Unison. Because whatever you do, the PPG has only 8 voices (max 16 oscs)... which still allows huge monophonic sounds!

    Bass Player and Synthesist.
    Virus TI2 Darkstar | Virus TI2 Desktop | Sub 37 | Voyager RME | Machinedrum | Analog Four | Digitone | MPC Live | NI Maschine+
    Mac OS 13.5.1 (Ventura) | Cubase Pro 11.0 | Ableton Live 9.6 | Logic 10.4 | MainStage 3.4 | NI Komplete Ultimate 13 | RME Fireface UFX+

    3 Mal editiert, zuletzt von ozon () aus folgendem Grund: oscillator, oscillator, oscillator

  • Updated set with new patches which are not of the choir variety.


    BTW, by reducing the LFO depth (and rate) of the Phaser to 0, the Phaser starts to act like a multi band resonator. IMO helps a lot to get some of the "hollowish" character of PPG sounds.

    Bass Player and Synthesist.
    Virus TI2 Darkstar | Virus TI2 Desktop | Sub 37 | Voyager RME | Machinedrum | Analog Four | Digitone | MPC Live | NI Maschine+
    Mac OS 13.5.1 (Ventura) | Cubase Pro 11.0 | Ableton Live 9.6 | Logic 10.4 | MainStage 3.4 | NI Komplete Ultimate 13 | RME Fireface UFX+

  • BTW, by reducing the LFO depth (and rate) of the Phaser to 0, the Phaser starts to act like a multi band resonator. IMO helps a lot to get some of the "hollowish" character of PPG sounds.


    Another trick I just found to replicate some of that "digital shimmer" that I hear in some classic PPG sounds is to fade in an oscillator of about a 25% or so pulse wave with some PWM and add dual Unison detune or Chorus. Maybe something like this is originally part of those PPG patches? Anyway, the caveat is that it uses up an oscillator to augment whatever PPG-style patch you're trying to program, and Unison or Chorus are of course, Global.
    Cheers! :)

  • oliAtBass - nice patches!

    Thanks!


    Would love to hear what you do with them.

    Bass Player and Synthesist.
    Virus TI2 Darkstar | Virus TI2 Desktop | Sub 37 | Voyager RME | Machinedrum | Analog Four | Digitone | MPC Live | NI Maschine+
    Mac OS 13.5.1 (Ventura) | Cubase Pro 11.0 | Ableton Live 9.6 | Logic 10.4 | MainStage 3.4 | NI Komplete Ultimate 13 | RME Fireface UFX+

  • Also, of note, the original PPG 2.2 manual states that the wavetable suffixed waves are, "tri, pulse, rectangular, and saw". When I investigated this on the PPG and Waldorf wavetables as presented on the Blofeld, to my ears they were only Tri, interpolating to Square. interpolating to Saw.

    Interesting. I have to check that with an oscilloscope.


    Oli, were you able to check on this and what were your results?
    Thanks! :)

  • Oli, were you able to check on this and what were your results?
    Thanks! :)

    Frankly, I had all forgotten about that. But I checked now. And...



    ... IMO those four suffix waves are missing completely (from e.g. Alt1, Alt, and Upperwaves). Which sort of makes sense, because the Blofeld offers them as a separate choice and allows to use a different Wavetable / Wave for each of the 3 Oscillators.

    Bass Player and Synthesist.
    Virus TI2 Darkstar | Virus TI2 Desktop | Sub 37 | Voyager RME | Machinedrum | Analog Four | Digitone | MPC Live | NI Maschine+
    Mac OS 13.5.1 (Ventura) | Cubase Pro 11.0 | Ableton Live 9.6 | Logic 10.4 | MainStage 3.4 | NI Komplete Ultimate 13 | RME Fireface UFX+

  • IMO those four suffix waves are missing completely (from e.g. Alt1, Alt, and Upperwaves). Which sort of makes sense, because the Blofeld offers them as a separate choice and allows to use a different Wavetable / Wave for each of the 3 Oscillators.

    Yes, the suffix waves are not on the two Alt tables. They were put together as a "best of waves" by Waldorf later down the line. But they should be on all other tables (except not sure about Upper Waves- I need to check-UPDATE: yes, suffix waves are on "Upper Waves") so make sure the Blofeld's "Wavetable Limit" control is set to off or you won't hear the suffix waves at all, (or the regular waves at the end of the Alt tables).
    The original question, though, was: are there only 3 suffix waves (tri, square, and saw, with interpolations) as I thought, or are there four (tri, pulse, rectangular, and saw) as stated in the PPG Wave 2.2 manual? The first 30 tables after the Alts are PPG tables, and the rest, except Upper Waves, are Waldorf's, but I believe the suffix waves are the same for all the tables that have them.
    BTW, Blofeld plays wavetables on the first two osc's only, but not on Osc 3 if I remember correctly. :)

  • Yes, the suffix waves are not on the two Alt tables. They were put together as a "best of waves" by Waldorf later down the line. But they should be on all other tables (except not sure about Upper Waves- I need to check) so make sure the Blofeld's "Wavetable Limit" control is set to off or you won't hear the suffix waves at all, (or the regular waves at the end of the Alt tables).


    The first 30 tables after the Alts are PPG tables, and the rest, except Upper Waves, are Waldorf's, but I believe the suffix waves are the same for all the tables that have them.


    True. I had forgotten about that.


    The original question, though, was: are there only 3 suffix waves (tri, square, and saw, with interpolations) as I thought, or are there four (tri, pulse, rectangular, and saw) as stated in the PPG Wave 2.2 manual?


    You are correct: On the Blofeld, the suffix waves are Tri, Square, Saw with interpolations. The additional range when deactivating the "Wavetable Limit" is from position 121 to position 127. So what you get are:

    • Pos 120: Last custom wave from table
    • Pos 121: blend between previous and next wave
    • Pos 122: Tri
    • Pos 123: blend between previous and next wave
    • Pos 124: Square
    • Pos 125: blend between previous and next wave
    • Pos 126: Saw
    • Pos 127: Saw



    BTW, Blofeld plays wavetables on the first two osc's only, but not on Osc 3 if I remember correctly. :)


    True as well. :thumbup:

    Bass Player and Synthesist.
    Virus TI2 Darkstar | Virus TI2 Desktop | Sub 37 | Voyager RME | Machinedrum | Analog Four | Digitone | MPC Live | NI Maschine+
    Mac OS 13.5.1 (Ventura) | Cubase Pro 11.0 | Ableton Live 9.6 | Logic 10.4 | MainStage 3.4 | NI Komplete Ultimate 13 | RME Fireface UFX+

  • Thank you for your research and confirming my tests, Oli!
    Although to some people I might be nitpicking, as a big fan of Access's and Waldorf's wavetables, I find it all very interesting!!


    So now the big question is: Did Waldorf change the way the suffix waves are now as opposed to the original wavetables on the PPG Wave 2.2 synth since the PPG manual states there are four waves in the suffix waves with, as it looks now, the missing one being a very narrow pulse? I'll put an email into Waldorf to see if they can solve this mystery! I wonder if Wolfgang Palm himself would respond to an email? :rolleyes:
    BTW, at the bottom of this page from Hermann Seib's PPG website you will find a link to a pictorial chart of PPG wavetables with the suffixed waves.


    http://www.hermannseib.com/english/synths/ppg/docs.htm


    Also, BTW, if I remember correctly, the suffix waves have a slightly different timbre than the regular VA basic waves on the Blofeld, so at least it makes a nice addition to the arsenal! :)
    UPDATE: After rechecking and not surprisingly, the VA basic waves have a touch more resonance than the wavetable suffix basic waves.

  • MYSTERY SOLVED!!


    From Holger Steinbrink at Waldorf:


    "Yes, we changed the four additional basic waveformes with the first Waldorf Microwave into three waveformes. All following Waldorf synthesizers with wavetable synthesis use the Microwave wavetable synthesis, not the PPG one. Except the recreation of the legendary PPG Wave 2, the PPG Wave 2 and 3 plug-ins."


    Whew! I can sleep now! :D

  • That whole discussion sent me again into experimentation mode and I started to explore the Blofeld (PPG) wavetables more thoroughly. By setting the PWM (wave index modulation) on the Blofeld to a rather slow pace and to scan the wavetable fully, I found out that many of the wave tables are actually samples of instruments or.... voices. Check out the tables 1-2-3-4-5 and 19/20. Those are actually spoken words!


    Because of that discovery, I was curious what the Virus wavetables would reveals if scanned slowly in the same manner. Some of the tables sounded a bit like bells or string decays, but nothing like the complete samples found on the Waldorf tables.


    However, I found (again) a new character of the Virus which I have not explored yet: Using evolving or constantly morphing sounds by means of slowly scanning the wavetables. Much fun. And very different from the well known and typical hypersaw sounds.


    Please find some experiments attached.

    Dateien

    Bass Player and Synthesist.
    Virus TI2 Darkstar | Virus TI2 Desktop | Sub 37 | Voyager RME | Machinedrum | Analog Four | Digitone | MPC Live | NI Maschine+
    Mac OS 13.5.1 (Ventura) | Cubase Pro 11.0 | Ableton Live 9.6 | Logic 10.4 | MainStage 3.4 | NI Komplete Ultimate 13 | RME Fireface UFX+

  • That whole discussion sent me again into experimentation mode

    Heh, heh! 8)

    and I started to explore the Blofeld (PPG) wavetables more thoroughly

    The first proper 30 and "Upper Waves" are Wolfgang Palm's original PPG tables, the rest (including the Alts) are Waldorf's several years later.
    Update: I see I repeated this info again here. Sorry- I forgot to take my memory pills! :rolleyes:

    Check out the tables 1-2-3-4-5 and 19/20. Those are actually spoken words!

    Yup! And of which you can isolate vowels or even syllables for some gnarly vocal pads!

    Because of that discovery, I was curious what the Virus wavetables would reveals if scanned slowly in the same manner. Some of the tables sounded a bit like bells or string decays, but nothing like the complete samples found on the Waldorf tables.

    An interesting observation! Pseudo-granulation. I know some of Waldorf's tables are a few single waves of an instrument sample with calculated interpolations (#60 Xmas Bells is from a recording of one of the employee's parent's door bell chime! Ha!) and some tables are generated all algorithmically, but maybe most or all of Virus's tables are all algorithmic?? :huh:

    However, I found (again) a new character of the Virus which I have not explored yet: Using evolving or constantly morphing sounds by means of slowly scanning the wavetables. Much fun.

    With a slow triwave bipolar poly LFO on a wavetable index of 64, you can get some really nice pads with some of the tables! :P

    And very different from the well known and typical hypersaw sounds.

    Indeed! :D

    Please find some experiments attached.

    Thank you!! :)