I've been hard at work with MMV2 almost daily for the last 6 weeks (and offline as a result) and that means the number of bugs has grown many fold. It's no longer practical to keep coming to this list and mention one at a time, so I've combined all my old bug findings with the new ones into one list. It's time to get Arturia off their asses and fix MMV2. Three and a half years is too long for this crap to remain in the disasterous state it's in. Anybody who cares about this product and has found bugs themselves should read this, and if those bugs aren't mentioned here, post them to this list, I'll gather them up in the near future, combine them with these and send them off to Arturia to do something about. I just can't take this crappy software anymore.
The following list is about bugs that are totally within MMV2, not so much about problems interfacing with somebody's particular computer or sound card, as those aren't experienced by all users. I've also included some comments about poor design issues or other problems that may not be bugs, but should be documented to show problematic or limited MMV2 behavior that shouldn't have been put there in the first place. Bugs specific to certain modules are placed under that module name. General MMV bugs follow after those under "General MMV2 Synth Bugs" heading.
(1) An osc has a total of 4 exponential modulation inputs, two on the osc and two on the osc controller. Connecting an audio rate osc or noise source into any one of those causes 3 of them to stop working. This is one of the most major module bugs in MMV.
(2) The sinewave can't hard/soft sync.
(3) Can't hard/soft sync more than one osc to the same source osc. Trying to do so keeps one osc sort of sync'd and the other not sync'd, with both making strange clicking sounds depending on their frequency settings.
(4) The second PWM nut improperly displays the setting of the first PWM nut when turned, so it can never be set numerically accurate.
(5) Saving a patch with an osc bank set to LFO mode may not get you the same sound when loaded back into the synth. In my case a polyphonic patch plays only intermittent monophonic notes. I have a set of patches that only work correctly after I manually cycle through the K1,K2,K3,K4,No setting back to the LFO setting after I load them.
(Poor Design Issues). The pulse width knob also changes the triangle and sawtooth waves into other waveforms. Unfortunately Arturia put the sawtooth wave at the leftmost position, and the triangle and square waves at the rightmost position. So you can't get sawtooth (at the proper pitch) and triangle and square waves at the same time from an osc bank. Mixing waveforms together on an osc is common, and when standard waveforms like sawtooth and triangle change into other waveforms as the pulsewidth is changed, it's annoying and requires bringing in another osc bank and syncing it to the first, using more oscs to acheive what should be done with just one. Those other waveforms aren't even very useful. Turning the pulsewidth knob gives you a hardsync'd sawtooth waveform, something that can already be done by simply syncing one osc to another.
(1) The envelopes are shaped wrong. Instead of a mild curve, they like a quarter of a circle with an additional flat part of about the same length that causes a delay or what seems like the envelope getting stuck before moving on to the decay stage. This causes all kinds of problems in getting certain sounds where a soft attack is needed on a percussive type envelope, or when using it to do large osc sweeps, or when using it to control the amount of modulation, and so on. And since the other stages are shaped the same way, release times that should have ended instead extend far beyond what they should causing more voices of polyphony to be used up causing MMV to make awful sounds when running out of processing power. Synth envelope stages aren't shaped like circles. It's typical on a Mg to control audio with an exponential VCA which somewhat counteracts the mild logrithmic curve of Mg env attacks making them closer to linear, which means Arturia has created the most un-Mg-like envelopes possible. Page 111 in the 2.0 manual shows what an envelope plot should look like. MMV doesn't behave anything like that.
(2) Envelope times are actually 7 times longer than displayed. Set an attack time of 100 ms and you actually get 700 ms for example. (Maybe I just don't understand the logic behind this because I'm not French, I don't know.)
(3) MMV envelopes are completely hosed when used with a sustain/hold pedal. The VCA2 envelope seems to be the only one that works correctly. Most of the time the VCA1 envelope stops triggering after all voices have been used up and only starts working again after the sustain pedal has been released. The general purpose envelopes are flakey, appearing to work in one patch, but not in others.
(4) VCA envelopes can't be triggered from the envelope follower module since the envelope follower trigger doesn't work unless the VCA envelope is already excuting its envelope. Chicken/egg situation. So that trigger menu option is bogus. Also bogus since the envelope trigger doesn't work anyway, see Envelope Follower section.
(5) The soft clipping only works intermittently and randomly when playing polyphonically. Many notes and chords played come through with no soft clipping. Soft clipping is acting like a monophonic function not useable in polyphonic patches.
(6) Can't sync env to an osc if the osc pulse width knob is in the zero position. The position of that knob should have nothing to do with syncing an env.
(7) The mod in nut on VCA2 gives a display in ms. Fortunately the numbers aren't actually given in ms which is meaningless for a mod input, it's just an improper ms tacked on after the number.
(Poor Design Issues) Attack and decay times can go to 38415 ms (multiplied by seven remember), that's 4.48 minutes, far more than anybody needs and beyond the 10 seconds of the original Mg. But rather than keep things simple Arturia made the release times go to 70000 ms (x7) for 8.17 minutes. Because of this lack of logic and the exponential workings of the knobs, no longer can release times be set to match decay times fast and easily by just putting them at the same knob positions, instead the popup numeric displays keep needing to be looked at, aggravated even more by the fact they often display settings for the wrong knob. There's a right way and a wrong way to do things, so why choose the wrong way? Seems to be the way companies do things these days.
(Warnings) LFO modulation and pwm inputs don't work correctly in polyphonic patches. I have confirmed that the LFO modules are monophonic accept for the env delay and fade portion. Thanks for telling us Arturia. Since this seems intentional, it can't really be called a bug other than there is no way to know about this until a patch doesn't work correctly. Think you can pop an osc bank into LFO mode and use that instead? Nope, that turns monophonic too. You need to use the osc bank [NO] option for keyboard tracking to get polyphonic LFO usage. But that causes all 3 oscs to consume processing power even if only one is used if the manual is to be believed. And it seems that the sinewave oscillators reset on keypresses (gee, why don't they reset when hardsync'd ?) most of the time rather than run freely in that mode, yet not always, depending on polyphony used. There is no way to guarantee reliable polyphonic LFO usage in MMV.
(Poor Design Issues) If the PWM knob is set to the zero position where you would normally keep it, then 3 out of the 5 LFO waveforms don't output what the panel graphics indicate. That knob needs to set to the rightmost position for normal operation. Why does Arturia do backwards stuff like this?
(1) Lowpass filter module is unreliable as a sinewave sound source since it can't seem to stay in tune, can all of a sudden change several semitones in pitch, and/or change to a different pitch when switched between monophonic and polyphonic modes. MMV is a digital synth, not real analog, and shouldn't be doing this.
(2) On the multimode bandpass filter (if I remember correctly), depending on the patch and the phases of the moon, turning the resonance knob to zero can cause MMV to stop producing sound. Probably do to floating point numbers going out of bounds and trapping out.
(1) Knob settings don't get saved with the patch. Unless a person wants to write down knob settings and manually dial them in anytime a patch is loaded, then this sophisticated module is uselss for anything beyond the 5 built in vowel sounds. This needs to be fixed!
Sample & Hold
(1) S/H module only works intermittently and randomly when used in a polyphonic patch. It doesn't even work properly with a monophonic input running through it such as a modwheel voltage in a polyphonic patch. It just appears totally absent when some notes are played. I confirmed this is a monophonic module and thus will not function properly in a polyphonic patch. This is totally screwed. Arturia seems to keep slipping monophonic functions into the MMV polyphonic section without telling the customer who thinks he has bought a polyphonic synth and then has to deal with patches screwing up because of it. This module needs to be polyphonic and fixed.
(1) The trigger function doesn't work with the exception of one rare time when running external audio through it and only when a VCA envelope was in execution. I have since tried several more times to get it to work from external audio and it won't, and I have never gotten it to work when running signals from within the synth through it. This needs to be fixed!
Keyboard Follow Section
(1) The knobs have missing gabs in range that can't be gotten to even with the mouse/right button fine adjustment. For instance, settings jump from 0.79 to 1.0 to 1.2 etc, with nothing inbetween. What about 0.80 to 0.99 for for example? In other words, you can't have keyboard tracking of 80% to 99%, 101% to 119%, 180% to 199%, and so on. Ridiculous. Negative keyboard tracking isn't supported here either.
(2) The K-Shift(what the manual calls Pivot) and Threshold parameters are switched. If you change the K-Shift parameter, you'll change the Threshold. And if you change the Threshold parameter, you'll change the K-Shift/Pivot. This section is confusing and is overkill anyway. I see virtually no use for a threshold parameter and even the pivot doesn't really buy you anything except tell you where the zero point is (that could be fixed permanently to middle C), and the fact that the Arturia programmers couldn't get it straight, and 4 revisions of MMV and 3.5 years later nobody has noticed, just shows how little understood and unused this is. It's still totally baffling how this bug has survided this long and not been noticed though. Maybe the fact is not one single person has ever used these controls in MMV's history.
(3) There are 4 low/hi keyboard trigger ranges that can be set, but all trigger menus in all modules only allow a choice to use the first 3. There appears to be no way to use the 4th one. How friggen hard is it to include the 4th one on all the menus? Didn't they know there were 4 of them when programming the menus? Do I really need to tell Arturia how to do their jobs?
(1) The soft clipping only works intermittently and randomly when playing polyphonically. Many notes and chords played come through with no soft clipping. Soft clipping appears to be another monophonic function not useable in polyphonic patches. Same soft clip problem as in the VCA envelopes.
(1) The delay time knobs go to 5000 ms of delay, but the module is only capable of 3000 ms of delay. The last 2000 ms of knob movement does nothing. (I've gone on many diatribes about the questionable intelligence of computer programmers, but how the hell is it humanly possible to write a delay module and not know about this? If 3 seconds of delay memory is allocated then give the knobs a 3 second range. Or if the knob has a range of 5 seconds then allocate 5 seconds of delay memory. It's really not hard to get the two to match. Do I really need to point this out? Sometimes it almost seems like a company wants to look totally incompetent to their customers.)
(1) The resonance knob has very little affect when the phaser is set to 6 stages. Depending on input material it can sometimes seem to do nothing. Seems to work correctly only in 12 stage mode.
(2) When used in conjuntion with the Delay module, the right VCA2 phaser input doesn't work correctly. It pans the sound to the right rather than keeping it centered as it does with the left VCA1 input.
(1) Same bug as phaser bug number 2.
(2) The rate knob also changes the amount. Thanks to Gordon Reid of SOS magazine for pointing out this ridiculous behavior. The amount knob is supposed to change amount, and the rate knob should change rate. So why then does the rate knob also change the amount?
(Poor Design Issues) The Chorus can be switched to type 1,2, or 3. The manual being of no help only says these are simple, medium, complex. What's really happing is that 1,2,3 refer to the number of chorus units. But when selecting more than 1, the others run at completely different rates. For instance, if you select 2, the second chorus seems to be running at maybe 2.5 times slower than the first. No wonder it's hard to impossible to get full thick chorus sounds. If you get chorus unit 1 to sound right, 2 is too slow and just adds an annoying swirling sound. Set unit 2 to where it should be and 1 is now way to fast and sounding out of tune. It just seems impossible to get a full sounding chorus sound without hearing all kinds of swirling and sweeping. The way to do it is either, have the other units run at slightly below or above the rate of the first, or have them run at the same rate but out of phase, like 0, 120, and 240 degrees apart for a continual thinkening effect like some old string synths. Add to that, that the stereo width has its own rate which seems to have the affect of panning the sound left and right rather than just stereofying it. I'd really like a decent chorus unit in MMV.
(1) In the little bit I tried to use the sequencer I uncovered many bugs and can't remember the specifics them all, other than so many of the trigger menu selections don't work and I've gotten strange sequencer behavior like it skipping from the first step to the last without running through the sequence and then getting stuck there. My main use for the sequencer was simply to hit a key on a keyboard and have it run a sequence once then stop. So simple a task, and going by the sequencer trigger menu options should be easily possible, but the sequencer ignored triggers from its own steps and envelope end triggers, so that simple task couldn't be done. My advice to Arturia is to actually try thiese menu items out to see if they actually work before releasing this pile of bugs on the public.
(Poor Design Issues) The sequencer knobs have 12 steps of resolution to both the left and right of center to allow 12 semitones up or down from the current key (far too small a range for many things, I don't understand this limit). Whether controlling osc or filter freq there are only these 12 +/- steps. So the dial displays 12 numbers left and right, correct? Of course not, because that would be common sense and make it easy to dial in the notes you want. Arturia has no common sense and the dials instead are divided into a total of 100 steps left and 100 right ranging from -1.0 to 1.0. So if you want a note to sound 7 semitones up, dial in something like 0.58 abouts and you have it. Be careful because a 0.4 might get you 0 semitones or 1 because 12 integers don't go nicely into 1.0. What the hell were they thinking. I guess I need to make up a conversion chart.
(Poor Design Issues) There is a row of numbers for setting the next step number for each seq step. Click on the number and move between 10 settings with the mouse and a redundant knob underneath also moves. There are 8 of these redundant knobs with the sole purpose of doing what can already be done another way. Why? What a waste of panel real estate. Why didn't they move those knobs up and give us a 4 rows of control voltages and a 32 step sequencer instead? I just don't get what they were thinking when designing this sequencer.
External Audio Inputs
(1) Unpredictable audio quality can range anywhere from sounding like horrible metalic noise, to clean audio with intermittent noise sometimes caused by nothing, sometimes caused by moving knobs or cables. There is no way to rely on external audio working correctly on any given day or hour.
--------------- General MMV2 Synth Bugs ---------------
(1) Polyphonic patches plagued with clicking/popping noises on their attacks. This is driving me insane. Because the problem mostly goes away in monophonic mode I spent time trying to figure out what's going on. Here it is. In mono mode the processing power usage indicator goes up. This means MMV is continuously running even when a note isn't being played, so all parameters are where they should be, and when a new note is played, the same voice is used. In poly mode, when the VCA release time goes to zero, the voice deactivates, and module settings are left hanging in whatever state they are in at that point. Then when a new voice is needed for a new note, one of the those previously used voices is used which is still set to whatever state it was left in when last used, and that junk appears at the start of the new note played. Arturia is starting the envelope and sound of the new note before junk from the last voice is cleaned up. That's not the way to do things. Clean up first, then trigger the new note, geez. Stuff like this really gets me angry, because it's so obvious it shouldn't be programmed this way and it's the kind of problem that can really make a synth useless. This is one of the most major bugs in the general working of MMV.
(2) Midi Controllers assigned to knobs aren't saved with the patch. This is totally screwed. If for example you assigned a foot pedal to control filter freq in one patch, then at some later time assign a foot pedal to control lfo modulation in another patch, it will no longer be controlling filter freq in your earlier patch, but lfo modulation instead. I guess it's now back to documenting midi patch assignments on paper and attaching them when a patch is loaded and removing them before closing the patch. Unacceptable. Arturia needs to understand that how a patch is controlled changes depending on the patch. This is totally baffling to me.
(3) Grabbing a knob and turning it often displays a parameter box for the wrong knob. Example, start turning the Sustain knob and up pops the display for LFO Delay instead. Then you have to release the knob and grab it a second time to get the correct display. With all the patching I'm doing, this is getting really old. In fact, MMV sloppily displays two parameter boxes many times, sometimes the second appearing ontop of the first or sometimes it appears in a completely different area of the synth, appearing at times if you hold the mouse over a knob for a certain amount of time, or other times if you move the mouse over the knob. The Phaser module is the worst at this kind of stuff.
(4) I have patches that bomb out when played too fast, or if played too high on the kybd, or if a cable is connected from one place to another, or if a knob is turned to a certain point. These can cause anything from MMV to stop producing audio, to displaying an Windows error dialogue box that can only be seen after closing the MMV window. Another person reported a bug with the S/H causing some kind of floating point overflow error. I believe all of these flakey bugs that cause MMV to stop working based on how it's played or patched are all probably do to values going out of range and MMV not able to recover until it's launched again. And some of these bugs will even cause you to have to exit a second time run it again to finally get it working. Sloppiness.
(5) The manual is a poor english translation, is vague and lacks substance. Want to know what the sawtooth and triangle waveforms change to when the PWM knob is moved? Want to know what the different Chorus Type settings actually do? Want to know if the LFO is monophonic or polyphonic? Want to know whether the effects are arranged in parallel or serial and in what order? The manual won't tell you. The manual says you can click on a cable to select it and press the delete key to delete it. Wrong, the manual also lies. Want to see a plot of all of the multimode filter types? The manual displays only 6 out of the 7, excluding the bandpass filter. Why leave out the bandpass filter? Because the manual is sloppy and incomplete.
(6) When installing version 2.0 I was asked if I wanted to install the version 1.2 VST. Is the VST really that old version or is this just another bug?
--------------- Other MMV2 Synth Problems and Limitations ---------------
(1) A single knob setting can suck all the power out of MMV. I loaded two patches for a bi-timbral sound. Just playing 3 notes caused it to run out of processing power and make awful sounds. The pwm was too fast on the lower notes, so in each patch I simply moved an osc bank freq knob a few octaves down lowering the rate from about 5Hz to 1Hz (just 4Hz). The osc bank was set to LFO mode so shouldn't have been consuming much power to begin with. Just that one adjustment allowed me to play 8 notes of polyphony with room to spare (could probably have gotten 10 notes). That means polyphony was cut down by about a factor of 4 just because of the setting of one knob. Something is wrong here. In fact, for a 2GHz computer, sometimes I'm getting a lot less polyphony than I'd expect, and some of it seems to be related to issues like the above, where just a single knob setting or cable patched from one place to another can drag MMV way down. I'm running the stand alone version. I can't conceive of running MMV as a VST with a DAW or with other plug-ins running. Software companies need to spend time optimizing their software for speed and efficiency because their own wastefulness could doom their own products.
(2) When connecting MIDI to control a knob on the MMV, there is no way to set the scale/range. Rarely does anybody want a foot pedal, mod wheel, ribbon cable, or breath controller to control a knob from 0 to 100%. Nor does MMV smooth MIDI to eliminate obvious stepping. These problems in addition to the assignments not being saved with the patch makes this whole capability virtually useless for adding realtime expression while playing from any controllers beyond the velocity/aftertouch/modwheel jacks, and of limited use when trying to control MMV from a sequencer. And aftertouch on most keyboards is neither polyphonic nor useable/controllable, so velocity and modwheel smoothed through an envelope follower are about the only controls you'll be able to use for realtime performance expression (remember the sustain pedal is unreliable in MMV). Extremely poor for such a powerful synth.
(3) I can see no reason for some of the limits placed on what modules can be selected for the second row on MMV. If I can select the formant filter for a 4th filter module, I should be able to instead select a 4th lowpass filter. Only one ring modulator is allowed, yet something as seldom used as an envelope follower (especially when audio input doesn't work well, nor does the env fol trigger work) allows four to be selected.
(4) Arturia claim to be emulating the Mg modular. But the MMV fixed filter bank isn't a fixed filter bank, it's an equalizer. This is very deceptive. They should provide a fixed filter bank like the Mg modular and a person could select which one they wanted like the phaser/chorus option. The formant filter isn't really a formant filter either, but a parametric EQ.
(5) The Glide is what I guess would be refered to as constant speed, as it doesn't change speed based on interval distance. There is no way to select constant time. I don't know if the Mg modular is like this, my MultiMg isn't. Constant time seems better. With MMV's constant speed, if you hold down a large chord, all the notes won't slide into their destinations at the same time, and is very sloppy sounding. And setting a comfortable glide for closely spaced notes will be too much for farther spaced notes. Setting a comfortable glide for far spaced notes may not even be heard at all on closely spaced notes. In addition, the glide is completely linear in its speed, so when it glides into a note it comes to an abrubt and very unnatural sounding stop. Would be nice if there was a selection.
--------------- Closing Comments ---------------
All the bugs I've come across so far are things that I have stumbled across while programming patches in MMV2 stand alone PC version. There could be still more that I haven't run across yet or ever. I've only listed bugs that are Arturia's own doing and should show up on everybody elses MMV2 running on their machines. I've left out anything that might be specific to a particular computer. As it stands now, MMV2 is quite a disaster.
When I first got MMV2 I reported the goofy envelope shapes and settings being off by a factor of 7 to Arturia. Their response was the envelopes on their Mg modular were shaped like MMV and also were off by a factor of 7. By talking to others I have found the first to not be true, and the second I find hard to believe based on the fact that Mg envelopes are about the only ones on the planet to so carefully show time settings around their dials rather than generic numbers of 0 to 10. I believe I was given false answers just so they neither have to fix the problems nor admit they exist. There's no reason to emulate idiotic behavior in a synth anyway even if the Mg acted this way. MMV envelopes aren't calibrated nor have the same ranges as the Mg nor follow the panel graphics around the knob, so their claim of Mg values being off by 7 makes no sense and is irrelevant in MMV.
Arturia has forums on their website. I have seen somebody who claims to work for Arturia appear only in certain sections, but never in any of the technical problems sections. No question asked about bugs or problems will be answered by them, nor does it appear Arturia has any interest in reading them. This total lack of interest in wanting to find out what might be wrong with their own synths nor willingness to correct the problems just isn't right. Instead they just keep packaging the same bug infested code into new products and are leaving a messy trail behind with a bunch of pissed off customers. I'm getting tired of buying software that doesn't work and this is where I will fight back. These damn software companies put 99% time and effort to create a software package, but don't want to put in the last 1% to complete it.
MMV has been through versions 1.0, 1.2, 1.5, and 2.0, and has been on the market for 3.5 years, and yet it still has major bugs that were present since the beginning. They included 6 new modules in version 1.5, 4 of which don't work correctly. And this is where I have another problem. I'm one of last people to buy this synth, so why am I the first to be finding all these problems. All this stuff should have been reported to Arturia long ago. Best I can understand is they are being sold into market with very little ability and with simplistic uses. If the only envelope shapes a person uses are a percussive attack/decay or sustained organ type and nothing else, then the totally messed up envelope shapes go unnoticed. Never plug more than one mod source into an osc, then jacks that stop working go unnoticed. Never use a formant filter for anything beyond the 5 included vowel sounds, then parameters that don't get saved go unnoticed. I could repeat this for every damn bug I've found and you will see that it appears a potentially powerful synth like this is probably not being used for much more than simple boring SH-101 type bass sounds or ugly techno filter noises. If you want to hear a harpsichord that's so awful it sounds like an 1970's organ harpsichord preset, or guitars that sound nothing like guitars, or horrible drum sounds, and so on, those are included as presets in MMV2, and also display the total wasted abilities of what MMV2 is really capable of and also show the sad state of synthesis/music these days. Now you can buy those awful sounds in Arturia's new Analog Factory synth. Yet another product they've worked on before completing and fixing MMV2. And this is part of the reason these bugs don't get fixed (beyond programmers who don't try their own code to see if it works), is because of the level of the market they're being sold to, Arturia can get away with it. A modular synth if for doing sophisticated things you can't do on a MiniMg, but MMV sure isn't being used that way. Now I have to be the one to beta test their synth 3 years and 4 versions later. Ridiculous.
If Arturia fixes these damn bugs, I'll go around recommending it to everybody. But as of now MMV is a disaster and in no shape to be used for anything beyond the most simplistic uses. I'm getting tired of wasting money on software that doesn't work, so I'm going to steer anybody clear of buying bug-ridden Arturia products on all boards I'm on until they get off their butts and fix it.