Waldorf Info – Things to remember Part 1 Basics
WALDORF… Ideas & things to remember
how many DSPs are in the XT?
IIRC they are the same type used in Kyma, correct?
Nearly, the Kyma uses Motorola DSP56309 and we use DSP56303, the only difference is that the 56309 has more on-chip RAM than the 56303.Waldorf V2.0 Beta team in 1994 and coding was done by Andreas Busse and Stefan Stenzel. Stefan focussed much on the Wavetable loader then. The machine use to code V2.0 was named RESI ;-)
it was done on a SAM68K system, and that had an operating system called SUSY…
CORRECTION: The system you are talking of was called KDE, as far as I remember (not to confuse with the Linux Desktop Manager).
This was used to program the Microwave 1 originally but I’m not sure if it was also used for V2.0 back then.
Resi and Wegy were two MIPS workstations and Stefan logged onto Wegy through an X11 terminal with a beautiful b/w monitor with 1:1 orientation, as far as I know. (Err, maybe he logged onto Resi because that was Andys machine…).
The Wave, the 4-pole, the Pulse and the Microwave II were developed on this system, doing 68k compiles.
didn’t know that we could actually edit the wavetables. is this possible on the XT?
It is possible with Emagic SoundDiver, not at the front panel.(planned!! to be programmed on the box as well in a later OS)
Peter Berg <firstname.lastname@example.org writes:>
had the sounddiver this feature before or after the end of the last year ? The wavegen tool for atari (68020), linux and riscbsd can do exactly the same and was written by me in the end of last year. I included even such a drumloop wavetable into the zip archive.Surprisingly a few weeks later someone made a drumloop wavetable with the sounddiver and got celebrated for this …
Moogulator can’t comment on this…!! Alternative for PC users: Philip Pilgrims stuff (see Soundprogramming)
how many voices can the microQ do and what about the DSP-power??.. go HERE!
here’s something about the Wavetables and where they come from..
>>R01 – R28 non-algorithmic (with real waves)
>>R29 – R32 algorithmic
>>R33 – R6x (All sync, pwm, fmm, k+s, fuzz and speech wavetables -> algorithmic
>>Chorus, Piano non-algorithmic
>Please explain. I thought these were algorithmically created, but stored in
The Chorus wavetable was an analysis I did for the Wave. If you load the
Wave factory set, you will find that there is a RAM wavetable containing my
original analysis (which sounds much better than Stefan’s ROM version where
he had to reduce the number of waves; the Microwave version is even worse,
The Piano wavetable is a sample from Stefan, I think he got it from his old
Yamaha TX16W but I’m not sure.
The sync, PWM, FM and Bell wavetables starting from R33 are created by
Stefan using the UPAW format.
The Karplus Strong wavetables where developed by Stefan and one of the
former Wave project manager, Claudius Bruese, as far as I know.
Wavetables 31 and 32 were created by Andreas Busse, the former development
manager of Waldorf.
Wavetables 01 to 30 (finally) were made by our boss Wolfgang Dueren and by
the original PPG Wavecomputer / Wave developer Wolfgang Palm.
Nice story: wavetable R14 wasn’t planned to sound like it does, it was only
an error of the computer they used to compute wavetables. They had to put
in the levels of each harmonic but without any normalize function. It was
one of Wolfgang Dueren’s first wavetables and he didn’t know that so he put
in all levels as he thought… the result was this awful clipping shit
which then became the most beloved and famous PPG wavetable ever. Nice, huh?
Another nice story: wavetable 28, the original choir wavetable, was also
done by Wolfgang Dueren. He had a book with lots of harmonic spectra and he
just copied the values written there. He had no idea if the wavetable would
sound similar to a sung voice at all. And this wavetable also became famous
and was used several 1000 times in these typical „PPG choir“ sounds :-)
Good listening examples: the albums „Construction Time Again“ and „Some
Great Reward“ by Depeche Mode and „West End Girls“ by Pet Shop Boys. Also
Was it what you asked for?
>>Was it what you asked for?
>No–again, I KNOW that the wavetables were created using different means,
>but the question was about a bug you found that affected certain wavetables
>differently. Forget about it for now, we’ll move on to other things.
Okay, we’ll do. But yes, there is a difference. The algorithmic ROM
wavetables are stored within the OS or so but the usual wavetables have to
be loaded separately into the E2ROM by the production department. So, there
is already a difference in how they are handled.
Now Stefan corrects all details
> The Chorus wavetable was an analysis I did for the Wave. If you load the
> Wave factory set, you will find that there is a RAM wavetable containing my
> original analysis (which sounds much better than Stefan’s ROM version where
> he had to reduce the number of waves; the Microwave version is even worse,
They are both the same and still the most waves consuming ROM tables in both
MW and Wave.
> The Piano wavetable is a sample from Stefan, I think he got it from his old
> Yamaha TX16W but I’m not sure.
> The sync, PWM, FM and Bell wavetables starting from R33 are created by
> Stefan using the UPAW format.
There are no ROM Wavetables using UPAW, FM and Vibes are usual wavetables,
Syncs and PWM are all done algorithmically.
> The Karplus Strong wavetables where developed by Stefan and one of the
> former Wave project manager, Claudius Bruese, as far as I know.
Only by me, but it was Claudius insisted on keeping them, I personally find
> Wavetables 31 and 32 were created by Andreas Busse, the former development
> manager of Waldorf.
> Wavetables 01 to 30 (finally) were made by our boss Wolfgang Dueren and by
> the original PPG Wavecomputer / Wave developer Wolfgang Palm.
> Nice story: wavetable R14 wasn’t planned to sound like it does, it was only
> an error of the computer they used to compute wavetables. They had to put
All waves done by the two Wolfgangs were calculated with a slight error in the
sine function used, I think the error produced those artifacts that led them to
the decision to use an analog lowpass filter, which they first thought could be
avoided with the wavetable technology.
Xmorph in the Q: Right now, it’s implemented as a pure scaling between the numeric values of
each variable in a patch. But, for things like waveform settings and filter
types, this is a small range, and the results will be „different“ – there
is currently no way to smoothly morph between a Comb and a Low Pass filter.
At some point in the transition, the filter type will simply do a hard
switch from one to another.
It might not be possible if Waldorf is doing filter modeling like they are
on the OSCs. LFO waveforms are also in this category.
in short: all CCs (Controllers) morph, „switch“ type parameters will switch over, not smoothly morphed (as on NordLead) filters etc are switched over as they were parameters.. wish: storeable „somewhere between“ a patch that is somewhere between the 2 sounds you are currently morphing between.. and: modulation matrix things are not morphed!!
switching the Mw 2 / XT off (use its power switch!) saves global parameters: if you switch it of by a main power switch (not the soft button!!) it will not remember the last sounds & some settings..
Hans Heerooms <email@example.com> wrote some good stuff for those who don´t know exactly about samples and wavetables (XT, MW (1&2), PPG) etc..
a wavetable synth is not a sampler, no even a 8 bit one, however the way we talk about ‚waves‘ and wavetables on the list makes it quite confusing.In fact a MW is more related to additive synthesis than to a sampler. Ok ,this is going to be a bit technical , but I think it will help to understand why the WAV converting is a tricky thing :Take a static sound : that is a sound with no variation in timbre or amplitude (let say : a very short perfectly looped sample). There are two ways to describe such a sound:
– The sampling way : for a lot of timesteps (=sampling frequency) take as good as possible (=sampleresolution) the amplitude of this sound.
– The additive way : use mathematics to describe the sound as a sum of basic sinewaves. More sinewaves will result in a better result, but often just 10
or 20 will do.
There is a mathematic link between this two methods, it is called fourier transforms.
Back to the waldorf synths : they use 8 bit waves with ‚only‘ 128 steps (actually only 64, the other are just copies of the first ones). The reason that the synths don’t sound 8 bittish is that by clever and secret DSP ing waves are fourier transformed in a sum of 64 sinewaves: The waves are NOT going to a normal DA converter but enter in some magical math machine. To convert a sample to a couple of waves for a wavetable the following must be done :
– Divide the sample in the number of choosen waves
– For every bit do a frequency analysis (with a lot of averaging over the
length of the samplebit !) . From the frequencyspectrum create a wave.
– Combine all the found waves in a wavetable.
What does this mean for ‚real life‘ converting :
– Samples can’t be to long , a wavetable has only 61 free waves. Maximum
time is (depending on sound) 5 seconds
– Inharmonic sounds confuse the frequencyanalysis. Inharmonic sounds are
things like the bowing sound on a string, a crash cymbal, an open hihat and
… much reverb (or other ‚cool‘ effects ) in a sample.
Conclusion : A WAV must be short and simple. In fact the drumloop I used was a rather dry recorded acoustic kit with a clear kick, a snaredrum with not all to clear snare, and low level hats. To be honest : getting such a cool loop was a bit of luck…
Strangely enough the sampling resolution of the wav file is not that important. Maybe treating a sample by a lowpass (non resonant !) filter can help.
For some samples the next works nice :
– Sample with as much waves as you can.
– Assign wavepostition to modwheel (full scan for full turn of the wheel)
– Play a note and rock the modwheel. Probably you will hear some unwanted
transitions in your sound. Now edit the wavetable : replace at appropriate
points (sorry : it is trial and error) a wave for a interpolated wave (-1 in
Sounddiver). By repeating this you can make the wavetable cleaner..
about the Envelopes Attackphase =0 (Clicks!)
Clicking (but not Amp Env to 0)
>I know that an instant, high level signal causes a click. Does the instant *removal* of a high level signal cause a click?
>Actually, this happens in the same patch that has the „dual-mode“ problem.
>I can hear a click on each note on. I thought it may be because the wavestart was set to „free“ but no. The patch is MONO BTW. What I am thinking is that it is clicking on the note off and not the note on. (Since it’s mono, and the signal abruptly changes) Actually, I’ll just switch it to POLY and set my release right and see if it goes away.
This is a known bug. It is on our bug list and will be fixed as soon as possible.
This is a past often discussed topic. The reason you get a click is that Waldorf is being true to what you should get.
ZERO MEANS ZERO. Thus, if you have an instantaneous change in the envelope, the output actually starts at a different point than zero, which causes an abrupt change in the output — this means that a spike of harmonics will appear in the output (on a spectrum analyzer, this HUGE change in ZERO time contains a HUGE amount of harmonics) — hence the click. At ONE, things start to calm down because, although the slope is extremely sharp, it is still non-zero time to change the output.
The Microwave II and XT have this „problem“ too … and I presume that all the Waldorf products will. They are being mathematically precise in what they’re doing — it may not make sense to the average guy, but it makes controlling and knowing what to expect easier, once you understand their thinking. If you ask for 100% you get 100%. If you ask for 0 you get 0, not some estimation of 0. :)
and: of course this is a feature like the click parameter in Matrix6/1000…
this is another FAQ about clicking vs. Attack Time.. (a FEATURE)
>I’ve noticed that the click that occurs with the attack at 0 only happens on some sounds; does this have to do with the phase the waves start in? I also seem to recall having made patches that didn’t click at first, until I screwed with something or other in the tuning or mod routings…
I’ve also had problems „de-clicking“ agressive bass sounds where the amount that I have to increase the attack to remove the click noticeably softens the sound… this mayt be unavoiable, but are there is thera anyway to predict what tuning, mod routings, trigger setting or whatever will affect the presence of a click?
>On a related note, if this problem happens on waldorf because the envelopes are so short, why doesn’t it happen on older synths when you bypass the envelope (set vca to gate, sustain to full and attack to 0, for instance) on older synths?
And why *do* you hear a click when you power down your home stereo while a CD is playing?
Think of a hollow waveform, e.g., sine. It is very smooth (in fact, the smoothest one) and doesn’t produce any harmonics beside the fundamental one. Now cut the first half of the upward cycle. What you get is an amplitude change from 0 to fully positive in exactly 0ms while the rest of the wave is again completely hollow. This 0ms change in the waveform produces a sound identical to when you hear a light switch in your audio system: A very short impulse. And physics says: a short impulse contains all frequencies in equal amplitudes. The result: a bright sharp click. When this impulse has a longer attack, it gets hollower and hollower until it is covered by the wave itself.
Now, the Waldorf Pulse: attack 0 means 1.9ms. 1.9ms is equal to around 523 Hz with this simple calculation: 1 divided by seconds = Hz So: 1 / 0.0019 is around 523 Hz.
This in turn means: you can hear a click in the attack phase that has a maximum frequency of 523Hz which is already easily noticable.
When you now turn the amp attack rate to 1, you have 2 x 1.9ms = 3.8ms = 261Hz. Ah, that’s hollower. But still too bright when you play a low and hollow bass sound. When you e.g. play a hollow bass sound with around 80Hz base frequency, you have to make sure that the attack envelope is no shorter than 12.5ms to prevent *any* click. And this is an amp attack value of around 6 on the Pulse!!
>>Anyone own an XT with Q-knobs?
>surely if its ‚just‘ a knob change Waldorf could supply the necessary retro kit for people who bought the original MWxt’s many moons ago.
No, the metal housing is different, too. The holes are smaller because the knobs cover the them now. No crossgrade Q-Knobs <-> old XT Knobs (the Q Knobs are much smaller !)
some about the Q by Eric Young::
Hello again, So I gathered this info on the order of the OS upgrades. This is an excerpt from Wolfram:
>updates will come in this order: global edit, multi edit, step sequencer,controller send/receive.
Heh, lets hope they don’t stop there! (on to the review…) Evolving Sounds: Well, I spent about 2 hours trying to get the classic Microwave2 ‚evolving sound‘, you know the ones that just float there in front of you, but they’re continualy revolving, glistening and so on.The reason they’re so easy on the wavetable synths, is becuase the tone
source is itself a sample which goes through spectral changes over time. The Q’s tones will not, unless you tell them to. They key to this is routing. Plan on using every slot within the 16 slot matrix to arrive at these kind of complex sounds. In fact a lot of times you’ll be using up slots just to compensate for other routings which have a undesirable effect at some point in time or on the range of the keyboard. For example, I made a patch which sounded *great* with a lot of resonance
late, but it sounded horrid early. So, first I turned the RES down, then I assigned ENV3 to F1(or F2)RES and gave it a very high ‚Amount‘ value. ENV was set to a slow attack and decay. This way, the RES will come in later and spit up all those shimmery lasers.
The way to add motion into a sound is to have lots of contrasting sources, but remember to keep them subtle! With such a powerful matrix, you’ll run out of LFOs in a hurry. But, as I mentioned before, make good use of the negative and positive mod Amount values. For example, lets say you want the sound to swim in panning. LFO1 for F1(Filter1) Pan, and LFO2 for F2 pan? Maybe, or you could have it set: LFO to F1 Pan with a positive mod Amount, and LFO1 to F2 Pan with a negative Amount. This will get the sound to pan swimmingly and only use 1 LFO.
In addition, the ENVs can have negative or positive amounts. I just wish they’d allow a ‚cycle‘ function. For example, attack to decay, then right before reaching sustain, just use the same attack and decay params in finitum. It could be a ‚ENV MODE‘ kinda thing…set it to ’normal‘ or ‚cycle‘. Now this would be great…but since its not there we gotsta work around it. But you know, this ‚cycle ENV‘ essentialy becomes a kind of LFO, just without ‚even‘ cycle speeds (for example a slow attck with a short decay etc). So, back to our ecample, what we’d do then is use ENV3 to give the slow attack, and use an LFO (set at maximum delay time) to create the clycling effect.
Are 16 slots not enough? Heh, get this: each section of the Q actualy has its own little ‚hard-wired matrix‘.
LFOs: allows keytracking, Fade and Delay
OSCs: allows Keytracking, FM Amount and Source as well as PWM Amount and
MIXER: allows balance for each of the following, OSC1-3 noise and
FILTER: allows FM Mod Amount and Source, Cutoff Mod Amount and Source,
Drive Amount and Keytracking, Pan Amount and Source, ENV Amount and ENV
AMP: Amp Mod Amount and Source, and VEL.
I’m sure theres a few I missed too, but you get the idea. Combining all these possibilties is the key to Q’s heart.
This is really cool: you can assign FM Source for any OSC to be any other OSC…So, I set OSC1 FM modified by OSC2, OSC2 FM modified by OSC3, and OSC3 FM modofied by OSC1! What does that sound like? Well, its very….ahh overtony. But heres the cool part, the above routings were not done in the matrix, they were done in the ‚hard-wired matrix‘ I mentioned above. Now we can go into the matrix and assign something like LFO1 to either OSC1 Pitch or FM Amount. Then lets get another slot and assign ENV3 to LFO1, this will make the LFO start slow and get faster then slower again (of course dependant on how you set ENV3). Get the idea? It goes on and on, and like all the best things in life, the more you put into it, the more it gives you back!
LFO Note: remember how I was complaining about not having longer delay times for the LFOs? I thought I may be able to assign an ENV to LFO Delay…but no dice. LFO Delay does not appear in the matrix, only in the hard-wired matrix (Edit buttons).
ENV Rounting Oddity: you can assign, in the matrix, something like ENV3 to AMP Attack….or even ENV3 to ENV3 Decay! Huh? I tried it with all different kinds of Amount settings, but it seemed to have *no effect*…strange. Anyone know about this routing oddity?
Levels: As I mentioned before part of what makes the Q so fat is its 5 sound sources: 3 OSC, Noise, and RingMod. The way to really make a sound evlove is by using their levels creatively…allow them to *individualy* come in and out. By using both LFOs and ENVS, you can have each of those sound sources have a life of its own. Assign either an LFO or an ENV to OSC1 Level, and give it a positive or negative number (remember you’ll want to fine tune this boost or cut, by tweeking the OSC’s master Mix Level, least you set a negative LFO value to an already low Master Mix Level, or vice versa). Go through each of the sound sources you want to have in your patch and set them to *not* correspond time-wise with the others. OK, thats not nearly enough, we’ll need some FM mod on a few OSC pitches, filters which come and go
on their own accord (assign an LFO to a filter’s cutoff, Res or whatever) and of course a generous helping of some Pan Mod. The result should be quite….charming.
Ok so thats prety much enough about Evolving sounds…bottom line is, they’re there, but you gotsta coax them out. But wait! All the stuff I’ve been talking about is using the various routing possiblities. That means you press the key and it does what you’ve told it to. Q can do more…send/recieve controller data (implemented soon). That means you can manualy tweeek knobs, and so on, and have them recorded to MIDI. Imagine going into ‚overdub‘ mode on your sequencer and making a recording pass * for each knob!* Bam.
Effects: Ok besides the effect parameters not showing up in the matrix there’s something else really annoying about their setup, but its possible that it will be cleared up in a future OS. I hope! It has to do with using the Q multi-timbraly with its 2 effects per 4 sound Intru set. There are 4 Instu sets each with 4 sounds= 16 part multi timbral, and 8 effects. I emailed Wolfram with this lengthy question, and as soon as I have an answer, I’ll let ya know:
…this is *really* important! I’m hoping like crazy that this is something which can/will be implemented in a future OS, and not limited by the hardware setup. OK, so we have 4 Instr. buttons, each contains 4 ‚Sounds‘. Pressing the ‚Shift‘ button acesses other ‚Sounds‘ within the Instr set. Each Instr set can have 2 effects *which are mandated by the 1st ‚Sound‘ in the set*. The other 3 ‚Sounds‘ belonging to a particular set must share the same 2 effects as perscrribed by the 1st ‚Sound‘ within the set. Good enough. Heres the problem: they also share the *Effect Mix* parameters as well! This is unfortunate. Sharing the 2 effects is fine, but please allow us to set the ‚Mix‘ levels *indepedantly* for each ‚Sound‘ within an Instru set!
For example, lets say I have 4 ‚Sounds‘ in Instru#1, I have the 2 effects set to chorus and delay. I want the 1st sound to be completly dry…both effects mixes set to 0. Sound 2, Mix set to 50 and 50 for each effect. Sound 3, just chorus…no delay. And lastly, sound 4 with both effect mixes set to full (127). It seems to me this should be possible. If we think of the limits of the hardware, we have 2 effects processors. We can send *anything* into it from a mixer. The effects parameters can remain constant, but the mix levels of the input can change. I know I’m being extreamly verbous, sorry! Simply put: As it is, effect 1 and 2 parameters *as well as Mix Level* are determined by the settings of the 1st Sound-out of 4- within any of the 4 Intru sets. But, shouldn’t it be possible to setup the OS in a way that *will* allow us to set at least *Effects Mix Levels* independantly for each Sound within an Intru. set?
Well, see…it seems easy enough to me. But it really would open the Q up a lot more. Otherwise you have to really figure out how you want to arrange each Instru block. For example, if you simply must have your newly created „Bubbles!“ patch come in with crazy amounts of delay and feedback, you’ll have to leave the other 3 spots within the Intru block empty (unless of course you have other patches you want that ridiculous amount of delay on!!). If Waldorf can do it, great, if not…no biggie.
I should have some Q MP3 demos up on Mark’s Midiwall within the week with some examples of what I’ve been talking about.
LFO Ocillation: For a moment I was afraid that the Q’s LFOs didn’t self ocillate…but no worries, they do! Just to clarify, this means that the LFO cycles on its own, regardless of recieving a ‚key on‘ trigger. Of course, this is really important for creating complex pads. I know this is fairly standard in synths but I just wanted to make sure. Also, any of the 3 LFOs can be assigned to Sync to Midi. Ooo yeah, like that.
16 Part Multi-Timbral: Remember how Waldorf upped the Q anti by making it 16 part multi-timbral? Well, this move shows in its MIDI setup. It was originaly 4 part, so there are 4 Intr buttons. Well, when they made it 16 part, they just allowed those same buttons to access 3 additional sounds. Unfortunatly, they don’t correspond numericaly with 1-16
channels. Instead we have 1.1,1.2,1.3,1.4 then on the next Inst button: 2.1, 2.2, 2.3, 2.4 you get the idea. Well it works just the same but, hey which one is MIDI channel 10..or 13?! Check out the following table. The numbers in the body of the table are MIDI channels, on the left is how Q *displays* the multi-timbral info, where x= either Inst 1, 2, 3, or 4 and the extension corresponds to its 4 multi-timbral slots:
Inst1 Inst2 Inst3 Inst4
x.1 1 2 3 4
x.2 5 6 7 8
x.3 9 10 11 12
x.4 13 14 15 16
So when you want to work with the sound on MIDI channel 10, you’ll need to access Inst2, then hit ’shift‘ and push the Inst button with the corresponding extension, in this case Inst3. It will be displayed as: Inst 2.3 got it? No. Ok, one more. Midi channel 13…Ok so its in the last slot of the Inst1 set which makes its numerical display read: 1.4 So to access it you’d need to first press Inst1, then ’shift‘, then Inst4. Access to MIDI channel 3? Hit Inst3, then ‚Shift‘ then…Inst1.
It will be displayed as 3.1
I know its confusing, but as long as you have this table handy (not included in the preliminary manual!) you won’t have to count buttons each time you want to acces a MIDI channel. The reason its set up this way is because, as has been mentioned, all the ‚Sounds‘ within each Inst set share the same effects parameters and Mix. So, for example, MIDIchannels 2,6,10,and 14 will all have the same effect setup. This is important, as you’ll have to plan ahead of time where to place your sounds. It would be nice if the Q display would give you corresponding MIDI channels…but with this Handy Table (only 1 cent/download!, haha) its easy enough.
NOTE on ENVs: Lets say you assign, in the matrix, ENV3 to OSC1 Level with Amount set to +63. Then you turn off OSC1’s Mix Level (least you get *two* OSC1 Volumes). OK, now ENV should have total control of OSC1’s ENV, right? Well, no, its still got to go through AMP ENV! Using the previous example: we set ENV3 to a fast attck with fast decay, just a pluck of sound. However, if AMP ENV is set to a slow attack…we will hear nothing!! While this will not impress some (myself included), it can lead to very interesting effects. For example, go back to the OSC1 Mix Level knob and give it some juice. What this is doing is supplying AMP ENV with its own copy of OSC1. Set the AMP ENV to have a quick attck, medium decay, and medium release. Now set ENV3 to have a very quick attck attack, a very quick decay, no sustain, and no release. What will it sound like? Press the key down and you’ll hear OSC1 give a short pluck (ENV3 and AMP ENV) at the beginning, and carry on a bit (AMP ENV). With some further permutations to the ENV3 routing you may increase the possibilities.
Either way, just remember that *ENV3 and ENV4 do not have an AMP of their own and still must pass through the AMP ENV*. What does it mean? Well it means they are suitable for creating sounds and effects *within* the time frame which AMP ENV is set to.
ADADSR: just to clarify this envelope term, ADADSR means Attack-decay-attack-decay-sustain-release. As I’ve mentioned before, the Q’s ENVs are only ADSR. Well, here it is, use ENV3 or ENV4 to add the second level inside of it. Basicaly have AD be AMP ENV, then the second AD be ENV3 (or ENV4), and then SR is of course the AMP ENV again. For example, (and for simplicities sake I will refer to using only OSC1, but of course you’d probably want to add any of the other sound sources as well) You will need to assign in the matrix ENV3 to OSC1 Level at +63. Also, set OSC1’s Mix Level (remember this *is actualy AMP ENV*!!) to a very low setting of say 100–we don’t want it to drown out its simese
twin over on ENV3!–. Now have AMP ENV set to A:8 D:35 S:35 R:45 and ENV3 to A:110 D:35 S:0 (or very low) R:35. The result is not exactly drammatic but certainly noticeable. It’d be *much nicer* if ENV3 and ENV4 were actualy conected to a separate AMP (note: this AMP wouldn’t need its own envelope, it could simply just power the ENV3 and 4 signals and send them into the Q master mix)…But hey, theres enough to work with here at least until we discover life on Mars! By next week, I should have some examples of an ADADSR patch on the net<—Mark’s Midiwall.
I want to point out that this whole discussion is concerned with using only *amplitude* to create ADADSRs. Very easily you could create a sweet ADADSR sounding patch using ENV3 to filter1 and ENV4 to filter2. Then set one filter to climax early while the other filter takes more time to develop. ;}
there’s lot more to say about the Q.. have a look at MIDIWALL.com , too!!!!
Q: if all voices play and another key is triggered
>does it select the one that uses the lowest volume?
>(as in the Wave (?))
>so will it steal the voice from the one that is mostly „useless“ (?)..
It has a very complex voice stealing algorithm that should make it almost inaudible if voices are stolen. It looks for both volume, age and note number of the voices to steal and tries to find the most „unneccessary“ voice.
By the way, this site do show a early (?) microwave painting used on advertising papers.
Just have a look at
to see a micro wave synthesizer diy kit from 1975. (nothing to do – but interesting.. (name??)
I don’t know if Wolfgang knows that, I didn’t know that. But the name Microwave came from a very famous NA musician who bothered Wolfgang to build a PPG Wave 2.2 in a small 19″ housing, afak. It was Michael Boddiker.
No-one noticed the hint with „fame-ous“… Afaik, Michael Boddiker did the soundtrack to the movie „Fame“ besides his work for Michael Jackson.
Now, the *real* history:
I talked to Wolfgang about how the name and the machine Microwave came: Wolfgang got really *bothered* by lots of people asking for a PPG Wave as a rack mount unit. Michael was one of the guys almost screaming for a 19″ Wave. He said that he wanted a „Micro“-version of the Wave. After PPG was closed, Wolfgang found „Waldorf Electronics“ to basically produce hardware for Steinberg. A year later, Wolfgang finally decided to do a 19″ Wave and Mike Mercer, vice managing director of TSi at that time, came up with the name „Microwave“.
Pulsar & Waldorf…
MW and XT on Pulsar? / Technical stuff..
remember that virtual analog synths often make a difference here to save processing resources. I like doing FM with the frequency modulation input with the Pulsar and I’d be curious if it’s of the same quality as i.e. FM on the XT. Sounds good anyway.It should sound similar to the XT, but I have never checked this.
..exept the wavetable OSC in the modular it is not planned to release an XT on Pulsar but you can build something near the MW.. (no sync).. but you have a big advantage: you find the ROM wavetables of the MW2 in every WaveOSC that you open in the Modular/Modular2(!!!) in Pulsar.. I could open 20 OSCs on Pulsar I and each has its own Wavetable.. in the MW,MW2,XT you have to use one Wavetable for both OSCs.. here you can have 20 OSCs with different Wavetables!! cool eh? ok, no editable Waves or UPAWs.. hope that will come, as the Pulsar has the Computer with it by default ;-)
john s cooper <firstname.lastname@example.org>
>>Bell sounds radically different from its counterpart in my Microwave XT.
>>is this a bug?
can you confirm this Wolfram or Rembert? i haven’t scrutinized *all* the wavetables yet, but most of the ones i’ve tried have seemed correct. only the Xmas Bell had obvious problems.
>I found „clipper“ sounding strang at higher keys (loosing treble a bit).Other than that I haven’t noticed any differences.
can you tell me (email privately) if you hear a difference between the Xmas bell wavetable, when compared with your Microwave II/XT?
after realizing this, i tried something easier: Wolfram’s CS-80 patch (simple, but I *love* it). i was able to reproduce this almost exactly, but the Microwave XT’s filters sound *much* better (bigger, brighter, fuller) than the filter I’m using in the Pulsar.
Did you try the new „24 dB V“ filter that came with the latest drivers and synths? This one is good though I still like to see the MWII/XT filter ;-)
yeah, this was using the new filter. definitely better than the old filter, but still nowhere near as nice as the XT !
i’ve also noticed that the sound coming out of the Pulsar oscillators lacks the intense full low-end which we get from the XT. tend to sound very thin.
>>you can hear some stepping between envelope values (especially when modulating a waldorf oscillator start-wave).
That’s an oscillator thing. I noticed that also during the review of the unit
what review are you referring to? if it’s a written review, i’d like to read it!
>and I was told the Pulsar-Waldorf-Oscillator was built to simulate older PPG/Wave-Oscillators that used to have a „stepped-mode“. So it’s more like a feature. However, I like the way the MWII/XT does interpolation better.
wolfram, stefan, is that true? when modulating the startwave of the pulsar oscillator with an envelope, should i hear „steps“? is this a problem with the envelope or a feature of the oscillator?
>For amp- and filter-EGs I use the ADSR-Vintage-module. As it lets you fine
good suggestion – i’ll take a look at it!
>BTW: I wouldn’t mind if you put that Pulsar-Microwave somewhere for download.
definitely, that’s the plan – i’ll be uploading it to the „patches/projects“ section of my website, as soon as i have server disk space for it: http://www.planetz.com/Pulsar.shtml
the modular patch file is over 2mb!! i really don’t understand how it can take so much space to represent a modular patch? are the modules themselves embedded into the patch file, or are they stored by reference to the external module files?
..while it may cover a lot of the functional territory of the microwave 2/XT, it’s certainly not a replacement (there’s no way i’ll be parting with my XT anytime soon). the pulsar synths (and my waldo modular patch) just don’t capture the magic of the waldorf sound. i don’t know if they ever will, but i should say that creamware’s synth designers (and algorithm programmers) are improving with each release. creamware’s new unknow 007 synth (a sort of juno-106 copy) is really very nice, very controllable/playable, good sound. some of the earlier synths had problems with realtime parameter tweaking (twirling the filter cutoff knob would increase in steps instead of smoothly, etc) – the unknow is perfect in this regard.
ConseQuence’s note: there is the Modular II Synth in the Pulsar.. that offers the U know filters and much more.. so now its a very different thing from now!! of course the XT is much handier.. than a modular synth..
re people’s comments about the pulsar modular: i agree that when comparing with the nord modular, it’s obvious that the nord is a more mature product.however, it must be noted that the pulsar is a soft+hardware environment with a much larger goal than just modular synthesis. as an environment for mixing, recording, effects, synthesis, etc, its modularity and flexibility are amazing.it’s not surprising that a box dedicated to the modular synthesis task beats out pulsar, but pulsar really does a pretty good job. and creamware is adding new modular components with every software release. re: virtual wiring – i agree with zon and mark – good riddance to those patchcords :o) re: lack of knobs, every modular parameter is midi controllable – just hook up a peavey pc1600, or a doepfer drehbank, and you’re all set.
re: SHARC power. the pulsar has 4 SHARCs (ADSP 21061). creamware’s modular
.. DSP environment transparently spreads dsp ‚atoms‘ across the chips, balancing the load of your project components. multiple pulsars (and SCOPEs) can be chained together using creamware’s S/TDM bus, transparently increasing the
overall DSP ‚pool‘. there is of course s me overhead in this modularity. i am often surprised at times how much good stuff i can get out of a pulsar with just 4 SHARCs, and other times disappointed with how little :o) ain’t that always the way. synth polyphony is certainly one of the more disappointing
areas. in a project with nothing else running but a mixer and a miniscope (essentially a polyphonic minimoog), you’ll max out your 4 SHARCs with only 10 synth voices. other synths allow much higher polyphony. i’ve put up lots of charts covering dsp usage of every component in pulsar – see my webpage: http://www.planetz.com/Pulsar.shtml
another price you pay for pulsar’s modularity is more PC/Mac ram requirements (128mb minimum), and the need for a faster cpu (the faster the better). many MANY people have argued that the flashy graphics are a waste of mem/cpu resources (and i tend to agree), but the counterargument is that their non-OS-specific UI will allow them port their software to other OS’s more quickly (mac, BeOS and linux are all underway afaik.. note: BeOS is almost dead imo so BeOS is kicked out of the program of almost all soft&hardware companies!).
> Pulsar has got way too powerful DSP chips than Terratec card or MW.
caution ! I’m not really shure of this. The sharc is a fast FP-DSP, but IIRC the BDTi-mark index rates the 53xxx much higher. In fact it looks like that the same algorithm written in fixpoint on a 53xxx executes faster than the floatingpoint code on the SHARC.
> OK, I am comparing DSP Chips, not systems, so I can safely say that a single 60 MHz ADSP21065, also known as Baby-Sharc, has at least about *half* the power of a 66MHz Motorola DSP56303, when both run the same algorithm,both fixpoint arithmetic.
> Although the ADi chips are nice toys, they can not do magic.
But the Motorola chips can!
> do you know if the frequency modulation input of the Waldorf oscillators
> in the Pulsar is of ‚control signal‘ or of ‚audio signal‘ speed? I
It is of audio signal speed.
NOTE: I tested the Modular II.. and at least it is a very powerful and nicely expandable DSP system that distibutes the necessary power to all connected DSPs (sharcs).. imo it can be more than the Modular.. if you consider Pulsar II or some SRBs it will overtake a Nord Modular..
technical help: some US users had some problems that can easily be reduced to one thing: the power supply:
> I purchased an XT two weeks ago. It has gotten hung up five times now while
>having a sound edited. Only by turning it off and back on again have I been
>able to free it up. There was no sequencing going on, just single notes from
>the controlling keyboard. I was wondering if anyone else has had this occur,
>or could this be a problem unit? It sure sounds good, and it’s got 2.14, but I
>also own an original MW which has never glitched in this manner. It’s
>confusing.Were you able to switch it off with the power button? If so, it didn’t
really „hang“, only the notes might have hung. And this is something you
can usually stop by pressing the power button shortly. If the latter is the
case, please try to figure out on which aspect of the sound you worked
before the hang occured, i.e., switching to dual mode, activating the
arpeggiator or so.
>I power up all my gear with a single switch, started dumping samples (via
>RS-422) to my Emax II, played some sounds to the Emax II from my D-50 via
>Vision on the Mac (Local off on the D-50, transmitting to Vision which is
>in Keyboard Thru mode). I then have a nice kit of samples in my Emax II
>and start to play with the XT and it’s hung. Totally, no power key
>responsiveness, i.e. „I would switch my self off in 9,8,7,6,5,4 seconds
>but I’m totally hung“. It was in Multi mode when it Hung, on the „Midi
>Multi WMF“ patch (the last one in the bank if I recall correctly).
I had the same problem with this multi last night. It uses the sounds A001
to A008. If this multi hangs, it might solve the problem when you feed the
MWII/XT with different or the same sounds on these places. This solved the
problem on my XT.
>BTW, My XT didn’t come with a U.S. Power supply, so I went to Radio Shack
>and purchased a 273-1653B 12vDC @ 1A power supply. I noticed that the
>german power supply is rated at 1.2A. Should I need that extra 200mA? I
>also noticed that the XT is rated to draw 900mA. Could this be my source
>of problems? Maybe the plug doesn’t quite fit tightly enough or well
>enough into the power receptacle? Maybe I should cut of the end of the one
>that came with my XT (German) and attach that to the radio shack „floor“
This shouldn’t make problems. 1A is quite okay. You shouldn’t use the
German power supply because it works with 220V instead of 110V. Just stay
with the one from Radio Shack.
The first three times the XT hung I do not remember what I was changing. It
happened 3 straight times in a space of about 5 minutes. That was one week
ago. It worked fine all this week till last night when it hung up twice in
quick succession. I tried resetting it with the ‚power‘ button as the manual
instructs, but there was no response, except for the MIDI indicator light.
There are no stuck notes during this time, the machine makes no sound at all.
It was in single mode, and I was not using the arppegiator. The first time I
was changing the wavetable on a sound with a slow attack on the amp envelope
and a release time of 56 , the second time I was changing the filter type on
that same sound. Usually I play 2-3 notes when programming long sustained
sounds like this.
This seems to be a glitch. My mw2 is the same. In fact we have added
„all note off“ messages in our editor following most sound program
switches. This seems to be the main cause. Our „genetics“ section (where
new patches are created a la frankenstein) now seems 100% immune from
the hanging when auditioning patches quickly and at the same time
running a sequencer. The MWI did not do this.
The hanging may possibly be a stuck envelope release….read between the
lines in the users manual on the functions of the off/on „panic“ button.
I am sure Waldorf will resolve it eventually. They always do…and more
This could of course be a bug, but unfortunately it does not happen
here. Maybe it is the same bug that caused the unit to sometimes hangup
when selecting Multis, we have fixed this one, expect the new version
soon after Frankfurter Musikmesse.
I had the same problem as I got my XT (long long ago..) it really disappeared after a week..
there were the same thing happening: normal run and while playing… somewhere it totally hung… I had to unplug the AC adaptor because the soft off didn4t work…
but after 3 or 4 hickups like these I never enountered these probs anymore…
it must be some sort of „start up the first time“ bug… the display sayed DSP Ram error one or 2 times… this seems to have to do with it…
but I have my XT several month now.. and it never happened again.. perhaps I could help you both
(Stefan.. this problem seems to be not just with my XT.. perhaps you can find it better with this info..?) …irgendwas beim initialisieren?? vielleicht hilfts den Fehler zu finden.. don´t miss to tell it to Steve… if you ve found out more..
Reorganizing Memory?? (XT)
general info: do it with sounddiver and send the sounds bankwise.. if you got „reorganizing memory“ the XT does not receive the sounds will displaying the message: workaround: just send it again afterwards!!.. someday the XT will be changed but currently just dump it again.. or better send them bankwise …
Yep we know that. We think the Microwave has problems with the massive
sysex dump. We may re-code so that uWavEditor sends 256 individual
sounds with a user defined (breather space) between them rather than the
68k bulk sound dump. A way to get around it is to send the bulk dump 2
or 3 times and watch if the Microwave has to „reorganize memory“….it
is confusing because once you send the buld dump a few times, the
Microwave does not show „reorganizing memory“???? It seems more like a
house-keeping problem within the microwave rather than a midi buffer
thing but we are not sure.
like wise there seem to be some unusual behaviour in Multimode with
respect to S/H LFO’s and Rates of 0. We will resolve as much as we can.
Unfortunately, this is still „normal“. Just dump the sounds again several times until you don’t read the reorganizing message anymore. Then all sounds are properly dumped. In Sounddiver: send sounds in blocks of 20 or so..! Or (others): Just resend the complete dump once again. The Unit (XT,MW2) will not listen to MIDI while reorganizing!!
how to do: There’s no „fix“, because it is not a bug. It is the way the XT/XTk has to
deal with the memory. We use FLASH memory which has the advantage that it
doesn’t need batteries. The disadvantage is that it is very slow when you
have to delete memory blocks. This is done when you see the message
„Reorganizing memory“. We might come up with an improvement once but we
didn’t say that it will find its way into an update V2.xx or so.
A workaround for you:
Just dump the sounds with any speed. If you see the message, proceed with
the dump to its end and dump the sounds again. Then you probably won’t see
this message again.
I had this idea: consequence wrote:
> maybe you could send sounds in blocks of say 20.. and if they are sucessfully transmitted.. send the next block..??
> maybe Stefan could implement a little message like CTS or something saying: wait! Im busy..
Good idea, whenever there is some reorganizing I send a msg like „Wait a bit…“ and afterwards a msg saying „OK, continue to send“, or I could instead send an acknowledge saying „Dump XX typ XX received and processed, OK to send next dump“, but I doubt many dump utilities will support such a bidirectional protocol.
This „reorganizing“ appears because sounds are stored in a Flash memory, which can only be deleted in whole blocks of 32k. While reorganizing, there is a buffer for about eight dumps, but this is often not enough to catch up the data that comes during reorganization.
> since 2.17 there’s a filesystem check.. maybe there could be something like a manual defrag routine?? (like on the EIV).. and further: it could be enough for not reorganizing when init’ed before Xmit??
There already is a manual „Reorganize“, this can help a bit. The rate of reorganizings per complete dump depends on how much the sounds differ from those that are currently in the machine, so sending the whole dump twice
should be safe.
There is also the „All Sounds“ dump, see the sysex spec. This dump can currently only be initiated by a sysex request and it will most likely be useful only for Atari ST users, the only computer that handles MIDI at full bandwidth. It is a lot faster and you do not have the reorganize problem.
official statement from wolfram@waldorf:
Unfortunately you have to live with this limitation. You **could** slow
down the sending of banks by a huge amount to prevent non-received events
but the message will show up anyway.
We **could** do a workaround to this limitation but it is quite some work
and if this would introduce bugs, this wouldn’t be helpful at all. So we
said that it is better to leave it as it currently is.
– Attack is exponential in Wave and Microwave1
– Attack is linear in 4pole,Pulse,X-pole,MW2,XT and Q. And in black
XT and XTK and XT with 30 voices and Blue Q and Pulse+ as well. – Decay and Release are always exponential.
Also, were any of the wavetables from the PPGs carried over to the XT\MWII?
Simple: all and every one. Two exceptions: no samples and no „upper wavetables“. However, it might be possible that we will see these upper wavetables at some time in the future on the MWII/XT, but this is not done, yet.
it must be sayed that theres not stepped mode… It isn’t implemented anymore. The „stepped“ mode just allowed to change a wave *during* the cycle compared to the „smooth“ mode where a new wave could only be selected on a new cycle. The stepped mode just introduced clicks when one wave was completely different to another wave.
Does the MWI os 2.0 LFO sync to midi?
No. The update speed of the MW1 itself is too slow to convert the data correctly to MIDI clock. But MIDI Clock can be used as a modulation source, which means that tempos above 120bpm create a positive modulation while tempos below 120bpm create negative modulations.
Microwave Basic Hardware:
8 CEM3387 VCF,VCA,VCP (later models) 8 CEM3389 VCF,VCA,VCP (early models)
„HOW DOES IT WORK“ – AREA (+ misc & etc..)
want have a look at the XT-Keyboard? Q Rack etc.. have a look on Waldorfs site or ftp!!..
because it is always confused: they are available as 10 and 30 voice versions! you can upgrade but you have to send it in. The MW2 will have the XTs DSP ram if you upgrade to 30voices.. so you´ll be able to use delay etc as on the XT…
where to get… UNCOOK.EXE,
available at http://www.gl.umbc.edu/~bthomp1/download.html
I own a MWXT and i
>would like to know how to calculate the checksum of a sounddump. I’ve
>tried to do so but for me it seems impossible to find a matching value.
>I hope to create a simple dos-proggie to change the sound-name and
>recalculate the checksum. Or, and that a better solution, is there
>already such a program (Win or DOS) to do so?
You don’t need that. If you really want to do that math, check the sysex
docs. But usually, the XT accepts a checksum of „7F“ as an „always valid“
checksum. So, when you want to control its parameters thru a fader box,
just insert 7F as value for the checksum and everything is fine. The same
is true for all sysex data types like sound dumps, multi dumps etc.
a technical (algorithmic) description of how the waveshaper works:
It is really simple (the following is a theoretical version, the numbers
are different in the MWII/XT because of DSP reasons):
* Just see the waveshaper-wave as a table with entries from 0 to 255 (you
know, 8 bit waves ;-)).
* Now, look at an output sample of the filter and round the output
amplitude to a value between 0 and 255, too.
* Now, go to the respective index of the waveshaper wave and output the
amplitude that is written into that table entry.
An upward sawtooth outputs everything identical to its input. Entry 0: 0,
entry 255: 255
A downward just inverts the signal in phase. Entry 0: 255, entry 255: 0
A square wave starting at full negative amplitude outputs amplitudes that
are only fully negative or fully positive. Very nice for hard distortion
sounds. Entries 0 to 127: 0, entries 128 to 255: 255
A sine wave compresses all amplitudes that go up to 50% of the maximum
level and reduces higher amplitudes. Full input amplitude is 0 output
amplitude. This is the reason why the sinx filter is so interesting. Entry
0: 128, entry 64: 255, entry 128: 0, entry 192: 64.
Time Quantize (XT) is an important parameter for low notes. What it does: the
waves are only 128 samples long and when they are played back with low
pitches, they are interpolated to sound smooth. However, this interpolation
is only good with dull waveforms like sine or so. A sawtooth wave instead
or Wavetable 14 (Clipper) must sound harsh and brilliant. To achieve this,
just turn up Time Quantize. This parameter lowers (on settings 1-4) or
disables (5) the interpolation so that you get a stepped output, resulting
in more high frequencies. 5 is the same setting as the MW1, the Wave and
the PPG Wave synthesizers had.
So, for analog type sounds, **always** set it to 5. For wavetable sounds,
experiment with it.
Just look at the factory sounds, find an analog patch (i.e. CS-80) and see
what Time Quantize does.
STEFAN about the S/H Filter:
First, there is a simple 12db Lowpass, then follows a waveshaper
that takes the possibly filtered sample as index into the selected
wave and reads the corresponding sample, with some interpolation
>> I’d like to hear what was *really* going on inside Stefan’s head when he
> wrote the other wave-shaping filters (esp. the S&H Filters!)
The S&H Filter does not perform waveshaping, it reduces the sampling rate
using Sample&Hold, a technique that will make every serious DSP engineer
shiver. There was nothing special going on in my mind as I wrote it,
except perhaps something like „OK Wolfram, I showed you the bitreduction
filter and it sounded really shitty, now I do the Samplingrate reduction filter
just to show you that it will sound even worse and you can stop getting on my nerves.“
Hi to all 2.26 Users,
you can reach the Demo-Mode by pressing the Play/Shift+Power-Button;
in Demo-Sound Mode you can listen to some new and modified MW sounds
(please try all the physical controllers like ModWheel, PB and AT).
There are two Demo Multis making some „rattle and hum“; by selecting the
3rd Multi „DemoSongMulti“ you can listen to a little demo track using only
10 voices (btw it used only 2.2 kb of internal memory).
The cook-book is a simply manual to get a first impression of the MW II/XT/-k
and the demo mode is especially created for the use with this book
(approximately 20 pages). It will be released in german/english language in
the near future and every Waldorf dealer will get an exemplar.
Since the be-knobbed Q is a programmer’s dream I thought I might relate how
it loads and saves patches using a sequencer.
>Saving: the easiest way is to set the sequencer to step record, hit utility ****once to get to the sound dump page, then hit utility again and the sound will be sent via MIDI*** and the current sound gets dumped. Ignore the numbers on the dump page, I tried to figure out what they mean to no avail (surprisingly the numbers can change on the same patch). Set the sequencer to the next bar and step record your next patch etc etc.
Loading: Have the Q set to the location you want the patch to load into, and
make sure the Q isn‘ in any edit menues. Run the sequencer with the Q
patches back at a tempo of about 60. If you have, say 20 patches, make sure
you have 20 slots you want top overwrite (C001,C002,C003 etc). Then as the
sequencer is running you can manualy turn the Q master wheel as the patches
load. As soon as you see the name of the new patch loaded, turn the wheel
so you arrive at the next slot, the next patch will load into that slot,
then turn the wheel again and so on. This is much faster then sending 1
patch, stopping the seq, finding the next location, and then starting the
seq again (although you’ll have to use this method if the slots you want to
overwrite aren’t in consecutive order….C024, A099, C019, B011).
Oops, forgot to add something! Its highlighted with **s (in other words you
hit the utility button 2x to send the patch dump) ;}
from Eric Young.. again!
some comparison Q <> Pulse:
My personal standard for How Analogue SHould Sound is not the Mini, or
the P5, or the JP8. It’s the Pulse. So my most pressing question about
the Q is: does it sound like the Pulse?
I’ve just spent about six hours comparing them.
The short form: a single Q voice is a near-perfect superset of the
Pulse. What does the Q sound like? It sounds *just like* a Pulse.
Plus, of course, FM, a second filter, stereo panning, seven other
filter types, all the effects, and 15 more voices. The all-important
low-end muscle is there, but it will take a bit of work to pull it out
of the Q.
The long form: I compared the raw output of all waveforms for a single
oscillator, the saw waves together, filter characteristics (using just
a single -24db in the Q), sync, ringmod, PWM, and some simple bass
Raw output – once you’ve corrected the volume in an external meter (as
differences in loudness are often perceived as timbral differences),
you find there are some subtle differences. *Very* subtle. The
differences are only apparent in the extreme low end. In the middle
and high registers, the outputs are indistinguishable. The
differences, when you can hear them, are not differences of
quality. The low ends of both machines are rich, heavy, and FAT.
When comparing the two, be careful – the response curves are not
identical for the volume and cutoff controls. With all settings equal,
the sounds are noticeably different. Don’t expect the same numbers to
mean the same things. Also, the amplitude changes in different ways as
you move up on down the keyboard on the two instruments, and it seems
that the cutoff has to be set a bit lower on the Pulse to get the same
sound. But once you make these corrections, raw output is nigh
Filter: as noted, identical settings won’t get identical results. But
a bit of adjustment yields sounds that are truly identical. Neither
Steve nor I could tell them apart. Resonance, like cutoff, doesn’t
have a value-for-value correspondence, but once corrections are made,
sweeping the 24db filter of the Q sounds *exactly* like the Pulse.
It’s amazing. Filter modulation via wheel, LFO, etc. is perfectly
liquid; there’s not a step or hitch to be heard. The only real
difference is that actually turning the cutoff knob yields stepping on
the Q, and none from the Pulse (which is to be expected from digital
Sync: to my amazement, this sounds *better* on the Q. They’re very
similar, but the Q is somehow livelier. The sync on the Pulse has odd
spots in the sweep range where nothing really happens. This is, of
course, due to phase cancellation at certain intervals (fourths and
fifths most noticeably). But getting into and out of these ‚dead
spots‘ is much smoother on the Q.
Ring mod: OK, here the Pulse shines. The ring modulator on the Q is
full of harsh artefacts and weird peaks in the beat cycle. I’d have to
say I didn’t really care for it. Whereas the Pulse has a creamy effect
that sounds less angular and just plain better. I hope the
programmers will take another look at this feature. While I don’t have
my XT here with me, I don’t recall it having such sharp ‚bumps‘ in its
PWM: sweeping the duty cycle of a pulse wave from 0 to 127 is
indistinguishable in the mid and high ranges. On the low end, PWM
sounds a touch wider and more dramatic on the Pulse. In fact, the only
area where I can say the Q doesn’t live up to the Pulse would be the
extreme low end of the pulse waves.
Three oscs together: adjust for volume. Press key. Buy new windows.
There are some subtle differences in sound, but I stress that they are
not differences of quality. The Q’s output has the monstrous depth
that we’ve come to expect from the Pulse. It’s all there, it’s just
not quite as close to the surface as it is on the Pulse.
Building patches: I tended to concentrate on bass, since that’s what’s
so near and dear to our hearts in the Pulse. I made deep, punchy
sounds on both instruments using one oscillator, and all three
waveforms. Once again, I have to stress that getting similar sounds on
the two instruments often involves settings that are numerically
different. On the Pulse, I set up a triangle wave with a cutoff around
30 and resonance at 70. Predictably, it shook the walls. After a bit
of fiddling, I was able to create the exact same sound on the Q. And I
do mean the *same sound*. The sound did change in different ways as
one moved up and down the keyboard, but I never more than a tweak or
two away from making them match again. So, while the two instruments
behave differently, they can be made to sound the same. We’re used to
having bone-crunching bass at our fingertips in the Pulse. It’s there
in the Q as well, but it takes a bit more doing.
I tried several different cutoff/res combinations, and found that in
the case of the triangle and saw waveforms, both the Pulse and the Q
could be made to perform identically over at least an octave. However,
some differences were apparent when making this type of sound with a
square wave. The Q seemed a bit buzzy and a touch thiner. At the
extreme low end, the square waves just don’t have the warm, silky
character of the ones in the Pulse. Of course, that doesn’t mean they
Another important thing to note is that I would often make a few
changes and notice that the sounds were very different. But when I’d
look at the level on the mixer, I’d find that one was simply louder.
(Yes, sometimes the Q would be louder than the Pulse.) Differences of
10db are easily mistaken for differences in timbre, so be careful when
putting these machines head to head.
Bottom line: the Q really *is* 16 Pulses. In certain very specific
situations, the Pulse will sound just a bit better. But if you are
skeptical about the ability of software to produce a truly convincing
analogue tone, doubt no longer. Waldorf has recreated one of the
greatest analogue synths ever on a DSP chip.
– There was a lot of hiss and grit in the sound, but I suspect
Steve may not yet have completely clean power. More later.
– The mod wheel can only access half of the range of many parameters.
Assigning the wheel to pulse width allows you to set a maximum of 64
(or -64) ‚units‘ of modulation. But the pulse width goes from 0 to
127. So the wheel can take the width from 0 to 64, or 65 to 127, but
it cannot control the full range of the paramenter. I see this as a
problem. An amount of 64 or -64 in the mod matrix should give the
source control of the full range of its destination.
– I’m assuming that the external sync options for the LFOs, arp, and sequencer haven’t been implemented yet.
I’ll save my other reactions for later – I just wanted to deal with the obvious comparison first.
And what is the difference between the phaser module in the NM and the comb filter in the Q?
The phaser in the NM *is* a comb filter. I was talking about this a lot with Q folks before I got my Q. The difference between them is implementation.
Think of listening to a vintage digital delay line, and then a better quality one from more recent days. They’re both capable of doing delays, but they sound different, and the newer one may have more functionality/capabilities.
The comb in the NM can bring you the same nature of what folks are raving about here, but the quality of the module is different („better“ or „worse“ is subjective in this) than that in the Q.
It’s very hard to put a tag on it. I’m tempted to again use words like „fat“, „biting“, „ballsy“, but I fear that I’ve used those words so much lately that they’re losing their impact.
You need to find somewhere to hear a Q in person
MW PC Problems
No, there are no performance problems as long as you adress the MWPC DIRECTLY.
It sounds to me as you are working with Microedit and adress Microedit from a sequencer via a virtual MIDI-device and from Microedit to MWPC via MIDI-1. This is not the way it works. If you want to play that box, adress it DIRECTLY over MIDI-1, cause that’s the home where it lives. ;-)
You don’t need Microedit at all, as long as you are only playing. With a real XT – when you play it – you don’t have an editor-program (for example Sounddiver) open as well.
> So is it possible to run a MWPC stand-alone without a PC (providing you connect it to the appropriate power source and hit it via this MIDI hardware interface)? I looked at the Terratec documentation and it seems the MWPC needs to receive an OS dump from the PC via a ribbon cable before it will work…thus the need for a PC.
No, this point in my post was misleading. You need the PC for the OS, perfectly right and AFAIK for MIDI too. I think the MIDI-1 port is not hardwired to MWPC. What I meant was you don’t need the Editor-Software Microedit after you have initialized the box. As Microedit looks exactely like a blue XT, some people mix it up and think it IS MWPC. They would sent ProgramChanges and Controller to the software instead of sending them directly to MIDI-1. That seems to cause the problems
note on sysex in the Q:I had pulled some patches out of my Q, and then posted them up on the
website for download. One of the things that folks noticed was that they
seemed to be slot specific, i.e., that they only wanted to go back to the
bank and slot that they started in on my machine.
I asked Wolfram about this, and he told me:
Hmm, when you execute a dump on the Q, the header should look like that:
F0 3E 0F 00 10 20 00
This means that the dump is routed into the temporary buffer.
If this is not the case, just select the sound edit buffer in SoundDiver
and send a request to the Q. If you want to edit the sound manually, don’t
forget to change the checksum byte to 7F (the not-evaluated checksum value).
I personally haven’t had a chance to work through this, but I think that
the key to getting non-slot specific patches out of the Q is to use the
DUMP option in the utility menu and not as a SoundDiver request. I say this
‚cause I had used SD to get these patches in the first place.
I just looked at the mainboard in my MWII (It also has a separate display/knob/switch board). The main board ie very small and has only a handful of components:
1MB FlashMemory AMD 29F010-70pc (2 @ $3)
256k SRAM UM 61256FS-15 (6 @$12)
32bit CPU Moto MC68331-CFC20 ($25)
24bit DSP Moto 56303 PV66 ($27)
Opto-Isolator Sharp PC900V ($3)
OpAmp? TI 7705AC ($2)
SMT device? 4330 (2 @ $2??)
ABOUT THE WAVE I (Wave 2 is scheduled to be out in Fall 2000).. Already with 16 voices, yes. In this state, it is already the biggest synthesizer ever built.
The internal hardware is as follows:
* 1 68k processor for controlling the whole machine
* 1 68k processor for controlling 16 voices
* 2 custom chips each creating 16 wavetable oscillators grouped into 8 voices
* 8 Stereo DACs each converting 2 voices to the analog domain
* 16 analog multimode filters with 24dB low pass and 12dB high pass filter
* 16 analog amplifiers routable to 3 stereophonic outputs plus 1 monophonic
And that’s only a third of the maximum number of voices although we
recommend to not upgrade to more than 32 voices because it gets hot inside
the machine ;-)
some ideas on the Wave 2: DSP based sound engine that can be front panel reconfigured to emulate
> different synth architectures (ie mode1 is virtual analog, mode 2 is
> additive, mode3 is FM, mode4 is sample and resynthes, mode 5 is granular,
> mode6 is modular analog, mode7 is wavetable…etc.)
> – From the selected synth engine mode, sounds would be created and captured
> digitally by the synth (sort of like a sampler and synth source all in one
> – The dream machine’s playback engine would play each sample back but using
> the technique of additive synthesis rather than sample playback so that LFO
> modulations don’t speed up when playing higher keys…etc
> – lots of patch memory (maybe the ability to store 1000 patches)
> – individual outs 16 trs analog stereo outs and 2 adat lightpipes
> – ergonomically sensible knobs and drum machine-like pads (for use a s a
> drum entry tool)
> – user assignable cc’s to each knob
> – a „touchpad“ mouse device that can be user defines to xmit cc for x and y
– enough voices, multitimbral.. 32 or more.. :-)
– nice modmatrix: a dedicated knob for source.destination,amount and slot number..
– mabe some sort of display like the IBK10control .. knob with a display that shows the current parameter..
it should have a high internal resolution above the MIDI 128-steps.. maybe some sort of „data windows“ that can be inserted that you can determine some values between 114 and 115 or so… or / and: 16bit (ok a bit lower would be ok ,too) resolution for all parameters.. especially for important parameters like FM,tuning,curoff,resonance.. etc..
very cool wavetable editing tools, full cycle waves and samples (wavestation meets wavtabing those samples/wavtableles including morphing the samples (based on that above mentioned technique.. maybe you know the Kyma/generator/Reaktor morphing/transforming stuff??..
Is the 4-Pole the same filter as the Pulse Pluse in?
It depends, there are two versions of 4pole out, the later ones
have the Pulse filter and the early ones don’t. Now don’t ask me
how to distinguish them…
The reason why you saw NIN playing a special edition of two black XTk was
because if we wouldn’t have sent those to them they would have used other
synths from other manufacturers. The OS they had wasn’t finished but it
worked for the show. What would you have done when NIN calls you and says
„We want two black XTk for the MTV Music Awards, can we get them?“ I bet
you would do the same ;-)
The website for the Cakewalk Studioware Panel for the Pulse is:
CLICKS in SYNTHS – the click bible and why and how.. by wolfram franke
recently there was again a discussion about clicks in the audio
signal when voice stealing occurs or mono mode is active or very fast
amp envelopes are set up.
As far as I know, in the MWII is a bug with mono mode, but let me
give you some general information about voice stealing clicks. I
explained it already several times in the past but as there are a lot
of new members on the forum, it’s good to say it again. The following
is quite an in-depth discussion but you should read through it and
keep it in mind (hey, you’re a Waldorf user, you just **have** to
know more than non-Waldorf users ;-) ). Here I go:
Chapter 1: The click in theory
A click is produced when a very fast level change in the audio signal
occurs. You can easily check that on your home stereo when you play
back a CD and switch the Source Selector back and forth between CD
and a source that doesn’t play anything.
The brightness of the click depends on the speed of the level change.
The faster the level changes, the brighter is the click. So, the
level change speed can be compared with the cutoff of a lowpass
filter. There is an easy formula for it:
Let’s consider a level change from full to zero (or from zero to
full) output from one sample to another on a machine that uses
44.1kHz sample rate. So, we first transfer the sample to milli
1 sample equals 1/44100 second, which is = 0.02267573696ms.
To calculate the cutoff frequency of the click, just use this formula:
Cutoff (Hz) = 1000 / Level Change Time (ms)
which in the example results in:
44100Hz = 1000 / 0.02267573696ms
Whoops? This the sampling frequency and, err, very bright.
Chapter 2: The click in the real world
Now, how could this knowledge help you and what has it to do with
Waldorf synthesizers? Easy:
When you play a sine wave sound, only the base frequency (the
fundamental or the 1st harmonic) is present. That means, when you
play note A=110Hz, no other frequencies are involved except this
Now, what happens when you abruptly cut the sine wave to zero when it
just is at its maximum level? You get the same effect as with your
From one sample to the next, the waveform is brought from maximum to
zero, resulting in the forementioned bright click.
The same applies when the opposite happens. On Waldorf synthesizers,
you can setup the oscillators so that their phases start randomly
when a new note is played. So, you never know at which level the sine
wave is when you hit a note.
Consider it would be at the maximum level, you would get an immediate
change from zero to maximum when the amp envelope’s attack rate is
set to 0.
BTW: the effect is the same, when you have a bright waveform but
filter it so that it is very hollow.
Chapter 3: In which situations does the click occur on my Waldorf synth?
There are several situations when you can get a click and when you
know where they happen, you can try to prevent them:
* Amp Envelope Attack. On digital Waldorf synthesizers like the MWII
and the Q, the Attack rate can be as short as 1 sample. This means
that the amp volume of a note can change from zero to maximum in one
sample, or in ms: 0.02267573696ms. This results in a very bright
On the Pulse, we chose a minimum attack rate of 1.9ms, resulting in a
click with a maximum cutoff of around 526Hz. When you own a Pulse,
you probably know of the 1.9ms number from the user’s manual, because
that’s the update speed of all CVs that are used in it.
So, when you hear a click on note start every now and then, just
increase the Amp Envelope Attack rate until you don’t hear a click
* Amp Envelope Release. Here, the same as with the Attack rate applies.
When you hear a click when you release a note, increase the Amp
Envelope’s Release rate.
If the click still persists, you should also check the Release rate
of the Filter Envelope. Maybe the filter closes very fast, which can
result in a click, too.
* Voice Stealing. We know that this is the most annoying situation.
But, the click helps you: When you hear a click at a certain position
in your song, you know that a voice stealing happened and you can
easily shorten or delete notes in the editors of your sequencer.
When you count the notes and say that they don’t exceed the maximum
number of voices of your synthesizer, just keep in mind that other
notes might still be in their release phases and therefore have to be
* Mono mode. In Mono mode, a click might occur when any envelopes
(Amp or maybe Filter, too) are set to retrigger on new notes. When
the Attack rate of a sound is greater than 0, they are brought to
zero so that they can go up to their full level again. This rapid
change to zero results in a click.
* Unisono sounds. Here, a click might occur even heavier. Unisono
sounds easily exceed the maximum number of voices and because they
steal not only one but **several** notes at once, a click can be a
lot more present. It is louder and happens more often. You should
check several points on unisono sounds to lower clicks as much as
possible: are the envelope rates set to reasonable values, are the
oscillator phases set to free, is filter keytrack set to 0% (because
this can also be a rapid change) and so on.
Chapter 4: Why does my synth xy (insert product name here) produce no clicks
Should I really answer that? Because it is slooooow.
Some japanese manufacturers (I don’t say names here) prevent voice
stealing clicks by fading out voices slowly before they start new
notes. Hey, brillant idea, why doesn’t Waldorf do that? Because it
ends up in a very bad MIDI timing (and those japanese synths are
**well-known** for that).
Furthermore, most of these synths are sample-based, which means that
their attack behaviour is stored in the sample that they should play.
So, a click on note start is also not possible because the sample
somehow gradually fades from zero to maximum.
If those synths allow you to change the sample start position, they
hopefully produce clicks, too (if not, they also have slow envelopes
which we don’t hope).
A couple of days ago, someone mentioned the Matrix 12 producing no
clicks on retriggering envelopes. Yes, that’s correct, because the
Matrix 12’s minimum attack rate is around 20ms. Or in other words:
its envelopes are among the slowest you can find in a synthesizer.
The same applies to all synthesizers of the Matrix series, because
they all used Curtis chips that had an automatic smoothing filter to
prevent steppiness. The older Oberheim synths like the 4-Voice were
Also, the Waldorf Microwave and the Waldorf Wave used those Curtis
chips, but when the Attack rates of the envelopes were set to 0, this
smoothing filter was temporarily switched off, resulting in an abrupt
change. Attack 1 there is the same as minimum attack on a Matrix
Chapter 5: Conclusion
You know that we at Waldorf could prevent clicks by increasing the
minimum envelope rates or allowing bad MIDI timing. We could also
prevent that the filter resonance can destroy your hearing ability or
that you could play a C major chord. But who are we that we could
decide what **you** want from a synthesizer. Clicks can even be
musically useful and add a kind of randomness to a song that brings
it to live. A very good example is the bad, ugly, annoying, but
famous and beloved keyclick on Hammond organs.
Recently I bought the latest Art Of Noise album „the seduction of
Claude Debussy“ produced by Trevor Horn and played by the creme de la
creme (even including Lol Creme of 10CC and Godley&Creme) of
musicians and I heard a lot of clicks during a couple of tracks. I am
even quite sure that they came from Waldorf synths but I don’t know
if. You can easily imagine that I had a smile on my face.
I hope you now have even more fun with your „clicking“ Waldorf synth.
info: PPG Wave (only because it is the predecessor of Waldorf..) disk format: there is a Linux driver to read the OS9 data
format, or at least a raw driver that can be customized or so.With Windows, you will probably don’t have a chance…related to the OS/9 that runs on the old Radio Shack Color Computers, which have 6809E’s in them. It was a Unix-like OS that
could multitask even on that little 2 MhZ (as I remember) machine. It would be cool if you could read the PPG disks using one of those . . . As far as I know, it was OS 9 (not to confuse with MacOS 9 ;-)) which was developed by some German programmers. There was a computer based on this OS and the Waveterm B used the same main board.Check with the experts: http://www.nashville.net/~antarct/ppg.htm
Here is a nice (soon to be filled) FAQ for the Q… have a look.. it’s growing.. (Maintained by Steven Noreyko)