Waldorf MicroWave as a plugin?

[DrGonzo]

8 bit wavetables through the CEM 3389 filter chip is very special and the reason the Rev A is more popular than the Rev B.

Curtis filters have already been taken on by multiple developers, and there's more than one sampler out there emulating legacy sample reconstruction hardware (Arturia's, TAL's, Amigo to name a few). Heck SurgeXT has a wavetable mode doing something very similar, as well as a number of modelled filters.

Getting it good enough for people who believe in pixie dust in particular revisions of a particular product is a futile endeavour anyway, and all these companies know better. I am personally more in the "who needs yet another wavetable plugin" b2b "Waldorf sucks balls at software" camp.

If this is all futile and pixie dust then why bother with anything ? All you need is your DAW's stock plug ins and record everything through the built in input of your computer.

I think the problem is that people don't hear the same.

I had a long discussion on Gearslutz about the the 303. I am not a die-hard fan of the 303 at all but I can hear the difference between the original, a good clone and a bad clone. This guy however, fully insisted you could program the exact same sound with any analog or VA. It was a frustrating conversation to say the least, but it made me realize that some people just can't hear the same as I can hear, then they dismiss it as "pixie dust".

But in a sense they are absolutely correct. They can't hear it.

[Shabdahbriah]

The classic Waldorf Wave, I would personally like to see resurrected (looking at you Behringer):

Wave-1.png

[zerocrossing]

Even Waldorf wont be able to nail the MW Rev A - filter in the digital world. They never did; yet still offering some of the very best digital filters around.

There are already a lot of great analog filter models. I don’t think there’s anything so unusual about the Microwave’s CEM. The thing most models don’t do well is audio rate filter modulations, but the Microwave didn’t have that. It used to be high resonance that tripped up emulations, but that’s been nailed a long time ago. Icarus has good analog models, Dune 3, Massive X, all have good analog modeled filters, in fact, I was able to get such good results running classic wavetables in Dune 3 that I didn’t feel bad selling my M and picking up the 3rd Wave desktop instead. I wanted that one all along but didn’t have space, so when they did release a desktop model, I went for it but didn’t need both the M and the 3rd Wave.

Seriously though, I’d sell the 3rd Wave if someone did a good emulation of how the PPG renders wavetables that had a similar feature set. I’m sort of obsessed with that sound, but I don’t like PPG Wave 3 due to its poor UI and general simplicity. If Dune had a PPG Wave wavetable mode, that would be perfect.

[drsyncenstein]

If you use Windows and can live with the filter emulation, the wavetable part on this is spot on: https://www.hermannseib.com/english/syn ... avesim.htm

[drsyncenstein]

I rather have no mw 1 than a mw 1 rev B and I rather have no MS 20 than an MS 20 MK2 the difference is THAT big to me. In case of the hardware Oberheim polys, I sold them all in favor of the Gforce OBX plugin, the first emulation that got it right for me.

I never heard the rev B in person. I got version A and the cem chips of version B in the glorious e-mu emulator III's filters. So I can't imagine that Microwave B sounds worse.
Since there is a lot of false shit about synth revisions i always assumed people meant version II, the dsp version. I'd be interested in a good A-B comparison so i can hear what it's like.
You know any?

btw, you sold an ob-x? Cause i like the vst, but not THAT much.

[gearwatcher]

Getting it good enough for people who believe in pixie dust in particular revisions of a particular product is a futile endeavour anyway, and all these companies know better.

If this is all futile and pixie dust then why bother with anything ?

I deliberately left my complete sentence above so that it's hopefully plainly evident exactly what is futile.

The thing is, you people will moan about differences that are at the level of component tolerances, often the type of difference that is there between serial number 1001 and 1002, and quite often based on some rosy tinted faint memories than actual comparisons. And use hyperbole like "sounds absolutely nothing like" to moan.

The truth is that you lot moan way more than you spend, especially on software - so why should software companies expend valuable and expensive engineering resources to cater for your whims?

Personally I don't have a horse here. I am not in the market for another wavetable synth, have abhorrent experience with Waldorf software, and don't expect that they will do a good job this time.

But for completely different reasons.

[gearwatcher]

Are you sure about that? For the 3389 I know only one, from Cherry.
Advertised with "CEM3389 FILTER is a low-pass filter with 24dB/octave slope module based on CEM filter chip CEM3389 very similar to the classic Oberheim filter sound."
And that's the biggest bullshit in the world. None of the classic Oberheims had a filter like the 3389. They had an early CEM chip that sounds pretty different.

The chip is a fairly standard in-wafer integration of the SEM filters and a casual glance at the schematics in various datasheets confirms that.

A sampler like TAL is not using wavetables, it won't give the same results.

Wavetables are nothing but sample playback. Waldorf style WT is looped sample playback with the pitch dictating the clock speed of shifting through the address space of the wave table ROM chip. That type of special sauce is present in TALs emulation of Emulator II which used equivalent approach.

And Arturia also emulated it in their Emu emu.

As an engineering problem - both have been successfully tackled by many.

To you, I'm sure, they'll sound "nothing like", read my reply above to see what I think about that tho.

[drsyncenstein]

The chip is a fairly standard in-wafer integration of the SEM filters and a casual glance at the schematics in various datasheets confirms that.

Wavetables are nothing but sample playback. Waldorf style WT is looped sample playback with the pitch dictating the clock speed of shifting through the address space of the wave table ROM chip. That type of special sauce is present in TALs emulation of Emulator II which used equivalent approach.

And Arturia also emulated it in their Emu emu.

You seem to confuse cem filter chips with sem filters. Sem is an Oberheim filter implementation, while a Curtis cem chip is a filter building block. The datasheets of the cem chips only show the connection pins and architecture, not what exact circuit is in the chip.

A wavetable is a set of multiple, different, single cycle waveforms.
So in a sense it's a sample indeed.

The clock just plays the waves faster or slower depending on the key pressed on the keyboard. That is what makes it possible to play the sample in a different pitch.
That is the part that Tal and the Arturia EII emulates.

To make it play back like a wavetable you have to be able to sweep from one single cycle wafeform to another single cycle waveform. While looping between those waveforms.

If you have a sampler where you can set the startpoint and the loop point to a random place in the sample and modulate the loop (with for example an lfo) while the sample is playing, you get the same kind of effect as scanning a wavetable.
Not many samplers can do that.

I could not find anything in the manual that suggests that TAL sampler can do that.
The EII can't do it either.

I would not know if all engineering problems are fixed. It's just that an emulation of A is not an emulation of B.

[DrGonzo]

If you have a sampler where you can set the startpoint and the loop point to a random place in the sample and modulate the loop (with for example an lfo) while the sample is playing, you get the same kind of effect as scanning a wavetable.
Not many samplers can do that.

I've always wondered about this. Why do you think that is? I can't see how being able to modulate the position of the loop to be that hard to do, or eat up too much resources.

[christian f.]

Getting it good enough for people who believe in pixie dust in particular revisions of a particular product is a futile endeavour anyway, and all these companies know better.

If this is all futile and pixie dust then why bother with anything ?

I deliberately left my complete sentence above so that it's hopefully plainly evident exactly what is futile.

The thing is, you people will moan about differences that are at the level of component tolerances, often the type of difference that is there between serial number 1001 and 1002, and quite often based on some rosy tinted faint memories than actual comparisons. And use hyperbole like "sounds absolutely nothing like" to moan.

The truth is that you lot moan way more than you spend, especially on software - so why should software companies expend valuable and expensive engineering resources to cater for your whims?

Personally I don't have a horse here. I am not in the market for another wavetable synth, have abhorrent experience with Waldorf software, and don't expect that they will do a good job this time.

But for completely different reasons.

You are the one moaning.
I own pretty much every notable plugin out, so your assumptions are incorrect.

My problem is that just about any company "emulates" a certain piece of desired hardware stating that it's the same or close to the real thing while it rarely is.
Most people that never touched or heard the hardware properly fall for the pretty GUI and marketing talk which I think is sad.
That doesn't mean that those are bad pieces of software, it means that it's false advertisement and it would have been way more respectable if they just gave it a different name and mention that it's not a 100% emulation.

I love software as much as hardware, but with hardware I do have a clear preference for certain revisions and if you ever going to spend a lot of money on hardware it would be wise to make sure to pick the best sounding one. I did my research and spent a lot of time to compare them one on one and then made my choice. If you can hear a difference and if that is important for your sound it's not futile. You just seem to be easily impressed and satisfied. Nothing wrong with that.
In the end, as I stated before...in 2024 it really doesn't matter anymore what you use. You can make good sounding music with anything if you take the things for what they really are.

[christian f.]

I think the problem is that people don't hear the same.

I had a long discussion on Gearslutz about the the 303. I am not a die-hard fan of the 303 at all but I can hear the difference between the original, a good clone and a bad clone. This guy however, fully insisted you could program the exact same sound with any analog or VA. It was a frustrating conversation to say the least, but it made me realize that some people just can't hear the same as I can hear, then they dismiss it as "pixie dust".

But in a sense they are absolutely correct. They can't hear it.

This is what it essentially is.
Not everyone hears/feels/likes the same.

I never heard the rev B in person. I got version A and the cem chips of version B in the glorious e-mu emulator III's filters. So I can't imagine that Microwave B sounds worse.
Since there is a lot of false shit about synth revisions i always assumed people meant version II, the dsp version. I'd be interested in a good A-B comparison so i can hear what it's like.
You know any?

btw, you sold an ob-x? Cause i like the vst, but not THAT much.

If you compare the A with the B you will hear the difference for sure.
It's very personal which one you prefer, but I never met anyone that picked the B over the A.

I always preferred the OB-Xa over the OBX and then replaced that with an OB-8 and Xpander. I also got the OB-X8 which I sold after 2 weeks. The Gforce OBX does everything I like about the Oberheim Polys very well and certainly good enough to not have to have those huge machines around.
I kept an OB-1 as my only Oberheim synth.

[DrGonzo]

I think the problem is that people don't hear the same.

I had a long discussion on Gearslutz about the the 303. I am not a die-hard fan of the 303 at all but I can hear the difference between the original, a good clone and a bad clone. This guy however, fully insisted you could program the exact same sound with any analog or VA. It was a frustrating conversation to say the least, but it made me realize that some people just can't hear the same as I can hear, then they dismiss it as "pixie dust".

But in a sense they are absolutely correct. They can't hear it.

This is what it essentially is.
Not everyone hears/feels/likes the same.

A similar thing just happened to me just a couple of days ago.

Apart from sounds, the only other thing I have the same level of obsession with is coffee.

I just replaced my old coffee maker with a new more advanced machine. I prepared a blind test for my wife so she could experience the enormous difference in taste. She couldn't taste any difference between the two.

But at least she didn't accuse me of tasting 'pixiedust'

[christian f.]

But at least she didn't accuse me of tasting 'pixiedust'

[gearwatcher]

You seem to confuse cem filter chips with sem filters. Sem is an Oberheim filter implementation, while a Curtis cem chip is a filter building block. The datasheets of the cem chips only show the connection pins and architecture, not what exact circuit is in the chip.

Nope. They show enough of the internal architecture to draw that parallel. Which is why Cherry drew it in their marketing materials and why pretty much everyone considers those Curtis filters to be based on Oberheim designs, and some are even internally wired for e.g. lowpass functions, without allowing the liberty that the more open-ended "filter building block" ones allow the designer, making them indeed "filters on a chip". And btw Oberheim also employed chips instead of discrete filters in their later designs too, which is exactly what the Curtis ones sought to replace on the market.

https://electricdruid.net/cem3320-filter-designs/

Being deliberately obtuse and "oooh, if it's not molecule by molecule identical, it is completely different" will not make you look clever.

Which brings us to:

A wavetable is a set of multiple, different, single cycle waveforms.
So in a sense it's a sample indeed.

Orly? Its's a sample "in a sense"?

The clock just plays the waves faster or slower depending on the key pressed on the keyboard. That is what makes it possible to play the sample in a different pitch.
That is the part that Tal and the Arturia EII emulates.

Which is THE ONLY THING that makes these samplers/wavetable reproducers have any diference at all from those samplers and wavetable reproducers that use sample interpolation and reconstruction filters.

Whereas this difference...

To make it play back like a wavetable you have to be able to sweep from one single cycle wafeform to another single cycle waveform. While looping between those waveforms.

...is entirely irellevant to the discussion about "special sauce" of PPG/Waldorf sample reproduction which boils down to the difference in playback being dictated by the design where speed of ROM scanning is how we deal with the pitch, versus the more common design of using interpolation and reconstruction which is employed in majority of other samplers and wavetable synths (those that don't use spectral wavetables which is another design completely different from all mentioned).

If you have a sampler where you can set the startpoint and the loop point to a random place in the sample and modulate the loop (with for example an lfo) while the sample is playing, you get the same kind of effect as scanning a wavetable.
Not many samplers can do that.

While this is completely irellevant to our discussion of whether or not the core engineering problem of emulating that specific design of sample reproduction has been successfully emulated -- in fact many still can. Like, for example all those granular samplers, or the one in Ableton where you can link start point to modulation source and have a fixed loop move with that start point.

Many libraries for Kontakt (which I don't own) claim WT capabilities -- which makes me conclude that it too can perform this type of sample scanning.

I could not find anything in the manual that suggests that TAL sampler can do that.
The EII can't do it either.

I would not know if all engineering problems are fixed. It's just that an emulation of A is not an emulation of B.

Again being facetous and deliberately obtuse, and deliberately avoiding the matter at hand.

[drsyncenstein]

O boy. There is so much bollocks in your reply that i won't go into it any more.