Waldorf MicroWave as a plugin?

[gearwatcher]

Either way, Surge already does both the ROM scanning wavetable emulation in perfectly bandlimited way:

https://github.com/surge-synthesizer/su ... wavetables

and it emulates SEM style filters in the OB models. For all intents and purposes you're only limited by your (lack of) skill and imagination at this day and age when getting sounds like these.

[chk071]

The more I think about this, the less likely it appears to me that it will be a microWave soft synth. Maybe a microWave soundset for the M marketed by Waldorf, or a microQ/Pulse 1 form factor cheaper "micro" version of the M.

Of course, I'd like to be proven wrong with those guesses, but, Waldorf really didn't appear to make huge efforts to their software as of late. Largo 2 is more than a year old now, and has only seen a single update in that time. And, the GUI still looks like something they let the developer do in his spare time, with zero experience in terms of contrast, choice of colours, font type and size, and usability. Which is probably true.

[drsyncenstein]

you probably read a lot about it. It is still conjecture and part truth. However the way you responds seem you are a bit too convinced of your own story. Would be nice if that could be skipped.

It's based on technical analysis by people who dissected these units with deep engineering understanding of them, and statements of the original designers.

Much more reliable sources than bunch of people swearing their golden ears can somehow infer what is and isn't "completely different".

That's why i said you probably read a lot about it. But that is not the point i was making. The point is grasping and reproducing that. And none of that says anything about if there are audible differences or not between synth revisions. Or between hardware and emulation.

Some softsynths are very close to what they emulate and some are just detailed enough to be in the ballpark of original units. It's not only component tolereances that makes the emulated hardware different from the original. You can call that golden ears, but i bet none of their developers call them perfect emulations in private.
Developers make trade offs to be able to run them on current computers. Or to be able to finish them and earn a living. They makes mistakes or fail at things. Some developers even admit that their emulation was based on a unit that was not working fully or was a diy clone of the original itself.
There is nothing wrong with people wanting something that emulates the last few percent.
Golden ears do silly stuff like buying very expensive esoteric cables for their loudspeakers, that's a whole different level of sillyness.

[IvyBirds]

I am pretty certain, as I have read it on multiple sources, that the original Microwave does what PPG Wave does, which is, similarly to the Emu II, it uses "pitch dictated EPROM scanning" (it would obviously be RAM on the Emulator) into an analog filter.

Simplified, it's a pulse corresponding to the pitch going into an increment input of an encoder, which goes into lesser significance address bits (higher significance ones selected the table), which effectively scans the ROM in such speed that a full cycle corresponds to the pitch fundamental. The value from the ROM is converted into a voltage in a fairly primitive DAC.

Not sure what you are reading, but that is not how the E-II works. The E-II uses DPCM mu-255. That's a technology borrowed from American Telecom Companies. That's an algorithm that gets similar results to 12 bit sampling with 8 bits of storage and ram

The E-II has no pitch fundamental, much like the telephone technology it uses wasn't concerned with pitch, the E-II isn't either. The E-II takes incoming audio and records it using the Telecom technology, then it stores that on a disk as a DPCM file at a 27.7khz sample rate.

On playback the DPCM file is loaded into system RAM from disk and assigned a root note. That is just there for math purposes, upon playback the CPU inside of the E-II changes the sample playback rate to alter pitch. Play it down an octave? It will drop from 27.7khz to 13.85kh, Play up an octave it and it will be 55.4khz. When you do this the pitch and the playback length of the samples changes. That data stream then goes through a DAC borrowed again from American Telecom

In Telecom use it uses an ADC to digitize your voice send it down the pipe in real time where it uses a basic DAC on the other end to convert it back to analog audio. E-mu recreated that only they also inserted a data recorder and playback system that plays back at variable speeds which also changes the pitch

The system because it uses off the shelf Telecom parts and designs on the digital parts made the E-II sound amazing and did so at an incredibly inexpensive cost compared to the Fairlight CMI or Synclavier II.

Essentially what E-mu did was take a basic analog subtractive synth design with 8 voices only instead of analog oscillators it plays back phone conversations at various speeds just like you were manipulating the speed of a tape deck, or record player

You wouldn't do that on a Wavetable synth, it would be to expensive to produce and not be needed for single cycle waveforms

Also should note that the Arturia plugin version of the E-II nails the digital side perfectly as the 8 Bit DPCM mu-255 is super well documented and they just needed to insert that into an analog synth model

[gearwatcher]

You've still explained exactly what I did, except you got hung up on sample encoding, but completely glossed over the part about HOW it reproduced the samples at different pitches - which is done EXACTLY as I explained it above (sans the cycling part which doesn't apply to samplers).

Also calling nonlinear PCM codecs "great sound" is certainly one way of putting it, but no, DPCM/Alaw/Ulaw are well known technologies - if they weren't lossy distorting crap, they would still be used - they were just a cost/bandwidth cutting measures, necessary evils until technology made them obsolete.

[drsyncenstein]

If IvyBirds would have also named the AM6070 companding DAC as one of the parts that made the EII, i'd given him\her 10 points.

What you wrote is this:
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.

You confuse pitch and timbre. The EII uses a clock of which the speed is divided down depending on the key, to play back the whole sample. Sifting is for wavetablesynths.
I'm not even sure that the MW does use divide down, it had a powerfull processor so more opportunities. But i'd be happy to have you post a link that proves us wrong here.

You don't need divide down for crappy aliasing playback, the E-MU EMAX did that with an algoritm of it's own. It's the 8 bit wavetables and the filters that make people like the Microwave.
And if you've ever used an EII, you'd notice that it's actually a very well behaved and pretty clean sampler.

Likewise: SEM filter = a state variable filter built from OTA's and FETS, plus a mixer for summing the different (bp/lphp/lp) types of filter outputs. Versus a CEM chip which is just Lego for filters.

[gearwatcher]

> You confuse pitch and timbre

The what?

Wtf are you even on about?

Also what dividers? I was talking about how variable speed playback functioned in these designs, versus the ones using a fixed speed ring buffer dac and software interpolation.

[drsyncenstein]

scanning wavetable = timbre
changing clock = pitch
variable speed playback = dividing down a masterclock running at the speed of the highest sample rate + extra speed needed for the maximum tranposed up range.

here is the EII service manual, show me the pages and figures where the variable speed vco for playing samples is described and i'll give it a rest:
https://ia800704.us.archive.org/8/items ... Manual.pdf

[IvyBirds]

scanning wavetable = timbre
changing clock = pitch
variable speed playback = dividing down a masterclock running at the speed of the highest sample rate + extra speed needed for the maximum tranposed up range.

here is the EII service manual, show me the pages and figures where the variable speed vco for playing samples is described and i'll give it a rest:
https://ia800704.us.archive.org/8/items ... Manual.pdf

Show me in the manual where there is a VCO or where anyone said there was one and you can stop embarrassing yourself

The Manual does reference many times and shows in multiple schematics the fact the E-II uses 6072 IC's. They are a monolithic 8-bit, companding digital-to- analog (D/A) data converter

You keep on saying the Companding doesn't matter but it right in the DAC just ahead of the analog filters and analog outputs you can't separate the two that is how the device runs , which is very different from the Microwave

The Microwave of course doesn't use them and is entirely different

[drsyncenstein]

Hmm. to who is the post above directed? btw, someone mentioned vco's on page 5.

[gearwatcher]

The custom state machine ASIC they refer to as the "microcontroller" sends samples from RAM into the input latch in front of the DAC at a custom rate. The latch holds the value, effectively causing the DAC to hold the output voltage, which is then fed into the filter.

The effect of that is the exact same stepped, non-iterpolated sample output as in the PPG and, by extension, Modwave which does the same thing with less discrete digital logic as the PPG according to its designers.

I made no claims about any VCO, or any dividers, and I didn't confuse timbre and pitch - you perhaps got confused interpreting what I wrote but that's your problem.

Yes content of the sample memory determines the timbre, yes, the rate of sample playback determines the pitch - however that is blatantly trivial and I would like to see where the f**k have I claimed the opposite.

And again - whether or not samples are stored in linear or non-linear PCM is absolutely irrelevant to the core point which is that playback of non-dsp-interpolated samples has been repeatedly successfully emulated in software by multiple developers. And sure, it's even the case for the more complex case of non-linear PCM. Does that somehow make the linear case less likely to be successfully emulated?

As for SEM integrated filters and Oberheim, do Google SEM3320 chips. Yes, again, Curtis filters were successfully emulated in software. Again by multiple actors.

[drsyncenstein]

Well, that is more of a reply than i expected.

This is kvr, things like sample rate and timbre and how that is achieved in hardware or software is not trivial for most.
And you replied in a way that suggested little knowledge and a lot of attitude. I reacted on that.
In retrospect that could have been because you feel that those things are trivial. The f**k you's don't help make it better reading or taking it more serious.
Maybe i misread parts, maybe you could write more clearer. Because i do think that you said otherwise than you now claim.

I don't care about the digital part being emulated, especially not for off the shelf parts from the eighties. I never claimed that could not be emulated. The Hermann seib ppg vst is a perfect example of it being emulated with very good results. But emulating the digital part of a microwave or ppg is not enough to make it a perfectly emulated instrument.
I already wrote enough about that.

SEM3320 does not give many results, some chinese stuff without datasheet. Pictures of chips.
Alfa RPAR cloned the cem 3320 years ago, cool audio reverse engineered them also. So nothing new. They are functionally identical to the originals.
But why do you call those software?

I never stated that curtis filters have not been emulated in software. u-he for example did the 3320 for the repro. I asked about a specific filter from the cem range that was used in rev a microwave. When is a filter in software emulated successfully?

[zerocrossing]

you probably read a lot about it. It is still conjecture and part truth. However the way you responds seem you are a bit too convinced of your own story. Would be nice if that could be skipped.

It's based on technical analysis by people who dissected these units with deep engineering understanding of them, and statements of the original designers.

Much more reliable sources than bunch of people swearing their golden ears can somehow infer what is and isn't "completely different".

That's why i said you probably read a lot about it. But that is not the point i was making. The point is grasping and reproducing that. And none of that says anything about if there are audible differences or not between synth revisions. Or between hardware and emulation.

Some softsynths are very close to what they emulate and some are just detailed enough to be in the ballpark of original units. It's not only component tolereances that makes the emulated hardware different from the original. You can call that golden ears, but i bet none of their developers call them perfect emulations in private.
Developers make trade offs to be able to run them on current computers. Or to be able to finish them and earn a living. They makes mistakes or fail at things. Some developers even admit that their emulation was based on a unit that was not working fully or was a diy clone of the original itself.
There is nothing wrong with people wanting something that emulates the last few percent.
Golden ears do silly stuff like buying very expensive esoteric cables for their loudspeakers, that's a whole different level of sillyness.

I’ve been at this for a very long time, and there’s one thing I’ve learned and it’s that everything is its own thing. Even reissues are different than the originals. Different runs of the same model can be different. Working in music stores, I heard it. Less true of digital gear, but often evident in analog stuff. You can buy one and then just be sure that you have “the one” or shut up, because that’s that. You can also dig into whatever tool you have and figure out how to get great sounds out of it.

[El°HYM]

Very sad ping - pong thread once again. So which software emulation gives me the Rev A - Filter 1:1 is what me wants to know.

[Uncle E]

None at this point, AFAICT.