How many of you are getting the Osmose?

[magog]

Its an eye-opener in regards how much processing is required to get things into a useful range, to achieve the right balance of compromises necessary to tame the beast. People will vary in terms of quite what number they want to set this stabilisation option to in order to get the right balance for themselves, and I'm glad Expressive E let us go beyond what the lowest useable value could possibly be.

Steve - thank you for this - that was a fascinating look into the challenges of MPE implementation. I had no idea about the internal smoothing and signal processing that has to take place to get musical results. Makes total sense.

In the Sonic State video they talk about how the Osmose allows you to set ranges and thresholds for the gestures so that you can adapt the device to your own style of playing, and that these are by default global, but you can store them in user patches as well. Personally, I'm a little heavy handed, so I appreciate that I can make the device suit my own playing style. And maybe I can learn a lighter touch over time and use the settings to help me do that.

And it's so cool to now learn from you that the Osmose allows adjusting the settings down to the point even where the gestures stop becoming musical.

[pdxindy]

but if I actually sit in front of an Osmose and draw live graphs from the MPE and MPE+ data, it doesnt really feel like the MPE version of the data is lacking the important subtleties of key position.

You keep saying that, and I would prefer you to be correct, but I have yet to hear a demo using a software synth (CLAP or VST) that has the tonal nuance of some of the Osmose presets.

[magog]

In the Sonic State video, I think they said that the MIDI stream sent externally from the Osmose is the same stream of data sent internally to the sound engine. Not sure if that's true or not or whether it's an oversimplification.

EDIT: I would hope that's the case, because otherwise, how would you record and playback your Osmose performance from your DAW?

[ThomasHelzle]

I think this is a bit of a silly argument that can only go in circles.

What is the difference between a mediocre violin player and an outstanding one? Milliseconds, tiny tonal shifts, tiny variations in pitch. "Feeling" is extremely subtle and I totally feel being much "closer" to the sound with Osmose as with my normal keyboard and a VST, even a good one (using 48kHz, 128samples buffer), and I'm not a very good player actually.
But that feeling of being "one" with an instrument I get a lot with the better presets in the Osmose.

Those adjustments for the sideward movement you mention - like Chris showed in the SonicState video - are mostly there, so that you have a certain neutral zone in the middle and especially with wide spread chords, don't instantly go into detune mode.
On the other hand, when I play with one, two or three fingers, I totally enjoy that rather instant coloration I get with some sideward drive. And of course I can't control that sideways movement with 64bit floating point precision - but again those involuntary nuances do a lot in acoustic instruments, even if you can't quantise them easily.

So this whole discussion feels a bit like science theoretically proofing that bumblebees can't fly, and the bumblebees not caring one bit.
There is so much more subtle stuff going on in acoustic music and a lot of the electronic stuff we use is dumbing it down to some good-enough mediocrity.

Inside the Osmose you get the full data.
In a DAW, you get what that DAW leaves untouched, many of them use smoothing algorithms and data reduction (Bitwig, I'm looking at you there...), so ATM we record Audio.

The argument about Midi being slow was answered by one of the inventors a while ago, who said, that via USB, that is basically moot, since you could in theory drive it through at any speed you want, so big amounts of data aren't really an issue over USB.
Of course, the OS and the drivers come into play here as well.

And I'm not even saying that you can't do great stuff with Midi and VSTs, I just think it's silly to dwell on the point and claim that all that additional data is useless or serves no point.

And for the pressure I definitely think there is a difference between 128 values and 14 bit - especially for the key glide - you can do VERY smooth and slow bends there...

But anyway, I just don't see much merit in this argument...

Cheers,

Tom

[ThomasHelzle]

In the Sonic State video, I think they said that the MIDI stream sent externally from the Osmose is the same stream of data sent internally to the sound engine. Not sure if that's true or not or whether it's an oversimplification.

EDIT: I would hope that's the case, because otherwise, how would you record and playback your Osmose performance from your DAW?

There are two virtual Midi ports via USB:
The first is sending normal MPE data and the second is communicating with the Eagan Matrix and sending MPE+ with higher resolution (this has different modes in the keyboard preferences as well). The latter is the data to record in your DAW if you want to play it back to the Osmose, so hopefully it's the same as internally, although you will have some latency added.

Not sure what the hardware Midi ports are sending.

Cheers,

Tom

[SteveElbows]

And it's so cool to now learn from you that the Osmose allows adjusting the settings down to the point even where the gestures stop becoming musical.

To be clear, its only the left-right key wiggling where they have provided a stabilisation feature that we can turn to either extreme. There are still adjustable curves for the other sort of movement, but stabilisation isnt required in the same way for those.

[SteveElbows]

You keep saying that, and I would prefer you to be correct, but I have yet to hear a demo using a software synth (CLAP or VST) that has the tonal nuance of some of the Osmose presets.

I keep going into tedious detail because I want to be very clear about what I am claiming and what I am not claiming.

I am very much urging caution in regards the assumption that giving other synths the ability to handle MPE+ is going to be the key difference maker. It should help in some ways but its only one part of the picture. Other parts of the puzzle involve the synthesis features particular synth engines have, taking care to craft appropriate patches that make best use of the expressive control surface to control the aforementioned features of the synthesis engine in a way that feels musical, expressive, nuanced and immediate.

I completely agree that demos do not currently exist which enable you or anyone else to determine exactly the full extent of what can be achieved when using the Osmose as a controller for other synths. I cannot compensate for that with all the words in the world, I simply seek to explore some of the theoretical details in the meantime. I have certainly claimed that using it with MPE can get closer than some people think, and that the MPE+ side of things doesnt make as much difference as people think. This is not the same as claiming that MPE+ makes no difference at all, only that you shouldnt look to the + to make all the key differences, or to explain why you havent yet seen a demo with a MPE synth that satisfies you.

The EaganMatrix synth is a very good synth indeed for coupling tightly to expressive controllers. It was an excellent decision to make it part of Osmose, it makes the Osmose feel alive, makes it a cohesive instrument right out of the box, it proves things to people without risk of them falling down at any of the other hurdles I keep going on about. And expressive control is at the very heart of EaganMatrix patch design rather than being an afterthought or something only a fraction of patch designers for that synth focus on.

I dont want to bore everyone silly by clarifying what I'm getting at endlessly. But I really want to be clear that I am urging caution in regards assumptions in two opposing directions: lets not assume that we cannot achieve wonderfully alive, connected, expressive results using Osmoses MPE output with certain other MPE synths and well crafted patches. And lets not assume that other synths gaining MPE+ support will be enough on its own to ensure that the results you've seen from EaganMatrix will be unlocked for other synths. And try not to read too much into a lack of demos or how long we might yet have to wait for more compelling demos. There are loads of reasons for that, and we've seen a somewhat similar state of affairs with other MPE controllers in the past. eg People want to take time to practice their playing before doing demonstrations, people have a lot of mileage to get from the built in engine and sounds before they spend large amounts of time pairing Osmose with other synths. And when some people do explore that stuff, who knows whether they will pick the very best MPE synths for the job or how long it will take them to craft the best possible patches to use with that synth and the Osmose together.

[SteveElbows]

I think this is a bit of a silly argument that can only go in circles.

What is the difference between a mediocre violin player and an outstanding one? Milliseconds, tiny tonal shifts, tiny variations in pitch. "Feeling" is extremely subtle and I totally feel being much "closer" to the sound with Osmose as with my normal keyboard and a VST, even a good one (using 48kHz, 128samples buffer), and I'm not a very good player actually.
But that feeling of being "one" with an instrument I get a lot with the better presets in the Osmose.

Those adjustments for the sideward movement you mention - like Chris showed in the SonicState video - are mostly there, so that you have a certain neutral zone in the middle and especially with wide spread chords, don't instantly go into detune mode.
On the other hand, when I play with one, two or three fingers, I totally enjoy that rather instant coloration I get with some sideward drive. And of course I can't control that sideways movement with 64bit floating point precision - but again those involuntary nuances do a lot in acoustic instruments, even if you can't quantise them easily.

I think you are missing the specific context of why I started getting into this detail. I've been trying to deal with specific assumptions about what MPE+ does and does not offer compared to MPE, and I think people go on to think I'm trying to make other points and claims that I'm not actually trying to make.

I'm actually in agreement about much of what you have said, and indeed I really wasnt so much into synthesizers at all before I got into MPE controllers, the small variations and expressiveness they offer makes everything come alive for me.

When it comes to subtle timing variations, I agree thats important, and thats exactly why I started going on about temporal aspects in a previous post. And why I was keen to point out that MPE+ messages compared to MPE messages are not really giving us something extra in this respect. If developers actually take heed of every point about timing that Haken make on their MPE+ page then that could actually give us something extra, but thats not because of the differences in midi data between MPE and MPE+.

And my point about left-right pitch wiggle stability settings was only to give an over the top example of the limits beyond which detailed data can become excessive rather than a benefit. And I actually missed out one of the settings needed to get the full degree of absurdity out of the experience, in addition to turning off stabilization you also need to turn up the bending range. But again please dont lose sight of why I brought this up, it was a deliberately silly combination of settings, and it wasnt to ridicule these controllers or diminish their value, their glorious advantages.

Let me put it very clearly one last time since I've just been talking about pitch again. With these posts I'm trying to deal with assumptions about what benefit MPE+ might bring to other synths. But some of us have looked at the MPE+ data that comes out of the Osmose and its not actually adding any extra resolution at all to the pitch messages compared to normal midi / normal MPE. When exploring this we also got someone who had a Continuum to see if they could get extra resolution out of the MPE+ pitches on that device either. They could not. Whether thats always been the case, whether its deliberate or only happened beyond a certain EaganMatrix version I cannot say. Certainly the MPE+ documentation suggests that CC87 messages can be used to increase the pitch resolution to 21 bit in the same way those same CCs are used to increase the MPE+ Y and Z to 14 bit, but it doesnt seem to be used. Not even when pitch range is set to something that should be MPE+ in an EaganMatrix patches settings.

[SteveElbows]

And why I was keen to point out that MPE+ messages compared to MPE messages are not really giving us something extra in this respect. If developers actually take heed of every point about timing that Haken make on their MPE+ page then that could actually give us something extra, but thats not because of the differences in midi data between MPE and MPE+.

Actually I'm being technically inaccurate with those comments. Because MPE+ does include the ability to send a few midi messages that tell the synths what the 'effective sample rate' of each expressive dimensions messages will be from that controller. And synths that take account of such things will presumably have a more sophisticated smoothing strategy compared to traditional approaches to this issue.

In practice I have not yet studied Osmoses MPE+ output to see if it is also sending those messages, and if so what values they are giving, whether they are the same as the Continuums or not.

When it comes to the Continuum, Haken say things like this about these things:

The effective sample rate for the most popular MPE controllers’ Z data is ~25 Hz, and ~500 Hz for the Continuum Fingerboard. MPE+ lets the controller specify the best amount of smoothing to avoid update noise (aka zipper noise or aliasing) but still retain the finger motion information in the data. The Continuum internally has an effective sample rate of ~3 kHz for Z; this data is processed to create the optimized ~500 Hz effective sample rate control stream for Midi transmission.

NRPN 100 specifies the low-pass filter cutoff for X (bend data) in 2 Hz units (eg value 60 means 120 Hz cutoff, corresponding to a 240 Hz effective sample rate).
NRPN 101 specifies the low-pass filter cutoff for Y (front-to-back finger position) in 2 Hz units.
NRPN 102 specifies the low-pass filter cutoff for Z (finger pressure) in 2 Hz units.

X low-pass filter cutoff at 120 Hz (CC 99 value 0, CC 98 value 100, CC 6 value 60)

Y low-pass filter cutoff at 60 Hz (CC 99 value 0, CC 98 value 101, CC 6 value 30)

Z low-pass filter cutoff at 250 Hz (CC 99 value 0, CC 98 value 102, CC 6 value 125)

From https://www.hakenaudio.com/mpe

[pdxindy]

I am very much urging caution in regards the assumption that giving other synths the ability to handle MPE+ is going to be the key difference maker.

No caution is needed... if my current assumption turns out to be incorrect, I'll change my assumption.

Other parts of the puzzle involve the synthesis features particular synth engines have, taking care to craft appropriate patches that make best use of the expressive control surface to control the aforementioned features of the synthesis engine in a way that feels musical, expressive, nuanced and immediate.

My thinking is thus: if it were only a matter of sound design, there would already be impressive sound demos using common plugins. If I had my Osmose and it was setup and connected to my audio interface, it would only take me a couple hours to design some sounds with my favorite soft-synths that took advantage of it as a controller. The question would be settled quickly.

Also, u-he has had their Osmose(s) for a while now and I doubt they would be excitedly talking about supporting MPE+ in their plugins if they were already having similar results with MPE mode and their plugins as is.

[SteveElbows]

My thinking is thus: if it were only a matter of sound design, there would already be impressive sound demos using common plugins. If I had my Osmose and it was setup and connected to my audio interface, it would only take me a couple hours to design some sounds with my favorite soft-synths that took advantage of it as a controller. The question would be settled quickly.

Also, u-he has had their Osmose(s) for a while now and I doubt they would be excitedly talking about supporting MPE+ in their plugins if they were already having similar results with MPE mode and their plugins as is.

Well lets try to retrieve something from the situation of me making a bore of myself, by steering the conversation more to the details of sound design for the Osmose (and other MPE controllers where relevant).

What sort of synths do you like for MPE patch creation? MPE X pitch stuff pretty much takes care of itself. And MPE Y is obviously different on the Osmose compared to other MPE controllers, in many cases its going to be used for final aftertouch stuff not dissimilar to what people often use traditional keyboard aftertouch to affect. I've gotten plenty of mileage out of doing pretty simple stuff with MPE Z pressure. But Osmose invites me to go further than I've tended to do in the past, especially if I am trying to rise to the challenge of 'competing' with what you've heard people doing with the built in synth. Theres lots more I want to say about this but as I'm sure you've noticed I struggle to keep my posts down to a reasonable size, so I will leave it at that for now.

As for u-he, I dont know if you have seen more of them talking about the Osmose than I have. I saw a few very short but enthusiastic posts around the same time as they first got their hands on the Osmose, and nothing since. So I dont have much to go on in regards what they have actually tried with it so far. Beyond that I only know that Howard Scarr started exploring sound design using the EaganMatrix itself, but there could be all manner of other things they are doing that they dont want to talk about publicly at this time for all I know. They certainly know far more than I ever will about whether any aspects of certain DAWs and plugin standards or their own plugins code gets in the way of harnessing the maximum amount of temporal resolution that comes from either flavour of expressive MIDI from the Osmose.

[ThomasHelzle]

All I can say is: "There is no spoon"...

All this theoretical assuming of what may or what may not be how relevant for this or that or who may do what is indeed becoming tedious.
If MPE+ turns out to be irrelevant, so be it. Let's see what over time will evolve.

Until then I'll rather play a tune on the Osmose.
This video captures more of the soul of things:

https://www.youtube.com/watch?v=3PMITkwyQPA

Cheers,

Tom

[SLiC]

This encapsulates the beauty of a 'self contained' MPE+ instrument- it just works, you can plug it in a PA and play; Its an instrument.

Beautiful.

[SteveElbows]

If MPE+ turns out to be irrelevant, so be it. Let's see what over time will evolve.

But I'm not claiming that MPE+ is irrelevant! It was just a discussion where I was very keen to point out that I dont think the lack of MPE+ supporting synths is the reason why we havent seen great demos of the Osmose being used as a controller with other synths. All my subsequent tedious waffle flowed from that very basic point. If I had realised that my point would be repeatedly misconstrued then maybe I would not have bothered. I tried to clarify what I meant by getting into extreme technical detail but it seems this made the misinderstandings worse, not better

To underline the point that I dont think MPE+ is completely irrelevant, I'll mention again that I did take the time to find a VCV Rack module that supports MPE+, and when that mode wasnt working I took the time to go through the code and find out what the issue was and fix it. And the author of the module was kind enough to publish an updated version that fixes its MPE+ mode. So if there is anyone out there who wants to experiment with MPE+, they can do so already in that environment. moDllz is the set of modules you need, v 2.1.2 or higher.

Meanwhile I think its normal that things have quietened down in recent weeks, once the initial flurry of people receiving their Osmoses had subsided. I am very much looking forward to the next batch ending up in more peoples hands, every time that happens we are going to see people demonstrating what they've been able to do when playing the Osmose. It will inevitably take longer for this to lead to more demos of people using the Osmose as a controller, but it will happen eventually.

[SteveElbows]

I do find it hard to know which demos to pick to mention on forums because I like so many, and peoples tastes vary.

Restricting myself only to stuff posted in recent days, I like this one:

https://www.youtube.com/watch?v=AQrsjhfWT_M