What
is MPE?

Everybody can use MPE: Playing with MPE features is much easier to understand than learning a new instrument. It is because MPE builds on those intuitive gestures that most people naturally make anyway when interacting with musical instruments.

MPE finally takes all the nuanced gestures that a player makes into account for sound generation that were ignored by most electronic instruments. The benefit is immediate: You don’t have to be a killer pianist or a rocket scientist to use MPE: just playing with the pressure and pitch of one single note can already be a gratifying experience.

A standalone instrument like Osmose allows you to start exploring directly. All you need are headphones or loudspeakers to have your first MPE experience. Admittedly, it gets trickier on the configuration side when you start connecting multiple devices and software to use MPE across your music production or performance setup.

To make MPE work across your tools, there may be some additional technical steps to consider that aren't necessary in non-MPE scenarios. That said, several major music production applications (e.g. the DAWs Ableton, Logic, Bitwig, Cubase, Studio One) already offer great workflows for MPE, and more and more developers of software and hardware instruments add MPE support to their products. As long as MPE is supported by the tools you are combining, you only have a few clicks to do and will enter a fantastic new playground. With this page, we try to help you with this.

MPE does not require MIDI 2.0. Instead, it cleverly extends the MIDI 1.0 protocol without the need to introduce new kinds of MIDI messages.

MIDI 2.0 represents a fundamental evolution of the MIDI protocol, building expressive control directly into its core. MIDI 2.0 introduces a host of new capabilities that go far beyond what MIDI 1.0 could offer.  It supports per-note controllers natively, offers dramatically higher resolution for pitch and dynamics, and enables bidirectional communication between devices. This means instruments and software can exchange profiles and properties, allowing for automatic configuration and deeper integration.

The relationship between MPE and MIDI 2.0 is both historical and functional. MPE serves as a bridge — a workaround that proved the demand for more expressive digital performance. It showed that musicians wanted to bend, slide, and shape individual notes in ways that MIDI 1.0 couldn’t natively support. MIDI 2.0 takes that vision and makes it native: instead of relying on multiple channels to simulate per-note expression, it introduces clean, scalable mechanisms that handle it directly. While MIDI 2.0 is designed to be backward-compatible with MPE, it ultimately supersedes it by offering a more elegant and robust framework for expressive control.

At Expressive E, we seek to introduce MIDI 2.0 to Osmose as soon as DAWs support it.

While any DAW can receive and send the MIDI messages that MPE uses, not all DAWs support MPE. Only a full MPE support guarantees that MPE data is correctly interpreted and can be comfortably edited. Here's a more detailed explanation of our table above to learn which DAWs already support MPE.

DAWs with MPE support:

• Ableton Live (version 11 and above): Extensive MPE support with individual note editing and expression capabilities. Ableton offer their own Ableton Push 3 controller that is MPE compatible as well.
• Bitwig Studio: One of the first DAWs to offer deep integration with MPE, Bitwig allows intricate control and modulation of individual notes since version 2.1.
• Apple Logic Pro: MPE-compatible with advanced support for editing and recording expressive MIDI data since version 10.5.
• Steinberg Cubase/Nuendo (version 11 and above): Cubase supports MPE and provides tools to edit per-note expression as part of their "Note Expression" feature.
• PreSonus Studio One (version 5.2 and above): Offers complete MPE support for expressive note control and editing.
• Tracktion Waveform (version 11.0 and above): started with basic MPE support and offers fully fledged per-note expression editing in version 13.5. Their Waveform FREE is a great way to begin with MPE music production.
• MOTU Digital Performer: MOTU added full MPE support in version 11.

DAWs with limited or no MPE support:

• ImageLine FL Studio: As of now, FL Studio does not have full MPE support. It can process some MPE data, but not to the extent of the DAWs mentioned above.
• Cockos Reaper: has flexible MIDI routing to allow working with MPE controllers and synths for a long time already, but offers no comfortable per-note editing or overdubbing of MPE performances.
• Avid Pro Tools: No MPE support at all. ProTools has a traditional focus on audio rather than MIDI. Compared to other DAWs, ProTools' MIDI routing is complicating workarounds for MPE setups. Also, their proprietary AAX plugin format is not designed to handle MPE.
• Propellerhead Reason: The architecture of Reason's "Rack" system wasn't designed for MPE. Manual channel routing is labor-intensive and there are no workflows for per-note expression editing.
• BandLab Cakewalk: Allows flexible MIDI routing for basic playing with MPE-enabled plugins and instruments, but has no native MPE workflow with per-note expression editing.

MPE - Ultimate precision

Thanks to its generous key travel and superior sensor technology, Osmose’s keybed detects everything from the slightest brush on its keys to the most intense performances.
Internally in Osmose, some per-note expression is transmitted with dramatically higher resolution—from 128 to over 16,000 steps—for ultra-precise control. 

Traditional MIDI - standard precision

You need to hit a certain threshold for triggering notes in the first place, and softer notes might not be registered. Only 128 steps of velocity intensity are registered (7-bit resolution).

Different DAWs differ in their MPE workflow, but there are some general steps to consider when following your expressive playing all the way to the sound that is generated.

Check compatibility:

• Check that your MIDI controller supports MPE. This is vital to translate your gestures to MIDI messages.
• Check that your DAW supports MPE. The type of MIDI messages used in MPE doesn't differ from those traditionally used in MIDI and can be handled by any DAW. However, only MPE support ensures that your DAW interprets the data as intended, allows overdubbing, and presents you with a logical MIDI clip, along with the tools to edit it ergonomically. 
• Check that your software instrument plugin supports MPE. Only then can every single note be shaped independently according to your playing.

Enable MPE: In your DAW, ensure that the MIDI port of your MPE controller is activated and set as the input source for the track containing your MPE-capable software synth. Often, you will also need to separately enable MPE for the track on the DAW's end as well as in the plugin's GUI itself.

Adjust Sensitivity: Depending on your instrument, you might need to adjust the sensitivity of pitch bend, slide/aftertouch, or other expressive controls to match your playing style.

While MPE utilizes only traditional MIDI messages that can be sent and received by any MIDI-compatible device, your tools will not react as expected if they don't support the MPE standard.

However, most MPE controllers can be configured to exhibit similar behavior to that of a classic piano/synth MIDI controller, disabling per-note expression to ensure downward compatibility.

The External MIDI Mode of Osmose goes further and offers various MIDI configurations that allow you to switch from MPE to classic MIDI, to Poly Aftertouch, or Multichannel workarounds in a blink. We think MPE is awesome in many contexts, but rest assured: using Osmose, the expressivity adapts to you when you need it, not the other way around.

For example, any monophonic synth, even if it doesn't support MPE, can be injected with new life when being played from an MPE controller as flexible as Osmose. Since you don't have to worry about the "polyphonic" aspect of MIDI Polyphonic Expression, you can make the synth react to pressure, aftertouch, and pitch bend gestures.

The wish to make electronic instruments more expressive is as old as electronic instruments themselves. Inventions like the Theremin, the Ondes Martenot, and the Trautonium all aimed to achieve a closer, continuous connection between the player's gestures and the resulting sounds as early as the 1920s and 1930s, also trying to overcome the physical limitations of acoustic instruments. Back then, synthesizers were monophonic and standalone instruments that couldn't be easily connected to each other.

To popularize a new technology and make it approachable for users, it is key that manufacturers of hardware and software  agree on a standard that they follow to ensure their tools work together seamlessly. This was the motivation behind the MIDI standard (Musical Instrument Digital Interface), which lets electronic instruments and computers communicate by sending and receiving data (notes, timing & dynamics). MIDI remains as relevant today as ever, despite being introduced already in 1983.

"MIDI Guitar Mode" could be considered the first precursor to MPE. Guitar Mode was never formally standardized but organically emerged shortly after the release of MIDI as a workaround for guitar synthesizers that required a method to allow for polyphonic (per-note) pitch bend, as opposed to the global (per-sound) pitch bend suggested by standard MIDI. It did this by granting every guitar string its own MIDI channel. Roland's GR-700 Guitar Synthesizer was the first to utilize it in 1984, but the concept persisted decades later.

Since the 1980s, computer engineer Lippold Haken has been working on a controller with a continuous neoprene surface that overcomes the segmentation of the traditional piano keybed, with its predefined, absolute pitches, and puts a clear emphasis on a delicate playing feel with fast, accurate sensing of the player's gestures. The first publicly available version of his Haken Continuum Fingerboard was released in 1999. The Continuum also allowed for polyphonic performances, sending the fingers’ X, Y, and Z positions (pitch, front-back, and pressure) to an external synthesizer via MIDI messages on rotating MIDI channels. This was essentially the same method that was later adopted as MPE.*

In the 2010s, notably Roger Linn and Roland Lamb pushed to gather support from the industry for a common communication protocol with the name "Expressive MIDI", "Multidimensional Polyphonic Expression" or "MIDI Polyphonic Expression" in an effort to introduce multidimensional controllers and instruments like their Linnstrument and ROLI Seaboard into the mainstream. Following this collaboration, the specification of MPE as it is used today was formally proposed in 2015 by Eigenlabs, Haken Audio, Madrona Labs, Moog Music, Keith McMillen Instruments, Roger Linn Design, and ROLI, with support from Apple and Bitwig.

MPE was officially adopted as a standard by the MIDI Manufacturers Association in 2018, marking a significant milestone in the success of MPE. Nowadays, more and more music production software, as well as software and hardware instruments, support the standard.

Since 2022, Expressive E's Osmose has helped push the boundaries of expressive playing and popularize MPE in music production and performance with its proprietary advanced mechanical keybed, which stays true to the form factor of traditional keyboards, making it one of the most accessible MPE instruments available.

*Haken actually uses a more sophisticated protocol that, after the release of MPE, was retrospectively named MPE+. It is also used inside Osmose. Read more about it here.

When working in a classic MIDI environment, all MIDI messages associated with a sound are sent on a single MIDI channel. Any modulations (such as pitch bends or changes in timbre) are sent per channel and thus affect all notes of that sound in the same way.

A so-called MIDI port offers 16 different MIDI channels. MIDI ports can be either input or output ports, and they are used to route MIDI messages between all your gear: Not only your MIDI controllers and hardware synths have MIDI ports, but also your software instruments / tracks inside your DAW.

In traditional MIDI, you would typically use the 16 different channels of a MIDI port to trigger different sounds, e.g., have a piano on channel #1, a guitar on channel #6, a synth pad on channel #7, and some drums on channel #10.

MIDI Polyphonic Expression (MPE) build on the 16 channels that traditional MIDI already established, but instead of using them to trigger different sounds, it reserves all of them for one single sound. One general channel is kept for global messages (typically channel #1), and then every single note that you play is granted its own MIDI channel in a round-robin manner (typically channel #2 to #16).

This way, per-channel modulation can be applied to each note individually, giving you much more control over your polyphonic performance. 

You already know your way around MIDI a bit? This table will help you understand how the MIDI parameters you know from traditional MID are implemented in MPE:

* when sent on the Global channel (typically channel #1), the pitch bend can also affect all notes at once. This is triggered by the pitch wheel/slider/touchstrip of your MPE controller instead of gestures of your fingers.

Yes, setting up Multi-channel MIDI is a common workaround that enables a non-MPE synth to act/sound as if it supports MPE. However, you need to do some manual setup to get it working, and editing this MIDI data will be somewhat uncomfortable. Also, wide pitch glides are only possible within the limits of the respective synthesizer. Not all synthesizers support the +- 48 semitones maximum pitch bend range used as the default in most MPE scenarios.

The workaround is as follows:

In your DAW or your multitimbral synthesizer, you need to copy the same sound as many times as high you want the polyphony to be. Each of the sound’s copies will play only one note and needs to be set up to react to only one MIDI channel. This enables the shaping of every aspect of the sound for each note individually.

This copy procedure can be done either by duplicating the same instrument track multiple times. Alternatively, if you're using a DAW and the plug-in you're working with features a multi-channel/multi-timbral mode (like Native Instruments Kontakt), you can duplicate the same sound within a single instance of the plug-in.

Your MPE controller must allow for a Multichannel MIDI mode. On Osmose, you would select the "multichannel" config preset in config menu of the External MIDI Mode. After this, you set the range of MIDI channels that shall be used, e.g. channel 1 to 6 for six voices of independent polyphonic expression in the adjust menu. Osmose will now send each new note on a new channel.

By the way: Osmose’s Multichannel mode will also work if you want to control synths that support the so-called ‘MIDI Mono Mode’ across several channels, used initially for MIDI guitar applications.

A DAW that does not support MPE can still record and play back the MPE messages sent from MPE controllers or instruments. This is because MPE uses standard MIDI messages, just like traditional MIDI. Just be sure to record and playback all MIDI channels on your MPE tracks, rather than filtering the channels down to a single channel.

However, without proper MPE support, the DAW won't be able to interpret the messages correctly, and you will end up with one or several of the problems below regarding MIDI editing and overdubbing:

• Note Expression cannot be edited in place, because your DAW doesn't understand that those MPE messages belong to a single note, not to the entire part.
• The visual representation of your performance in the editor is very intransparent, because notes are misinterpreted as not belonging to one performance, but to different instruments. 
• Adding notes manually during editing or overdubbing is hardly possible, because your DAW doesn't know it has to re-allocate MPE notes across the channels upon every replay instead of just leaving them on the channel they were recorded on.

When sending MPE messages to a synth that doesn't support MPE, you will likely run into the following problems:

• the synth will trigger only every few notes or not at all, because it expects notes on its MIDI input to be sent on only one channel. You may be able to configure the synth to accept notes on all channels, but even then...
• Per-Note Expression will not work. The pressure, timbre, and pitch bend information intended to control individual notes will all compete to influence the global parameters of your synth, resulting in glitchy sounds due to conflicting messages.

You can check if your MPE controller offers a downward-compatible Monochannel Mode if you wish to play non-MPE instruments from it.

Touché and Touché SE don't support MIDI Polyphonic Expression (MPE).

It is because the controllers don't trigger notes themselves, but are meant as an addon that allows you to add expression to existing instruments in your setup. Since there is no possible way for Touché to differentiate which note of your real-time performance or a pre-recorded MIDI pattern should be affected, gestures on the touch-plate of Touché will always affect all the notes you play at once. It is expressive playing, but it cannot possibly tick the "polyphonic" box required for MIDI Polyphonic Expression (MPE).

This doesn't mean, however, that you wouldn't be able to use Touché (SE) as a MIDI class-compliant controller alongside your MPE-compatible controller/instrument. In this case, the MIDI CC messages sent by Touché (SE) could be used to manipulate the sound on a global basis, while messages from your MPE controller can remain individual per note.

Haken Audio’s EaganMatrix, the sound engine in Osmose, has been around well before the MPE standard was introduced. MPE+ is what Haken Audio calls their optimized extension to the MPE standard nowadays. If you wish to learn more, you will find more information on MPE+’s technical specificities here: https://www.hakenaudio.com/mpe

On the practical side, this help article explains what you have to bear in mind when sending MPE messages from your DAW or other MPE controllers to the EaganMatrix : Triggering the EaganMatrix from an external device