Improved playback using Noteperformer and other external programs

• Dec 5, 2015 - 19:59

Musescore's playback sound quality has mproved greatly with the addition of Fluid sound fonts, but as always things could be better. Other score creation software like Sibelius still do a better job with rendering sound, and many Musescore users are forced to export files to Sibelius to get the best possible score playback. One way around this problem is for Musescore to provide some level of compatibility with NotePerformer or other external sound synthesizer packages. It could be as simple as getting the Noteperfomer people to make a version of their program that can read Musescore files, or a more direct link like Sibelius can do. Well, something to think about....


Comments

I'm not familair with Noteperformer, but assuming it can communicate with Jack, then MuseScore already supports it - just set MuseScore to use Jack for playback in Edit / Preferences / I/O (and of course install and configure Jack appropriately if yuou have not already). Jack makes it possible for MuseScore to playback using most MIDI software, so if there is a program you like that produces better sounds than you can find in soundfont format, you are already welcome to use it.

In reply to by Marc Sabatella

Thanks for the reply. I believe that Noteperformer ( http://www.noteperformer.com/ ) goes far beyond simple MIDI playback, and instead uses a more advanced analysis of the score to do its sound rendering. I've heard the results and it is quite nice...but I'm not going to buy a copy of both Sibelius and Noteperformer to play my own works quite yet...too many $$$$!! Noteperformer itself is relatively inexpensive, and I would buy it if it could directly work with MS.

To use NotePerformer, I believe MS would need a tighter integration method that Jack offers. I would recommend that Musescore programmers have a chat with the NotePerformer people to see what would be needed to make the two programs work together. This might be a way for Musescore to finally get world class sound rendering.

-J

In reply to by jayW

Probably the most important first step would be for them to make their product free and open source, also more fully cross-platform (eg, Linux). In general, one should not expect developers of free and open source software to go too far out of their way to support proprietary third-party programs. Although to be sure, we do support some proprietary formats for import.

I've often thought there would be a market for a program like I imagine Noteperformer to be. Hoepfully someone with a commitment to free & open source software comes up with an alternative, perhaps using MusicXML as the interchange.

In reply to by Marc Sabatella

Actually NotePerformer is a VST plugin, which is a standard interface. Support for VST 2.0 or maybe even 3.0 could be justified on the basis that it provides access to a large number of libraries and effects for rendering. Whether these are free and open source or not should be irrelevant unless there are licensing (GPL?) restrictions, in which case maybe MuseScore should review the license, rather than expecting companies to abandon their revenue stream. It costs a lot of money to hire recording studios, engineers and performers to generate sample libraries.

But even then, I think the opportunity to do something much better than MIDI exists. NotePerformer has to resort to numerous tricks in order to properly render articulations. It measures the time between MIDI note on and note off events to try and figure out whether a note is staccato, or legato etc.

This should be deterministic. There needs to be an interface where all the metadata of the notes are communicated to the rendering engine.

When rendering a note, the synthesizer needs a sample for the attack, for the sustain, and for the decay. Without time-travel, it's impossible to know when to start playing the decay sample for the end of a note - it has to happen prior to receiving the MIDI note off signal. Either MIDI needs a mode where you can specify the entire duration of the note at the start, or we need an alternative.

Seems to me with MuseScore getting ever more mature, it could be time for someone to think up a new system for this which would allow much better rendering than MIDI does.

In reply to by Marc Sabatella

Hi, Marc!!!

I know this question is not in the main issue, here, but...

I don't know to use JACK, with MuseScore (at least).

I want to know if you are using JACK and if you have some "final" opinion if the sound playback quality is better than when we use MuseScore without JACK. I'm talking about the expression (legato, etc).

Do you? ???

In reply to by jotape1960

Connecting something else to MuseScore via JACK doesn't magically make MuseScore recognize markings for playback that it wouldn't otherwise. So if, say, slurs don't have any playback effect within MuseScore, then they won't in the DAW or other program/device you are connecting MuseScore to - we aren't sending info we can't use ourselves. But things like staccato that work within MuseScore, will work in the DAW, because we already do the work of turning that into appropriate MIDI messages internally.

Not sure how the new single-note dynamics coming in MuseScore 3.1 will work - this uses CC messages and I don't know if they go out via JACK or not. Probably they do, but I haven't tested it.

In theory, one could program up an implementation of slurs that used the legato controller or other MIDI message and it would work in the synthesizers that support this, and I don't think Fluid normally does (my memory and understanding are extremely fuzzy on these points). So I could imagine some future in which you can get playback of some marking via JACK & DAW software that you can't get internally. But I'd also imagine if we went to the trouble of implementing the necessary controller messages, we'd also implement them in our synths.

Do you still have an unanswered question? Please log in first to post your question.