[MU4 Issue] Dynamic range between FFF and PPP is to wide (comparing with MU3)

[DmitryArefiev]

Describe the bug Dynamic range between FFF and PPP is to wide.

Should be investigated more. Probably PPP and PP can be increased a little bit..

https://user-images.githubusercontent.com/90187801/183133218-0b7feb37-1551-4069-a619-fdf0d36f8970.mp4

To Reproduce

1. Open dynamic test_MU3.zip
2. Playback in MU3 and in MU4

Platform information macOS, Windows, Linux

Additional context Add any other context about the problem here.

[Tantacrul]

When @bkunda is back from vacation, this should be investigated with him too. (Although, @DmitryArefiev - if you're happy to make suggestions, do. I can double-check also)

[bkunda]

I've just started looking at this and I'm wondering if there are any values someone could point me towards (volume level, I'm thinking) for both MS3.6 and the current levels in 4.0. This would help me make some more concrete suggestions. @DmitryArefiev @vpereverzev

From what I can hear, ppp seems to be too soft relative to pp, and there is not enough differentiation between mp and mf.

You can hear this quite clearly in a slightly reconfigured version of the test file given in the issue description (I've just changed the dynamics to step linearly from ppp through fff):

dynamic test_MU3-2.mscz.zip

[Tantacrul]

I would first find out how volume / dynamics is specified (is it a % value?)

Secondly, I wouldn't bother looking at V3. I'd be inclined to look at something like NotePerformer and then figuring out 1. if they're relative dynamics make more sense and 2. translate that into however our existing system is defined.

[Jojo-Schmitz]

The values are stemming from MIDI, 0-127, silence to full throttle. At least as far as the user value' is concernd, the 'offset values' are relative to whatever the current absolute value is

[Tantacrul]

Since MIDI is not our primary system anymore (we are MPE-first, MIDI second), I'm not sure 0-127 is still the way we define dynamic volume anymore.

For sure it is translated into this value when interfacing with MIDI samplers (Fluid, VST, etc.)

[bkunda]

@vpereverzev @RomanPudashkin My starting recommendations for this would be:

1. To increase the volume of ppp relative to pp, and
2. To make more a difference between mp and mf.

If there is another MPE-specific way of defining these relative volume/velocity levels, then that info might help me make some more concrete suggestions. On point 2, perhaps the velocities of all dynamic levels from mf-fff could be increased, or alternatively the levels of pp, p and mp could be decreased – in either case, to support more of a distinction between mp and mf.

Happy to discuss further in a quick call if that helps.

[Tantacrul]

@DmitryArefiev - would be good to revisit this.

[erinic04]

Here's a video demonstrating that at least with Muse Sounds, the dynamic range is too wide

https://user-images.githubusercontent.com/21023353/203227913-824e323e-fb85-4f11-b40a-1033c8fd9313.mp4

The Mixer should never look like this while playing notes, especially when it's only pianissimo.

(could be a combination of this and #13842)

[rgreen5]

See also, #14418.

[Tantacrul]

@DmitryArefiev - when you first logged this, I think it was for MS Basic only. Do you still think it's an issue?

(I think we should leave Muse Sounds out of this, since related issues are logged elsewhere)

[Tantacrul]

@bkunda

[bkunda]

I've re-tested this with this sample test score and I would say that points 1 and 2 of my recommendations above still stand, however I'm not sure the issue is so serious to be considered a release blocker. I think this is something that could be tweaked over time.

[Tantacrul]

Yeah, let's move to Post Release