Instruments are not playing their correct sounds

• May 4, 2021 - 21:58

I have this symphonic score that I'm working on and the woodwinds and brass are not playing their correct sounds. They only play string sounds, and, interestingly enough, they aren't just playing arco; some play pizzicato and, some, even tremolo. I tried to "swap" out the instruments and, despite doing so, they still play the string sounds/techniques. Does anyone have any suggestions on what to do, because I'm heavily relying on the sounds of the instruments so that I can "conduct" the performance of the piece the way that I want it to be "performed". I posted the file as well.

Attachment Size
Also Sprach Zarathustra.mscz 177.95 KB


That's a really big score. Difficult to pin point the problem with a quick look, but indeed some of the wind instruments are playing string sounds, and some are doing pizzicato, some are arco.

This is a great example of where the user interface of MuseScore could be improved a lot. I managed to see most of the instruments in the mixer - but the mixer spread across about 4 screen fulls of a 27 inch iMac. There's a lot of wasted/useless space in the mixer. Sliders look nice for small scores, but for something this big knobs or even just digital numbers would take up much less space. Instead of having around four screenfuls of sliders, it would probably be possible to get several hundred knobs on a screen. See the rather crude example here. A good designer should be able to get more than a hundred nice looking knobs on a screen. Here there are about 50. Also, they don't all have to be visible at once - they could be in banks - so grouped together, but with their visibility and position variable.

Knobs for mixer.jpg See possible example - just a mockup.

Trying to scroll around the score is also difficult - though that might be a limitation of my mouse and setup.

I tried to see if the instruments are assigned to the "correct" sounds - but that was inconclusive.

Are you using the standard soundfont?

In reply to by dave2020X

I am using the standard font (well, the HQ one) and can reproduce the issue.
Locking into the mscx shows no traces of a non-standard font.
Piccolo uses port 4 and channel 8, no other instrument does.
The score uses 132 instruments, but 311 channels and ports (due to some instruments having more than just one). that is clearly more than 256, so there got to be duplicates, and quite many.

Edit: no there are not, the score uses ports 0-20, so 21 ports!
Not sure whether that is legal though, the Mixer doesn't allow for more than 16

This code seems to indicate that 16 is the max. for either:

                              int nm = getNextFreeMidiMapping();
                              midiPort    = nm / 16;
                              midiChannel = nm % 16;

In reply to by Jojo-Schmitz

Interesting - up to a point. Seems to get too technical for me - but there are other versions of this on the site which seem to manage with fewer instrument sounds in the mixer - and which don't have this particular problem. Also I was surprised about this - as I would have expected this score to not be in the Public Domain - but it seems that it is in some countries, and is perhaps not in some others.

Is the current one a new version by the OP, or adapted from one of the other versions which are shown in IMSLP doesn't indicate a copyright problem though, so perhaps it is OK in most countries now -…

It does look as though some technical limitations have been exceeded. Probably MS should refuse to process this, or give a very strong warning - which would then allow users to fix or avoid the problems.

The overall message seems to be to reduce the number of sounds (channels, VIs, articulations - whatever) required to represent the instruments in the score. It may not be necessary to duplicate sounds for each scored instrument. I think that can be done in MS, though I'm not sure how. It is something which is possible in many DAWs - though again I don't know how exactly.

In reply to by Jojo-Schmitz

I reordered the instruments ports/channels so now they only use Port 1; Channel 2 - Port 12; Channel 14 and I placed all of the instruments that I didn't want to make a sound on Port 1; Channel 1, so now the total amount used now are 189 ports and channels. Unfortunately, this did not work and the problem still persists. I am using the default font in Musescore as well.

It took about 1-hour to reorder each instrument (and their articulations/techniques) so their port and channels don't overlap and it was all for nothing... I think I'm gonna go cry now ',:')

Edit: Updated the file so you don't have to zoom in 800% just to see what's going on, but now it takes a few seconds to scroll to the bottom. I swear I should have a record or something because now the score would be about 4ft tall if I printed it out.

And to answer a possible future question/s. No. I can't just remove all the extra instruments. I need to be able to control the individual dynamics of each instrument to get accurate sounds and to get the interpretation of the piece that I want. Plus it's helping me to start working with the confusing mess that is DAW's and so I can start using shortcuts so I can start writing music in Musescore as fast as I can type words.

In reply to by LukeyDupti

Sorry the fix so far hasn't worked.

Still sounds a bit like a technical problem in MS, but if you could reduce the number of instruments or channels it might help. That's beyond my level of knowledge and expertise in MS - so keep pushing the technical experts.

Re DAWs - there is a way of working in which tracks are grouped together. This still allows individual tracks to have levels adjusted, but also allows control of groups. It is perhaps unlikely that you'll want to micro-manage the levels of all the instruments all the time, so maybe some group level control would do that for you.

Example - you want a trumpet to sound out over other brass instruments. You have a string accompaniment which starts off quiet. You want to turn all the brass down so as not to swamp the strings - but the trumpet should still stand out. You don't need to control all the individual instruments. You might want to have short periods of swell in both the brass and string sections - not necessarily at the same time.


Unfortunately I don't think the Mixer in MS works in that way at the moment - but my point is that grouping might reduce the number of instruments (or channels) you think you need.

In reply to by LukeyDupti

"No. I can't just remove all the extra instruments. I need to be able to control the individual dynamics of each instrument to get accurate sounds"
I know what you are trying to do. I think the problem is that it's not going to work. You think that if you have 50 string parts that it will sound like a full orchestra. But there is only one solo violin sound in the default font. So if you load that sound for each player in the 1st violin section, all you get is that sound louder, not the sound of a section. Why? A string section sounds like it does because it is made up of, in your case, 16 different players that all play differently. Their vibrato and tone quality are not the same. Even if you pan them to 16 different places, it will always sound like one player, just louder.
And consider that even if you succeed in getting all 130 or so instruments to sound just the way you want , it will only do on your system and to your ears.
Not to mention that there is no solo horn or trumpet sound.
Good luck.

In reply to by bobjp

The reason is more to do so with a limitation with Musescore. You can not control the individual dynamics of the voices on a single staff and even on a split staff. Parts/phrases that are supposed to be emphasized won't be able to without making another phrase to that isn't supposed to be emphasized stand out (and vise versa). I'm very well aware that no matter how many instruments that I add it will always sound the same. That's just a limitation of virtual playback. Like I said before i'm doing it more so for a detail thing and so that the end result isn't so stuck together like it normally would. It also is a aid into what parts would sound like in a more realistic setting (ex. 1 player would have to play twice as loud to match the volume of two players)

I wasn't planning on doing this at first, but I see it as a kind of challenge to help understand the piece as a whole and how the individual parts work together (and to learn the shortcuts lol). To sum up this whole situation, I'm putting in extra to get extra.

In reply to by LukeyDupti

"(ex. 1 player would have to play twice as loud to match the volume of two players)"
At moderate levels that wouldn't present a problem. Depending on how you measure it - that's only a 3 dB (6dB) difference. More likely issues would be that one player amplified just does not sound like two players - due to a whole bunch of interference effects. A string section sounds different form a single violin even if playing exactly the same set of notes.

However, conductors don't normally go down to the level of individual players, unless there is an obvious problem, which with a professional orchestra is unlikely.

In reply to by LukeyDupti

Sure, break up the string parts. For example, you could have two or three violin 1 staves. As you know, there are three group violin fast expr. sounds. Violins, 2 violins, and strings. They all sound different and work well in different situations. Strings fast expr. can be applied to any string part. You can control the dynamics and phrasing of the two or three 1st violin parts you have independently.

In reply to by Jojo-Schmitz

Yes. There is also Strings. The OP hasn't said yet, but I'm guessing that part of the reason for 50 string parts is to mix individual players just like a real orchestra. The problem is that there is only one solo sound per section in the font. In my experiments, a large number of violin staves playing the same part doesn't sound like a violin section. It just sounds like one player playing very loud. And maybe that is the effect the OP wants. He wants to be able to control the volume of individual players. Heck, even a conductor can't necessarily do that in real life.

This is, of course, the beauty of the virtual orchestra. I get it. We have much more control. But because there are channel problems, I might suggest a bit smaller scale tests to work out the problems.

And maybe I'm way off base. I have downloaded the score. I started to assign sounds. It's just such a large project.

In reply to by bobjp

I'm not trying to reproduce the sound of a violin/string section. I'm trying to accurately reproduce the sound levels of the whole orchestra. 1 player will never be as loud as a whole section and a whole section will never be as quiet as a single player. That's all that i'm trying to do, and not just for the string section, but for the whole orchestra!

The other reason is because this particular piece doesn't just split the players into a regular divisi, it splits them into desks and the players often play at different levels and as I said before, Musescore doesn't allow you to adjust the dynamics of a voice on an individual staff so this is kind of a workaround (a really really big one at that).

In reply to by LukeyDupti

I think there are a few points to note. Firstly you may have hit a limit in MS. Secondly you are trying to do something which is probably completely unrealistic. Conductors don't normally try to micro-manage every note played by every player. If there are solos conductors might try to encourage players to play in particular styles, and they may also balance instruments in sections, but mostly they will I believe trust the players to do something reasonable.

You also mention trying to "accurately reproduce the sound levels of the whole orchestra" - and I think that is an unrealistic expectation. There are very few systems which can get anywhere close to that, though I have been to opera productions where the offstage brass instruments were actually playing in a room in another building and the sound in the auditorium was delivered by very high power and quality amps and speaker systems.

I would also like MS to be able to do better, but it is what it is, and some things are not possible now, and some things may never be possible.

In reply to by dave2020X

I've halted the process of interpretating the performance of the score and the detailing of all the dynamics. For now, i'm just purely focusing on the transcription (since I can't really do anything besides that) which will take a few months. Hopefully by then the issue will be fixed, or Musescore 4 will be out, and I can resume work on the interpretation/performance.

I do want to mention–I said this in a previous comment–that I put all of the instruments that I didn't want to play on the same channel and port which brought me out of this "limit" and the sounds still are playing incorrectly, so I believe that this more than an issue about "limits" and may be the discovery of something more adverse.

In reply to by LukeyDupti

> "I do want to mention–I said this in a previous comment–that I put all of the instruments that I didn't want to play on the same channel and port which brought me out of this "limit" and the sounds still are playing incorrectly"

Would you mind sharing this file for closer inspection?

In reply to by jeetee

All of the brass and wind parts that are showing, except the ones with small staves, are all on the same port and channel. You can click any of the notes on them and they'll make any of the three string technique sounds (arco., pizz., trem.,). The small staff brass and woodwind, however, each have their own port and channel, including the sub-channels for brass, and, despite this, still play string techniques. I left the small staff flute and trombone staves shown so you can test, but, if you want to test any of the other instruments, just unhide them. The ones with small staves are the ones that will be playing the actual sounds in the finished transcription, and the normal ones are for decoration/representation of the notes.

Attachment Size
Also Sprach Zarathustra.mscz 244.95 KB

In reply to by LukeyDupti

OK then. Let's consider just the 1st violins for a moment. You have 8 desks and 16 players. Is there ever a time when all 16 are playing together at 16 different dynamics? I'd be surprised. Could the number of desks be reduced to 8 or even four? At least just to get the idea working? Perhaps think about figuring out the minimum you need. Then expand it. The original idea is intriguing, but might be beyond MuseScore.

In reply to by bobjp

I'd be surprised if there is ever a time when different violinists are playing at the same dynamics! That's just one reason why groups of vioinists sound different. However it is likely that all the violinists in a group are trying to work at the same dynamic level.

Another thing to consider is that sometimes in a large orchestral texture some instruments playing certain notes just become inaudible. Having more instruments with micro-managed dynamics for most purposes would seem to be not only impractial but also undesirable.

In reply to by dave2020X

This has nothing to do with dynamics.
It is about the tuning of each instrument (which differs when human players play the same note) and the effects of this difference on each other. //Of course, the different vibrato pitch/rates of each player also affect this.

I don't recommend following the method in this example while you already have a good Violins or Strings patch available. I wanted to show this just to explain it technically.

It is easy to simulate. Create 4 violins. Enter the same notes in each. Select each of their staff individually and set (cents) from the tunings area in the Inspector to the following different tunings:
Violin (1): 0
Violin (2): 5
Violin (3): -7
Violin (4): 19
(These 4 instruments with different tunings are equivalent to a group of 6 violins)

Listen and review the attached file.

An effect will be similar to the sound of the Violins instrument group. The only problem is that they play the same vibrato at the same speed and at the same rate. : /

// The timbre of each instrument (having different waveforms and different phases) also has an effect on this, but we cannot simulate it right now.

PS: Just adjusting this setting is not enough to simulate a group of strings. The demonstration here is only to explain one of the important factors.

Attachment Size
example-z01.mscz 15.03 KB

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