V3.6.2 - All staves sound like piano

• Sep 4, 2022 - 19:56

Loaded a MusicXML file into MS and saved as a MS file. All of the staves sound as a piano no matter what they are labeled and what is showing in the Inspector. Tried resetting a few of them in the Inspector as well as the Mixer with no luck. I see from Internet searches that this problem goes back a long way and includes a number of break-fixes. Are we at another break point? Thanks.


All of the staves sound as a piano no matter what they are labeled...

Right click on the staff's 'label' (i.e., instrument name) and select Staff/Part Properties. Click 'Change Instrument' and choose the instrument that the 'label' states.
This will re-assign the correct sounding instrument to that staff.

In reply to by littlereg1

Check the Mixer: Most instruments (Bass, all 3 acoustic guitars, piano, upright bass and even drumset), have their sound set to Grand Piano.
At least here. Are you using any non-standard soundfont (in synthesizer)?

You very certainly did not do what Jm6stringer recommended and you claimed to have done... as that indeed would have fixed the sound

In reply to by Jojo-Schmitz

OK, I attached the zipped XML file. I see that all your instruments work in the file you sent although the parts have been changed. For example, the acoustic guitar is now an octave lower than in the original score. Are these two things related? I supposed I could go through the entire score and fix it but this shouldn't be necessary. However...the vocals are all still piano sounds for me - even from your file. Are you saying that the vocals are performing correctly for you? Thanks.

Attachment Size
CCM Retrospect.zip 496.2 KB

In reply to by littlereg1

The xml file stems from Sibeluis:

   <software>Sibelius 18.6.0</software>
   <software>Direct export, not from Dolet</software>

Try using the Dolet Plugin, that usually works much better

In that xml there's nothing to indicate which sound is wanted, nor which instrument it is:

   <score-instrument id="P1-I1">
     <virtual-library>General MIDI</virtual-library>
   <score-instrument id="P1-I2">
    <instrument-name>Soprano (2)</instrument-name>
     <virtual-library>General MIDI</virtual-library>

Because there is none (or because there are 2, "Soprano" and "Soprano (2)"?, MuseScore defaults to the very first instrument in the list: Grand Piano

Here's how it should be looking:

   <score-part id="P1">
      <score-instrument id="P1-I1">
      <midi-device id="P1-I1" port="1"></midi-device>
      <midi-instrument id="P1-I1">

In reply to by Jojo-Schmitz

I do not see this as the issue. Regardless of the source of the initial input, the application should be operating in its own system under the auspices of its own architecture. If the Properties of a staff are assigned to, say, "Trumpet", then there is no way that the music should sound a "Trombone". The question is not, "What are the nuances of the input XML file?", but rather, "How can MuseScore display one instrument and play another?"

What this means is that the application is architected without consideration of WYSIWYG. The Properties widget is not coded to check its display, which is typical of software development: one hand not knowing what the other is doing. (Users have only the UI as the means of communicating with a computer, yet UI considerations are consistently the least attended-to aspects of software, even after 40 years after the advent of the micro-computer.)

Once the import has finished, no matter how flawed that input data is, there is no reason why the user cannot manually adjust the settings in the tool once the file has been saved to standard format. After that save, the data schema should now exist in MS just as if it were entered by hand - the input file is now of no consequence, operationally. Thus, if instrument mismatches occurred during import, then users should be able to easily reinstate these in the UI. The fact that we can't is wholly an issue of MS, not Sibelius or XML format, and the fact that this "bug" (a warm-fuzzy term smuggled into our parlance in place of its proper term: "failure") has existed for multiple releases now means that MS developers neither know nor care what causes it.

In my case here, MS recognized most of the staves in the original score and assigned similar if not exact labels to them as well as the voices I had originally assigned. I can understand if there was "confusion" on the import, however, and that now cleanup is required. Once I went into the Inpsector or the Mixer, however, this "default" Grand Piano you mention should have simply gone away once I manually assigned the instrument I wanted.

There is never ever a valid reason why a user interface displays one thing and executes another. Never. In our case here, if MS insists on performing piano when the user has selected a voice instrument, then the button dropdown should display "Piano" or render a detailed error message as to why this is happening and what the user can do to fix it. If the user cannot fix, then a bug report can be written up and the issue addressed, but at least the UI remains consistent and trustworthy.

The fact that the button dropdown continues to display the incorrect instrument and that there is no help from the application is completely the fault of the application, not the XML import.

In reply to by littlereg1

MuseScore can't guess an instrument, just from something that is meant to be its dispay name (in any foreign language) and it can't guess which to take if there are multiple instrument names mentioned.
Garbage in, garbage out, Sibelius needs a better XML export here (and the Dolet plugin is exactly for this, the Sibelius builtin XML export is only to move scores from one Sibelius version to another!)
In those cases Grand Piano is the sane default/fallback.

In reply to by Jojo-Schmitz

If you read my previous note, you would see that - more than once - I agreed that the import might have flaws. Rarely do competing products fully synchronize with one another. However, once the data is in MS and saved, then it should now be MuseScore data, not XML. At this point, the user can enter the app and begin to clean up the data.

Computer software should never choose values for the user without informing the user. If MS can't figure out what instrument to assign during import, it should leave the instrument blank and tell the user so. However, there is no excuse as to why the user cannot change the default, and there is no excuse why it displays one thing and executes another.

Why on earth would the XML file still be of use inside MS once it is imported? Why would users want it to be? Why shouldn't they be able to trust the displays they are given by the application? If the default is "Grand Piano", why display something else?

These are questions that have nothing to do with the quality of the import file.

In reply to by littlereg1

MuseScore can't leave the instrument blank, without corrupting itself. Even if it would, it would sound as, wait, Piano!
The MusicXML file is not in MuseScore, not in its internal structures representing a score, _after the import. The import does translate it into MuseScore's inter data structures and those mazcgh the mscz file format

In reply to by Jojo-Schmitz

Ever since you sent your sample file, I have been working solely with this file - not my own. I cannot get the vocals to perform anything but piano. The drum staff is indeed percussion instruments, but its a high-pitched tom and a whistle, not a kick or snare. (As mentioned, I didn't expect that the import would render the percussion staff properly, anyway.)

In reply to by Jojo-Schmitz

Oops, there are mid-score instrument changes, like the "Bass (2)" text in measure 2 and the "(5)" in measure 3 and many more of those, probably even some invisible ones. You'd need to change instrument for all of those or just delete them all.
Yet another Garbage In, Garbage Out issue from Sibelius' MusicXML export I guess

In reply to by Jojo-Schmitz

I fully expected that the import would be rife with anomalies. I would have been quite surprised if MS would have interpreted the instrument changes correctly, brought in the ossia staves, etc. What I did not expect, however, was that once engaged in the cleanup, MS would not follow my direction - regardless of what the import had told it.

I did not invent the system that mapped sounds to MIDI instructions and then cut off this data from notation software. In fact, the makers of Finale and Sibelius were explicit in their claims that their notation systems could "run" a digital performance or recording session "just like live musicians". Allegedly, each of the symbols used in traditional music notation would have MIDI instructions that would "tell" the instruments how to perform, all the way down to articulation and dynamics. The upshot was that if users were willing to go through the pain of assigning their notation to voices, then the notation software would "take it from there".

That these basic instructions are ignored by vendors allegedly using "standard" formats (e.g., MusicXML) is not surprising in the least. Integrity has never been a hallmark of the computer industry. However, what is suprising is that this malfeasance is simply embraced by vendors and users alike, who seem to revel in spending hours on end troubleshooting software and undertanding geekspeak instead of doing things like, well, composing music.

In this case, MS dev would clearly understand that the world of computing is a massive hodge-podge of non-standard, incompatible technology, and would architect its application accordingly - especially the UI. They would make MS "third-party-proof", meaning that if users can at least get something into the application, then they can user MS's tools and fix things - still saving themselves the time and energy needed to recreate scores from scratch. (It takes FAR less time to switch from a default Grand Piano to a Guitar than it does to figure out why the assignment didn't happen in the first place.)

It appears as though, however, this is not to be. Instead, since MS's WYSIWYG cannot be trusted, users will have to spend the time they saved importing auditing each and every aspect of the rendered file. What's missing? How as MS interpreted the source data? Will the widgets work? Etc.

Gentlemen, a few observations.

The OP stated that he loaded a MusicXML into MS and had problems. Yet the file he posted is an xml. I loaded the xml into Sibelius and exported as both an MXL and an xml.

Doesn't make much difference because all three loaded in MS the same. I.E. The vocals played back as vocals and everything was piano. I used the mixer to change everything but the drums to their proper sound. No problem.

I'm not sure where the idea comes from that saving the file as a MS file would cause the proper instruments to load, comes from. I've loaded plenty of MS files that I had to correct. Should I have to? Quicker to just do it than complain about it.

Sibelius has its own extension that is saves files in. Just like MS does. But newer versions won't load in older versions. So MXL is needed in that case.

Besides which, MXL or xml is less than perfect. Always. Expect to have to fix things.

The fact that the OP can't change sounds in the mixer doesn't seem to be a problem with the file. It works for me.

In reply to by bobjp

Exactly, my friend. If you read my posts you will not see one word of insistence that either importing or saving the file in standard format would automatically fix import issues. Rather, I have maintained that after importing and saving, I cannot get MS to perform the instruments that it is showing in its own Inspector. Thus, I too have maintained that there is nothing wrong with the file.

I'm not familiar with the nuances of "XML" vs. "MusicXML", etc., but I've attached a screen cap of what I'm using from Sibelius so you can see. Because neither my own import nor the sample sent to me works, I have maintained all along that the issue is NOT with the file. It is with the software.

There is another point that I've made - repeatedly: I can change sounds in both the Inspector and the Mixer - I've sent screen caps demonstrating this. It's that no matter what I switch to, the actual sound coming from my speakers is Piano.

I am not only "willing to fix things", as you observe, but I never expected not to have to fix things. What I've been saying is that I can't "fix things".

Attachment Size
musicxml.jpg 21.96 KB

In reply to by littlereg1

In that case, case closed - the application is aware it is playing piano sounds, and it displays this information quite clearly in the mixer sub-channels for the instrument changes. So really, it's just a matter of knowing where to look for this information. Don't look in one place - the main sound for the instrument as a whole - to find information about something else - the sound for one of the mid-score instruments changes.

In reply to by Marc Sabatella

So if the Mixer says "Choir Aahs" and it sounds like a piano, as well as all the other changes also sound like a piano, then it's working properly? If so, this is a strange application. I don't know of any other place to look for this information except in the Inspector and in the Mixer. Sorry.

In reply to by littlereg1

Again, as explained, the mixer says "choir ahs" for the overall track sound, yes, but if you expand it to see the various subchannels within the track, you'll see they are set to piano:

Screenshot 2022-09-05 1.35.45 PM.png

As mentioned previously, this is because Sibelius is not including the proper information in its MusicXML files - it's one of many known bugs in Sibelius' MusicXMl export that is worked around by using the Dolet plugin.

In reply to by littlereg1

OK. I've done a bit more testing.

I loaded your file, changed the sounds and they playback properly. However, when I save the file and reload it, everything plays back as piano regardless of what the mixer says. Just as you describe.

So I did the same thing with the MXL I created from Sibelius. Load into MuseScore, change sounds, save as. Reload and sounds played back properly.

This leads me to believe that there is some kind of problem in the original file. At least some kind of misreading going on. No idea.

In reply to by bobjp

...and this is the point I'm making. After saving in MSCZ format, how could it possibly "change back" once you reload it? At this point, any changes you made and saved should be saved, but you say they are not. Each of your use cases should have turned out the same after saving. You are at least one step ahead of me since I can't even get them to change.

In reply to by littlereg1

The very first reply you got, from jim6stringer, said:
“ Right click on the staff's 'label' (i.e., instrument name) and select Staff/Part Properties. Click 'Change Instrument' and choose the instrument that the 'label' states. This will re-assign the correct sounding instrument to that staff.”

You replied that you had done that …. but you didn’t. You “Tried resetting a few of them in the Inspector as well as the Mixer with no luck.” So your actions were quite different. The Staff/Part dialogue box is not the same as the Inspector - try it.

I’ve found by sometimes bitter experience that being casual with computer programs can lead to unexpected results. I’m away from home and my PC at the moment so can’t experiment with your file.

In reply to by littlereg1

I don't think you did anything wrong here - it's just not enough, because again, Sibelius inserted a whole bunch of instrument changes, and the change here only affects the overall instrument, not each individual instrument change.

if there's any way to go back to Sibelius and re-export it with Dolet, things will go much more smoothly. But if for whatever reason you are stuck with this buggy MusicXML file, you can at least remove the spurious instrument changes by adding a real one via Add / Text, then right-clicking it, Select / All Similar Elements, and pressing delete. That will repair at least that portion of the damage caused by the buggy MusicXML export from Sibelius.

In reply to by Marc Sabatella

Marc: The vocals in my score go back and forth between "Aahs" and "Oohs" and the upright bass toggles between plucked and bowed. I did not expect MS to necessarily pick up on this, but did expect that it would either a) simply default to the instrument that the score begins with, or b) allow me to make the changes manually. Neither of these occurred.

Still, when I've provided examples of the Properties from the screen caps, I have taken them right from the part of the music in question, not the score as a whole. I've attached a snip of the score to this post where the vocals come in for the first time. You can see the blue staff label I right-clicked on to get the Properties dialog to appear. This is what I clicked on to get the snip I sent in my previous note. This is exactly what I get when I click on the precise note (also highlighted in blue).

This tells me that this single note - not just the part, not just the entire score, but the note itself, is incorrectly assigned. What I am expecting is that despite how "confused" MS is in re to the initial import, once the file is in MSCZ format and I click on a single note and bring up the Properties, once I change the voice that is assigned to that note, then that note should now sound as I assigned it.

If this doesn't happen, what this means is that MS is retaining some kind of information from the initial import that it would not normally retain had the score been created from scratch in MS. I do not understand this. I'm expecting that if I do a "Save As" in MS, then it saves as a MSCZ file, rids itself of anything "XML" and is now in a state where the user can take over and fix the anomalies.

I am going to attempt this Dolet export but I guess I need to establish an account with MusicXML. I'll post the results here. Thanks.

Attachment Size
soprano1.jpg 0 bytes

In reply to by littlereg1

As mentioned, the attachment didn't work. But I know the properties dialog in general affects the entire staff. You have independent control of the names and transpositions for the different instrument changes via this dialog, but the sounds are controlled from the mixer, so you'd need to go there to fix it if you don't want to remove the the changes as I suggested.

It's true also that MusicXML files can cause MuseScore to create scores that are different from normal scores created directly, and if those files are very buggy as is often the case with non-Dolet exports from Sibelius, it can mean some work cleaning it up later.

Anyhow, Musescore isn't in charge of the MusicXML spec or how it should be interpreted. It's up to Sibelius to export what you want. So you really need to focus your energies there.

In reply to by littlereg1

Sounds like not a good sign, no reason a (hopefully) valid MusicXMl file should take longer than an (known to be) invalid one from the same score. But it would be interesting to try importing it back into Sibelius, or any other programs. Could be something unexpected but nevertheless valid that is giving MuseScore fits. Feel free to attach it here so others can check.

In reply to by Marc Sabatella

OK, after about 40 minutes, it loaded the "paper" portion of the score, but when I went to click on anything in the interface, then the app went to "Not Responding" and I had to kill it after a couple of hours. I've attached the zipped XML file for your use.

On a side note, I'm not expecting that folks twist themselves into knots about this. I already could have recreated the score from scratch in MS based upon all the time I've spent on this import issue. Let me say that MuseScore is an awesome product for what it does and I get that it's freeware. I also get that - by definition - this means that MS is a forever WIP. I've only objected to the UI issues that leave users in the dark as to what's going on. If MS struggles with XML imports, then just drop this from the feature set until it works properly, or at least inform users that if they choose to engage it, then they are on their own for the time being.


In reply to by littlereg1

Outch, that file is mighty big.
But loads in less than a minute here
And sounds like Piano too... :-(

It isn't from Dolet though:

   <software>Sibelius 7.5.1</software>
   <software>Direct export, not from Dolet</software>
   <encoding-description>Sibelius / MusicXML 3.0</encoding-description>

And almost the same size as the one you attached earlier, a tiny bit smaller even.

In reply to by Jojo-Schmitz

Thank you, sir, but I'm not surprised. I spent two hours setting this up in Sibelius and did not really expect that it would work. I followed all the directions to the letter, got the requisite EOJ messages, confirmed the install, and followed the instructions for the export. MusicXML doco assured me that although "Dolet" does not appear in the UI of Sibelius' export tool, if it is listed in the plugins area, then it "takes over" for the standard Music XML format when user engages it.

That we're dealing with yet another UI failure with software is not surprising to me.

To all: I'm done with wasting time on this - mine as well as that of you gracious folks. Thanks for all your help - I hope that you learned a thing or two yourselves. Much appreciated.


In reply to by littlereg1

OK. I hope all participants in this thread read this post.
It is difficult to know where to start.
The OP's original file from Sibelius is a very large xml. It is not an mxl. the size probably has a lot to do with the original music in Sibelius. The xml is not buggy or faulty. Or something that needs to be reported to Sibelius. It is what it is. The fact that MuseScore has problems with it is not a poor reflection on MuseScore. And the Dolet makes little or no difference. The fact is that MuseScore and this xml don't mix. It happens.

In a post above, which it seems nobody read, I noted that Sibelius can also export mxl. This is a much smaller file. I was able to load it into MuseScore, change fonts, save as, reopen and my changes played as I intended. So it isn't that MuseScore can't deal files from Sibelius. It can.

I posted this file for people experiment with.

Does this answer the original question? Can we get this xml to work? No. And at this point I think we just need to move on.

In reply to by Jojo-Schmitz

...and as I stated, I can't seem to get this to work in Sibelius. This whole "plug-in" scheme that has evolved over the years is a euphemism for "essential features that are lacking in an application". I have spent almost as much time troubleshooting plugins in the past 20 years as I have "regular" applications.

I rarely come across a plug-in installation that is constructed with the user in mind. Instead, user is supposed to blindly follow the directions and then run the feature. That Sibelius has ignored the Dolet install - even though it displays the plug-in in its own list - is not surprising. The audacity of both providers is breathtaking: making me pay $750 for a piece of software that MuseScore - a free application - runs rings around, and this organization that makes me sign up for an "account" and go through a half dozen screens just to get to the download of something that doesn't even work.

Avid believes that people who use Sibelius are actually gamers (not musicians) who love to figure out why rehearsal marks have suddenly jumped to a second page even though they are anchored on the first, or why their sounds library keeps disappearing, or why Siblelius won't even integrate with their own Pro Tools. If Sibelius was doing its job, MuseScore wouldn't even exist.

I went back through the de-install and install process, however, and clicked on the "other" MusicXML export option (neither uses the term "Dolet" in its label - another blatant disdain for UX). The file is larger but I opened in WordPad and it says "Dolet 6.6" in the header, so I hope this is what you were looking for.

In reply to by Jojo-Schmitz


You wrote;
Loads in less than a minute and sounds line... believe it or not...

sounds line ??
A typo...no problem...so

My first thought:
Loads in less than a minute and sounds fine... believe it or not...

Then, what follows is: "Piano!" rather than "Bingo!"
Aw shucks...

Correction reads:
Loads in less than a minute and sounds like... believe it or not...

BTW: Kudos to @littlereg1's tenacity.

In reply to by littlereg1

MuseScore handles millions of MusicXML files just fine, no reason to disable the feature just because buggy ones don't load in exactly the way you expect, or because you weren't aware of how to find the problem. That's why there is also this free support forum - so you can ask, and we can help, and meanwhile the millions of other people loading MusicXML files with no trouble aren't inconvenienced.

In reply to by Marc Sabatella

I'm not the only one who is "not aware of how to find the problem". See Jojo's comment above and:

https://musescore.org/en/node/3273 [2009!]

Besides, none of this addresses the fundamental question: why does MS display one instrument and play another? This can't be laid at the doorstep of the import - not once the user saves the file and then attempts to change the sound in MS.

In reply to by littlereg1

As already explained, MuseScore is not displaying one instrument displaying another. It absolutely positively shows you the correct sound for each individual instrument change once you open up the channel strip as I demonstrated, even including a screenshot so you could exactly how to do it for yourself. Again, whether you save the score or not in no way changes the fact that there are spurious instrument changes you will need to fix or remove, and the mixer provides the method to fix them if you so choose.

While I applaud your resourcefulness in finding other threads that mention problems people have with playback, none of them are even slightly related to the Sibelius bug you are discussing here. Most of theseare a matter of people loading non-standard soundfonts and not realizing that means they will then need to select the sounds in the mixer. Others were cases of someone importing a non-standard MIDI file that didn't set up percussion correctly, another was just someone who had quibbles with the sample being used in the harp, a couple were talking about a bug in a much older version that was fixed years ago, one was even not about a release MuseScore but an experiment build made before MuseScore used MIDI playback or provided any soundfont at all.

None of these have anything whatsoever to do with the SIbelius bug we are discussing here or with the need to open the channel strip to see the sounds being used when instrument changes are present.

In reply to by Marc Sabatella

I have - again - attached a screen cap of my mixer settings. MS is showing the voice as "Soprano" and the instrument as "Choir Aahs". It is playing "Grand Piano". Please tell me where this "hidden" Grand Piano sound is coming from since you seem to be able to see it and I can't.

This really should not be an issue. There should never be the case where the UI displays one thing and the functionality that it is designed to represent does something else. Rather, the code of the software should be written so that when the UI displays a voice, the application checks its playback engine - the part of the application that is specifically tasked to render sounds - and says, "I have a mismatch. I have a way to fix this with user intervention and will state as much to the user."

With the application checking itself in this manner - the UI connected to the functionality, "bugs" like this one should never occur. The fact that the UI believes one thing and the playback engine believes another is not the fault of the import or the user. It means that there is a major disconnect inside the application itself. The import could be full of errors, but once MS renders the information, it should now be the domain of MS - just as if the user entered it by hand. The XML file is now beside the point and the file could be deleted outright making no difference to MS.

For every other import process on the planet that I have ever engaged, this is how it works. I've imported PDFs into Microsoft Word, for example, and they have all sorts of formatting errors. I painstakingly go through the documents and fix the misformatted data one piece at a time, but never have I encountered Word saying, "I know you see the word "cat" in the display, but really, behind the scenes, it's "dog" and no matter what you try to do, you can't change it." I've worked with all kinds of "standardized" formats (and I use this term extremely loosely) with genealogy software, video and photo editing software, email clients, office applications, etc., and I've never run into an instance of post-import incompatibility.

With MS able to display one voice and play another, this points to the possiblity of breakdowns for any reason whatsoever, not just importing. It means that despite the fact that you claim that none of the examples I provided exactly mirror my case here, these disconnect issues are real and could easily be addressed by fixing the application so that it is not permitted to contradict itself. Even if users clicked on "Choir Aahs" and the dropdown bounced back to "Grand Piano", this would at least be preferable because the link between the UI and the playback engine would be locked and troubleshooting simpler: Choir Aahs is not available.

When I opened my Dolet XML in MS, all of the sounds were Grand Piano. This is acceptable - even the incorrect displays - because all of it could be the problem with incompatibility. As I have stated, I did not expect that the import would be perfect. I saved the file as a MuseScore file, then went to the Acoustic Guitar staff, opened the Mixer, and changed the instrument to Steel String Guitar. The dialog now displayed "Steel String Guitar" and upon playback, that's exactly what sounded.

In other words, I am now completely inside the application - file format, score, sounds, etc. Whatever content imperfections there are as a result of the import are now in the file for me to fix. There should be nothing hidden behind the scenes that should prevent me from making these changes and updating my file. In other words, MS should now have "taken over" the file in terms of WYSIWYG, and anything that it could not take over would be displayed to the user. MS would thus never compromise the integrity of its UI.

I went through the same process for the vocals, but none of these "took". Same sequence of pokes, same display, entirely different results. Even if I am doing something "wrong", the fact that I can execute two identical processes and get two entirely different results means that the tool is at fault: I should either be able to change nothing...or change everything.

Whenever a user validly issues a command to software and it simply does not respond properly, it's most likely a bug that needs attention. However, whenever software displays on-screen acknowledgement and acceptance of an instruction that a user issues but refuses to execute that acknowledgement, this is not a "bug". This is a design flaw that should have surfaced in thorough unit testing, UI audits, and code QA sessions.

Attachment Size
soprano2.jpg 101.42 KB

In reply to by littlereg1

Your screenshot doesn't show the relevant portion of the mixer. As explained previously, you need to expand the channel strip for the relevant instrument in order to show the subchannels it contains for instrument changes. To do this, simply click the little arrow at the top of the channel strip. Here it is circled in the screenshot I helped you out with earlier:

Screenshot 2022-09-05 1.35.45 PM.png

The fact that info that is normally not relevant is shown behind collapsed panels that you can expand when you do need to access them is a pretty standard software model. But yes, it is not obvious that this is what you need to do here, and this is already improved for MuseScore 4. But it's also not rocket science, and millions of other have successfully used this. You can learn to as well, I'm confident. We're always here to help. You just need to ask!

And since you seem to have extensive experience with and knowledge about software design principles, keep in mind MuseScore is open source,e and you are always welcome to contribute your talents to make an absolutely amazing program even better!

In reply to by Marc Sabatella

It has become almost impossible to tell who is responding to whom.

Greg. You have a newer version of Sibelius than I do. Sorry you have had so much trouble with it. I don't recall having problems you describe. As for plugins, I don't use many of them. One that I do use is the Strummer. It makes your guitar parts sound more like a real player using up and down strokes. Very easy to use.
Does your version export compressed xml? That's the setting that rendered an mxl file. 357 KB. That is the file that I got settings to work with. No Dolet required. And I do have it. Again, I opened the mxl in MuseScore. Used the mixer to change sounds. Except for drums, and BTW your drums don't work in my Sibelius. Anyway, I saved as a MuseScore file. I oppend the MuseScore file in MuseScore and found that the changes I made were retained. I thought that was a start for what you are after.

BTW, in my version of Sibelius, instructions "pizz" and "bow" don't work. The sound your file loads is "Upright Acoustic Bass" which has no 'bow" (even adding the proper term 'arco' in Technique text, above the staff, doesn't work). I suspect other things done improperly, also.

Jojo. I only recently acquired the Dolet. I can't say that it makes much difference.
I have transferred xml files for full orchestra, back and forth between MuseScore and Sibelius for a few years now and have absolutely nothing like the problems with this questionable xml. These problems seem to me to be caused by problems with the original music created by Sibelius.

In reply to by bobjp

I actually tried v8 - the subscription model - but since I had all the same problems as I did with v7.5, I wasn't about to pay by the month for a bunch of bugs. I spent $1100 for the Pro Tools bundle that included the direct box and more than 40GB of sounds and none of it would work. Pro Tools wouldn't even recognize my FocusRite sound card.

The point we're all missing here is that users shouldn't be able to break the software. I wasn't the brain-dead Sibelius designer who never understood that in the real world, there doesn't have to be a one-to-one correspondence between a staff and a voice (instrument). Notation software is not simply a quaint example of a DAW where staves are equivalent to tracks, but this is how the makers of Sibelius and Finale have treated their products.

Avid "promised" that Sibelius could act as a broker for creating sound demos. They never said that this is accomplished with an avalanche of plug-ins. Even so, Sibelius' playback and record engine regularly ignored the notation as well as the plug-in instructions, and despite spending hours meticulously inserting articulation and dynamics in my scores, my demos turned out poorly.

The reasoning given by Avid support and other pundits on the Internet is that music notation is passe and "everyone" goes direct to DAW now, preferring wave images and buttons to staves and notes. Fine...but then don't lie to the public and tell them that a DAW will "fully comprehend" the notation.

This is where MuseScore comes in. With v3, this application has reached a near-full maturity and I don't know why anyone other than top-shelf professionals would spend a dime on the other products. The makers of Sibelius and Finale should be embarrassed at how the makers of MS have exposed their hucksterism.

I'm not "mad" at MS - only disappointed. I was expecting that once the import had finished, that in the case of mapping issues, one of two cases would emerge: 1) MS's back end would be passive waiting for me to re-assign characteristics of the score in the .mscz file; 2) MS's back end would have actually been corrupted so that when I attempted to fix the mapping it couldn't obey the commands.

In particular, Case 1 would mean that the instrument showing in the Inspector or Mixer would display "Grand Piano" and wait for my instructions. I would change it to "Choir Aahs", MS would then reset itself to that voice, then the playback engine responds accordingly. In Case 2, when I attempted to change the voice, the playback engine would not permit me to change to "Choir Aahs". Instead, the dropdown - dependent upon its back end display - would simply be dead. This would let me know that I have a more serious problem and probably have to rebuild the voice.

Instead, neither of these happened. MS displays the correct sound but plays another, and any attempt to change and save it fails. Yet, for all the other voices in the score, I can change the instruments just fine. My point all along is that the application should never contradict itself regardless of how data was brought into it. The imported score would simply be a flat file of notation symbols and MS would read it as such. User would then save the file in .mscz format which would "wake up" the score, disconnect it from the import file entirely (and all its problems), and permit the user to begin clean-up operations.

At no time should the UI be untrustworthy - it's the only means of work that the user has. All functionality in an application should be designed so that it checks the parts of the program that are tasked with rendering displays and connected accordingly. Mismatches would be brought to the attention of the user, and any anomalies between the UI and the functionality would cede authority to the UI - no matter what the original import had directed: "If the Mixer says Choir Aahs, then you will play this instrument. I don't care what the original import instructed you to do."

Can you imagine piloting a commercial airliner with an altimeter that said one thing while the plane was actually flying at a different height?

In reply to by littlereg1

As far as anyone can tell, no one has broken a thing. It's just a thing you haven't yet learned how t do, so you're finding it confusing - not uncommon with any new software. I've already showed you how to find the information you were having trouble finding in the first version of your score, and also showed you haw to fix it. If you'd like assistance with the new version also, we're happy to help with that too - we just need you to attach it and tell us exactly where the new problem is.

Again, with any new software, it's a given that sometimes you might look for something in one place and not find. But that doesn't mean the software is broken, it just means, there is a learning curve. Were here to help.

In reply to by Marc Sabatella


I've attached a screen cap of the "expanded" Mixer per your instructions. No, I only showed you the collapsed version previously in order to save space, but since you need to see this (my bad) I took a bunch of screen caps and bundled them together. As you can see, as I cxontinue down the line, none of them are assigned "Grand Piano". Although the image shows only the first five channels in the Mixer, I can assure you that I went through the entire rest of the set for vocal parts and all of them are assigned "Choir Aahs" or "Choir Aahs Expr" instruments. I cannot explain why, if the image you provided in from my score, is showing Grand Piano. mixer.jpg

In reply to by littlereg1

This must be a different copy of the score than the one you attached previously, because my screenshot shows, the previous one absolutely shows Grand Piano - other than my red circle, that picture was not edited in any way whatsoever. Another obvious difference is, the levels are not the same as they are in the copy you uploaded previously. And, your current picture is missing the "normal" subchannels that were present in the original (eg, one literally labeled with that word).

Here is what the previous versions shows after expanding the first channel:

Screenshot 2022-09-06 9.08.08 PM.png

As you can see, not the same score at all. If you attach your current copy - the one that those screenshots came from - we can take a look to see what changed.

In reply to by Marc Sabatella

It's the same score, alright. I only have one. In the image below you can see the Sibelius file (arrow), and my three attempts at exports. #1 is the original Sibelius Music XML; #2 is the export after the first Dolet install; #3 is the export after the Dolet re-install. Now, we're dealing with another variable, since I've been informed that the standard Sibelius Export is unacceptable to MS (#1); export #2 is just a copy of #1 since the install didn't take; and #3 is actually a Dolet export. (There are only two .mscz files, however, because I saved over the first one because the two XML files associated with the first one are the same.)

I haven't done anything else to the files. Isn't the Dolet export supposed to be different? score.jpg

In reply to by littlereg1

Sounds like you are saying the picture you showed above came from a different score - one you created by importing the Dolet file. That's the score we'd needyou to attach to help you with that one.

In general, yes, Dolet does a much better job of MusicXML export than the native facility in Sibelius. It's not just a question of being unacceptable to MuseScore - the MusicXML generated by Sibelius is simply not correct MusicXML according to the published specification, in countless ways.

In reply to by littlereg1

I just did a search of this entire thread for ".mscz" and the version you posted on 9/4 is the only one you posted. You posted some other things, but I'm looking for is the actual MSCZ file corresponding to the screenshot you posted the mixer from. And according to your file listing screenshot, that file was modified last modified 9/7. Make a copy of that file now before uploading, so that it stays unmodified even if you continue to experiment with it, so that we can always be referencing the same score when we need to.

In reply to by littlereg1

Greg, I have more questions and observations.

This xml is huge. 14,000 KB. None of my orchestra xml files are remotely that big. Is this because of all articulations and dynamics you entered? None of which seem to show up in Sibelius. And were all these done properly? I only ask because of the "Bow" in Expression text rather than Technique text. Bow won't work anyway.
If there is incorrect or conflicting information in the Sibelius file, I'm not surprised that a DAW or MuseScore will have trouble with it. Did you try a file with just notes in it? So that the DAW could make changes, as intended?

I don't use that many plug-ins. I'm not offended by them.

I don't think it takes a rocket scientist to figure out which button to push to get Dolet export.

Sibelius recognises my Focusrite.

It is true that not all instruments respond so a phrase mark. In which case there are settings to make that happen. As there are for many things that Sibelius seems to ignore. Why are they "hidden"? They aren't any more than various settings in MuseScore are. You have to know how to use the software.

And finally, I was incorrect. The Dolet download is indeed different. Aside from being 10,000 KB larger, I was able to get it to work in MuseScore. Meaning that I was able to load it into MuseScore, change instrument sounds, save as, reopen and have the changes stick. Why MuseScore can't do that right away seems more like a definition problem in the file. Which is what we have all been trying to work around. But I am more and more convinced that it's not a wonky xml as much as it is a problem with the original file in Sibelius.

I know this seems like I'm defending Sibelius. The truth is that I am using MuseScore because Sibelius probably won't work on some future version of Windows. There are things I like better in MusesScore, and things I don't. I await v4. Which I can't use because it gives me garbled playback unless I use my Focusrite. Which I don't always have room to set up.

In reply to by bobjp

People in this thread keep bringing up the file size of my export as if I could somehow get into the core code and pump it up somehow. I'm just a dumb user - all I do is click the buttons I need to render the results I want. The score is 45 pages in length and contains 11 staves - is this so outlandish?

What's surreal about all this is that this score was created largely because of the advice of other pundits responding to other failures of Sibelius' playback engine. Avid admitted to me that there are "inconsistencies" in the quality of their samples, so a few pundits suggested that I double the acoustic guitar parts to have greater control over a jimmied stereo mix. However, I created this score for live musicians who need to actually read the parts, so there is not only a sound demo use case associated with this file, but also a printing one.

No one at Avid could tell me how to combine staves into a single part while maintaining the integrity of the sounds. I mentioned this in an earlier post. Thus, if I needed, say, the Soprano voice to sing "Aahs" in some spots and "Oohs" in others, I would have to create separate staves with the separate instruments, then merge these together on a score to create parts, then create a copy of the score for the conductor to use since it would be ridiculous to create a score with 50 staves on it simply because I needed to change sounds.

This is also something I mentioned earlier: after examining the issues with my score, most of the responders advised to either use a DAW for the sounds and use Sibelius only for "paper" parts.

Avid's "help" files (and I use this term loosely) were the typical salad users get from makers of software: an incessant listing of "you can" items instead of "you should". It's rare indeed to find documentation of best practices because help systems are almost always feature based instead of task based. Yet, once the user has painted himself into a corner - exercising the functionality exactly as it "works", then they come out with the "you should haves".

I will admit that I did not use a dedicated computer for my music creation - yes, my FocusRite had to share duties with everything else on my machine, such as providing sound when I watched videos, playing songs from iTunes, etc. Yet, Sibelius is architected to be the "sole proprietor" of a Windows machine, and when I brought up the soundcard properties dialog in Pro Tools, it was empty. (I hire a professional Pro Tools engineer who reassured me - at $100 an hour - that he could set up my machine "no problem". After more than 20 minutes of trying to get something into the audio properties dialog, he said, "Well, this is not good." He abandoned the session and refunded my money.)

As for Sibleius recognizing my FocusRite, yes, it did, which is why I did all my work in Sibleius. However, this was against the advice of the pundits who disparaged Sibelius' sound engine and were not surprised when I mentioned all sorts of issues with poor samples, digital drops, voice EQ issues, little control over mixing, etc. Thus, the (temporary) move to Pro Tools.

"Export Music XML" is not the same as "Export Dolet XML". As is typical with Sibelius' shoddy UI as well as the doco from the maker of the plugin, I had no idea that "Export Music XML" was Dolet. The doco from the plugin states the process is File -> Export -> MusicXML -> Export (and the plug-in "takes over" for the Sibelius process), not Plug-ins -> MusicXML.

Yes, there are a lot of scoring anomalies in my score. This is because I followed the marketing hype from Avid about Sibelius being able to "understand" scoring. It was supposed to know about things like articulation marks, dynamics marks, etc., so this is how I generated the score. I was told that if I was able to insert these markings, then Sibelius would not only "see" them, but would instruct its playback engine to perform them. (BTW, MS allegedly is capable of mid-part instrument changes, just as I created in Sibelius. That they didn't take with my import should at least be somewhat noteworthy.)

Things began to quickly deteriorate, but I had no idea about post-notation clean-up and DAWs. Rather, I stubbornly stuck to the concept of the notation running the playback show, so - one-by-one - I began to download and utilize plug-ins, workarounds, and tricks that indeed made the file quite a hodge-podge of functionality.

It did not surprise me that MS would not fully transcribe this score. I fully expected to have to go back into the .mscz version of the score and fix, reassign, and whatever else was required to create a valid MS score. What I did not expect was that, because of Sibelius and/or XML issues, that MS would refuse to let me do this clean-up.

MuseScore's own documentation describes what this clean-up process consists up in re to instruments:

Change instrument
You can change any instrument in a score to a different instrument at any time. The following method updates instrument sound, staff name, and staff transposition all at once.

Right-click on an empty part of any measure OR on the instrument name and choose Staff Properties...;
Click on Change Instrument... (under "Part Properties");
Choose your new instrument and click OK to return to the Staff Properties dialog;
Click OK again to return to the score.

Where do you see in the above that user has to go into "sub-mixes" or "import formatting" to perform this task? Nowhere. The process should work regardless of data input - it should have nothing to do with Sibelius or any other format. Why would you expect that the average user - following the instructions above - would have any reason to believe that they wouldn't work?

As for the "just notes" approach, this is actually what I had intended to do from the beginning, but I didn't want to destroy the original score. I'm not skilled enough to configure Sibelius for "just notes", but I imagine it means keeping the staves and removing the instrument label assignments. I'm going to try this and see if it works.


In reply to by littlereg1

Do you have the Sibelius Reference Manual?

If you had your Focusrite working in Sibelius, then minimized it and opened Pro tools, the Focusrite won't work in Pro Tools. You have to tell Sibelius to release the audio engine when Sibelius is in the background. Or close Sibelius altogether.

There just isn't any good reason to double the guitar part. Stereo?

Also, no good reason to write multiple soprano parts. You do it with an instrument change that doesn't have to be visible.

When I open your original xml in Sibelius, I see almost no dynamics, no hairpins, no phrasing. Almost no instructions for real players at all. What happened to all the ones you put in?

Your drumset loads woodblocks, so there is no sound. I suspect notation problems with the drum part.

From the Dolet instruction file.
"To save a Sibelius score as a MusicXML file, go to the Plug-ins menu and select the MusicXML submenu. Then select the "Export MusicXML" item." Seems plain to me.

The quality of any notation software playback is a matter of opinion. Some things I will write in MuseScore. In fact the last video project I made, I used MuseScore because it had the blend I wanted. Others, I prefer Sibelius. In either software, you can't just throw instruments together and expect good results. You can't just plop notes down and have good music. You have to know what works and what doesn't. It's easy to blame playback sounds for poor sound, when the real culprit may well be else ware.

I'm afraid I haven't experienced this list, "poor samples, digital drops, voice EQ issues, little control over mixing,"
Yes there is no individual channel EQ, but why would you need that. You wouldn't have that control over real players.

Most DAW users I know don't bother with notation. Those that do, use notation to get the basic notes into the DAW. If any major changes to the score are made, they can be applied to the notation for real players.

You might be surprised how good notation playback, regardless of the program, can be. But you have to know howto use it.

And one more time for good measure, do you have, and use, the Sibelius Manual.

In reply to by littlereg1


I opened the attachment you posted above (Contemporary Christian Music A Retrospect_dolet_again.mscz) and observe that when the Bass (not the Upright Bass) is expanded in the Mixer, it shows more than 200 child tracks!
Also, the 'Short Instrument Name' is missing on some pages, and different on others ('M' on page 5).
Right click on one of the 'instrument change' texts (e.g., either 'Bass' or 'Instrument change' in measure 2).
Select > All Similar Elements in Same Staff
Press 'Delete'.
Soprano and Alto will keep subchannels and sound as voices ("ahs"). You just need to change the Bass from Grand Piano to whatever sound you wish.

If desired, you could select all the instrument change texts instead of only the Bass, but patience is required for all to become selected.

In reply to by littlereg1

Thanks for the example! I can confirm there is a problem here, although I'm not 100% of the cause. I think it's that there are so many instrument changes - the bass part in particular has what seems like over 100 of them! - it's actually maxed out the number of theoretically available MIDI ports and channels, so channels have to be reused across instruments. I note if I delete just the bass staff, the others staves play as expected. I've seen this happen in a handful of other very extreme cases, but it's incredibly rare to have that many instrument changes in a score. It's unfortunate MIDI doesn't handle that better, but perhaps MIDI 2.0 is better.

Anyhow, the solution is as before - those instrument changes don't belong and should be deleted via the method I described earlier.

MuseScore 4 handles channels and instrument changes very differently, so it's not a problem there. Thus, not worth filing a bug report.

The critical issue was with all those instrument changes.
Back on Sep.4 you attached Contemporary Christian Music A Retrospect.mscz which is what I used to come up with this:
Contemporary Christian Music playable_w_drums.mscz

It may still need work with formatting (unnecessary frames, page breaks, etc.), but other issues included the guitars. Checking Staff/Part Properties showed that some didn't have strings available(!), and the pitch range was silly. 'Change Instrument' and selecting a guitar from the 'Select Instrument' dialog re-assigned the instrument(s) correctly.
Also in that dialog, a guitar with the word 'Tablature' shown in parentheses must be chosen for the score's TAB staff.
The piano staff had "fake" pedal symbols which I deleted.

In reply to by Jm6stringer

I do not dispute that the end product of the score has turned out to be a mish-mash of instruments in the export file. Yet, it seems as though the folks on this thread are inferring that I deliberately went about constructing it this way. Nothing could be further from the truth. Yes, I have the Sibelius Manual and I have yet to find anything in it that instructs how to make single parts - which are performed by single players in real life - perform different "instruments" other than to insert the instrument changes "mid-staff". Keep in mind that humans were around before computers and they can sing both "aahs" and "oohs" just like an upright bassist can switch from plucking to bowing.

The reason you're not seeing some of the articulation and dynamics that is in the file is simply because neither XML import style brought them in. It's possible, however, that the import may have interpreted these insertions as "instruments" which is why you think you see so many in the MS file. For example, none of my pedal marks came in:

The import also inserted a bunch of instrument changes where there are none. Compare the original Sibelius file:
...to the MS interpretation:

As for the drum sounds, it's as simple as it gets: kick and snare, and these are scored traditionally - the kick is on the first space of the staff and the snare on the third space. The snare note with the "X" head in Sibelius is defined as a brush instead of a stick when it's on the third space, at least it was in the library I purchased from Avid. Why the import did not understand this and inserted "woodblock" is beyond me.

The page breaks in Sibelius are without issue. The import did not bring in the layout of the score as it is in Sibelius: instead of two systems per page, the import brought in only one. As to the "silly" pitch ranges - none of these exist in my score with the exception of a handful that stretch the vocal ranges a tad. (This was done because this is a pop score not a formal choral arrangement, and ranges for these voices has always been subject to debate.) The guitar part was scored so that the pitches would sound as I wanted them to, then a tab line was created for the guitarist to actually play.

I can't help it if the import refused to place the original scored notes in the correct octaves:

These are just a handful of examples of the disparities between the original file and the import. There are tons of others. I could have brought all these up in my original post but didn't. This is because, although I was disappointed in the outcome, I was not surprised. I expected a lot of cleanup and I sure got it.

No, my only complaint was that MS was not letting me perform the cleanup on the basis of its display. If users are forced to dig into the application to find the nerd knobs, then importing - at least for me - is a senseless exercise. Instead of being able to select the instruments I need simply from the Inspector, I have been advised to go into the Mixer and the score and start deleting all these instrument references one at a time. This is a senseless exercise and defeats the purpose of the alleged effort savings that importing is supposed to provide.

In reply to by littlereg1

It's probably just how Sibelius does things, where all articulations are apparently treated as instrument changes. I would guess that is actually get different sounds as opposed to merely fiddling with note duration and velocity. So the export - both native and Dolet - are honoring the intent of the file as Sibelius handles articulations, but really, they should be smarter than that that and allow the importing program to handle articulation its own way. Maybe there is some option in Sibelius to make that happen.

Anyhow, MuseScore is simply honoring what's there in the MusicXML file - those additional instrument changes are present, so yes, they get imported. I suppose an option could added to MuseScore to ignore those for the sake of malformed Sibelius files, but again, I've shown you exactly how to remove them, it takes five seconds, so I suggest you just do that and move on.

BTW, no one is blaming you. We're simply pointing out that this case is extraordinarily unusual, and that's how come it isn't handled more elegantly - it simply isn't something that comes up very often despite millions of MusicXML files having been imported over the years.

As for the other unrelated issues - best to open a separate thread to let us help you with those. I can tell you the bass issue is almost certainly just a matter of bass being a tranposing instrument. And the drumset issue is probably a matter of the MusicXML file not include the appropriate definitions.

In reply to by Marc Sabatella

I'm sorry, Marc, but if these are the instructions you game me: "But if for whatever reason you are stuck with this buggy MusicXML file, you can at least remove the spurious instrument changes by adding a real one via Add / Text, then right-clicking it, Select / All Similar Elements, and pressing delete.", I can't get this to work. I added an instrument change to a staff, right-clicked on it, and this is what I get:

This may take "five seconds" for you, but even if this was working for me, it takes MS about 10 seconds per poke for me. The application keeps going to "Not Responding".

What am I missing here? Thanks.

In reply to by littlereg1

As I said, after right-clicking, go to Select / All Similar Elements. "Select" is the sixth element in that menu. Click it to reveal a submenu, one of the items on that is "All Similar Elements". So, only two seconds to go from here!

Not sure what you mean about "not responding", that sounds unrelated - you score responds fine for me considering its size. Do be sure you have enough available RAM; close others applications that may be taking resources etc. Best to start a new thread about that problem if it persist.

In reply to by Marc Sabatella

OK, I'm just used to different semantics for submenus: Select -> All Similar Elements. The only mid-part instrument references that I could find in the score were Bass, so I followed your instructions and everything worked: the vocal parts are now voices and I did not have to re-select what is in the dropdown. The only staff that needs fixing is the drum kit, since it appears as though MS doesn't have anything the equivalent of "Drum Set". (As an aside, it sure would help if that massive instrument list were either sorted in raw alpha order, or grouped by instrument group w/headers and horizontal dividers.)

I'll go through the MS doco and learn how to create a drum set. (I've written tons of scores in MS but only head charts, so this is new territory for me...not a problem.)

The rest of the score has converted my ossia staves into standard, and assigned an instrument label to these. The staves that had multiple voices in them (not multiple instruments) that I used to indicate performance instructions for repeats have also been converted to just text. I'm going to experiment with playback (now that the playback button works) and see how MS "reads" these instructions.

Something that may be of interest to you is that once I got rid of the rogue Bass parts, the Not Responding issue went away. I have an 11th-gen i7 Windows machine with 32GB of RAM and a 1TB SSD, so there should normally be enough firepower for my purposes.

Thanks again!

In reply to by littlereg1

MuseScore definitely supports drumsets, both for native MuseScore files and for MusicXML import. But Subelkius is well-known for not including the necessary information in its MusicXML exports, so it's common indeed for MuseScore to need to guess about this and often guesses wrong. I thought Dolet fixed that but maybe not always. Anyhow, as mentioned elsewhere, it Sibelius is using non-standard pitch assignments, but this can also be worked around easily enough in MuseScore. After enabling drumset in the Mixer, right-click one note, Select / More / Same pitch, then use the arrow keys to move through the different valid standard General MIDI pitches defined for the drumset, stopping when you get to the one you want. Like, I'm guessing those timbales are meant to be bass drums.

In reply to by Marc Sabatella

Thanks, I've noted these instructions. I fully expected to have to fix my drum kit in MS, but since there are only two instrument sounds - kick and brush, I'm hoping that this will not be a difficult exercise. Where the nuance is, however, is that Sibelius let me first establish a sound (e.g., Brush), then I could assign it to various drums in the kit simply by changing the position of the "pitch" in the staff. I was saved the exercise of having to create separate lines for the snare and each of the toms, as well as leveraging the ability to use only one staff for the part that the instrumentalst would have to read. This was one aspect of Sibleius' "intelligence" that was most welcome. I can fully understand, however, that this isn't something that would translate well into some sort of generic export, and even if MS can also do this, I expected to have to rebuild it.

In reply to by littlereg1

Sorry, no one is blaming you for anything. We're just trying to figure out why there are so many problems.

To that end, I took one of my orchestra scores in Sibelius and copied it over and over until it was 60 pages and over 400 measures. The resulting Dolet export was only 6,000 KB. A forth of yours. It opened in MuseScore with only minor problems. So I can't help but think that there is indeed some kind of problem with your file. As in way too many things in it.

Change between ooh and aah with expression text. If you start with aah in the mixer, at the point where you want to change to ooh, hit CTRL+E and type oo. At the point you want to go back, hit CTRL+E and type ah. This is not in the manual as far as I know. Nor is it a plug in. I don't know where I heard about it. It may have been mentioned as an aside in a Pipe organ add on discussion.

As a guitar player, I have many thoughts about your guitar parts. But not here.

I would probably have to see your Sib. file to tell much more.

In reply to by bobjp

If Sibelius is like a typical "document generating" application, it has the potential to amass unwanted instructions ("dirty data") in its datafile. I had more than one session with Avid where the tech support conclusion was that "due to excessive editing" (whatever that is), the file had become corrupt. My Sibelius file that we've been working with here has been punched and kicked countless times as I've made many changes, changed plug-ins and plug-in versions, switched sound engines, attempted to link to Pro Tools, etc. I would notice that as I saved the file, even though there were still only 45 double-system pages, the .sib file would continue to grow.

Even so, doesn't the XML file itself reveal where the bloat is coming from? In my experience with XML, it's just a file of functional text, a primitive code. When I opened the file in Excel, there was about 65,000 lines in it. Wouldn't any XML file that has this many lines in it be similar in size to mine? The "things" that are in it are not of my doing directly. I've described the contents of this file in a previous post. I have seen scores in blogs and forums that are far larger than mine and much more complicated. I have assumed that someone familiar with how MS reads XML files could easily see what MS doesn't like about the file.

I had worked with expression text in Sibelius and experienced a small delay or cutoff between the instruments that was not there when utilizing the mid-staff instrument changes, so I kind of abandoned this concept altogether and never even considered it in MS. Thanks for your suggestion, however, and I will try it.

As for the guitar, I know a lot of what you would probably say, but I must reiterate that I had multiple purposes for the guitar part: printing, tab, and sound. Again, this is a use case not a "feature", and it is very hard to find doco on use cases, especially unusual ones like mine. The reason for all this trouble? The guitarists that I work with cannot read music or anything where they can directly translate what a composer wants directly from the paper. I am saddled with the additional responsbility of matching an exact MP3 sound all the way down to each string and fret so that these guitarists do not fall back on familiar chord structures and unwanted strumming rhythms. I need them to play exactly what I write - and record.

We could indeed go to another thread to discuss details, but the topic is relevant here because of what is possible to the user in terms of creating scores and what these XML import/export operations "dislike" about all this. Is there a best practices manual for writing scores that are intended for XML export? I wish I had seen one of these prior to going through all this.


In reply to by littlereg1

I can't speak for any other program, but I can say what makes sense when created a score in MuseScore intended for MusicXML - try to stick to defaults as much as possible, don't make manual adjustments to try to make something look like something it's not, don't use a symbol that normally means one thing to mean another, etc. So for instance, don't attach a symbol to one note and then drag it to look like it applies to another - MusicXML export won't capture that. That sort of thing.

In reply to by littlereg1

Interesting. I can't say that I've ever experienced any delay using expression text. My computer is not half of what yours is.
It is probably a safe bet that your guitarists will not play your chords just the way you have scored them. I probably wouldn't. I may or may not add the low A string on your D chords. I would probably play a six string G chord. You would never know the difference. I think what is far more important is the rhythm and direction of the strum. I.E. up or down. For example page 14 , measure 35. Are those all down strokes, or down down up down down up. Big difference in sound.

In reply to by bobjp

Yes, and I fully expected to have to do some significant adjustments to the guitar sound once we went with live musicians. However, as I've mentioned, they need more to understand what they need to sound like than what exactly they need to play.

Because of the serious digital drops that I was getting from the Sibelius mixer production, I had to develop a number of scoring workarounds in order to get the overall parts to "emerge" in the mix. (This is a main reason why advice given to me was to forget Sibelius and use a DAW for my demos.) I would experience anomalies like a piano run playing a simple 10- or 12-note scale and along the way a section of the scale would simply vanish, as if the pianist had simply stopped playing. Adjusting the faders, dynamics in the score, etc., were either of little consequence, or they created fresh mix problems. I found myself playing Whack-a-Mole with the Mixer.

I found - interestingly - that adding notes to parts that I didn't necessarily want would "pop out" the ones I did want. For example, adding an addtional lower note to a guitar part took away the tinny sound, or making a scale run on the piano octaves instead of single notes actually brought out the "missing" upper notes. I know this is all engineering stuff, but I hadn't expected this to be such a battle. (I first engaged all this in the world of R2R one-inch tape 50 years ago, and we never faced blatant holes like this in our recording. A little working with the mix was all it took to at least get the sounds to "be there" even if the levels and balance were not what we wanted yet. One would expect that these issues would have been anticipated by the developers of digital music tools.)

The reason that the guitar parts in this score are sometimes ill-advised or even unrealistic is that when I attempted more-conventional voicings, the sound from the part was tinny and choppy. (Even in the final that you're seeing it's pretty bad, but still leaps and bounds ahead of when I started.) This goes from the strumming strokes, as well. I never could get the samples to respond to "up" or "down" strokes, and was told that the particular sample I was using did not include this data in it. (At the time I composed this score, I did not know that if instructions for articulation, dynamics, volume, etc., are not built into a sample, then it won't respond to these when inserted in the score. This drove me crazy because most of Sibleius' synth sounds would not even respond to volume changes, and the upright bass sounds had only one duration for plucking - no matter what I had composed in the score: staccato, 128th-note durations, etc.)


In reply to by littlereg1

The more I read about what you've gone through, the more fascinated I become.
What version of Sibelius did you start this in. I see you used version 18 at one point.
I only have 7.5.1
I have used Sibelius since 4.1, and I can't say that I've ever had the problems that you describe. I'm trying to figure out why.
I don't know for sure, but I'd be surprised if response to dynamics is encoded in the sample. I would think that would be more a function of the Sibelius Player.
Did you solo the piano and heard that the piano did indeed drop put? Or did it get covered up by something else going on? As I said, I can't remember ever having anything just not sound.
Any guitar, real or virtual, will tent to sound tinny if only the top four strings are played. I experimented with a few measures of your guitar part. I made them five and six string chords, added the strummer plug in, and the sound was much better. There is only one default six string sound. The Strummer just works.
Which makes me wonder if you are using the default 35 G sound file. If not, that would explain a lot.
Yes, the acoustic bass has one sound. Much, though not totally, like a real player. If you need an "arco" sound, you'll need a different instrument.
I also did tape mixdown and am just as happy to not have that to go through any more.

True. If you want professional quality recordings, a DAW will be the way to go for some time to come. In the meantime, I have had very good results with both MuseScore and Sibelius. Certainly you should be able to get a satisfactory demo recording with out all these problems.

In reply to by bobjp

I've experienced digital drop in other areas of my music world, not just with Sibelius. Digital waveforms are squared, whereas analog are curved. (I've had sound engineers tell me that when digital came out and really began to be "the thing," guys that had literally decades in the business had to re-examine a great deal of their knowledge to make things sound good.) Evidently, digital mixes can produce overtone series that can collide with one another at just the right moment and, unlike their analog counterparts, the tools reproducing the sounds have to "decide" which to favor.

I experience digital drop in the band I'm in. I play keyboards along with one guitarist, vocalist, bassist, and drummer. We'll do a sound check and tweak everything just right so it sounds good in the mains as well as the in-ear monitors. Then...the guitarist will step on a pedal, introducing an entirely new set of soundwaves. "What happened to the keyboard?" someone will ask. "Sounds fine to me!" says someone else. Reason? Slightly different personal monitor mixes.

Digital drop has caused what I call "nuclear arms escalation" when we've rehearsed or performed. It starts with something (partially) going away, so someone will raise a level, which in turn causes someone else to raise a level, and off we go. Somehow, a frequency series collision occurs between, say, a Rhodes sound pitch D3 and an electric guitar chorus pedal pitch F4...and only one can "win". (We've actually conducted experiments with this where I'll play, say, the D3 in quarter note increments and set the levels where everyone can hear. Then, we ask the guitarist to stomp on his chorus pedal and play F4 and the Piano D3 almost completely goes away.)

The problem with the notation software playback is exacerbated by the arbitrary nature of the samples - they're all identical...and thus unforgiving. For example, in my score, you'll have noticed that there are times when all three of the "female" voices are singing exactly the same pitch. Well, at this point, you'll also notice that all of a sudden, the volume of the three vocal lines has just tripled. All the while they were singing different pitches (lines or harmonies), there was a relatively stable volume output, but as soon as they blended on pitch, now the vocals are all screaming at you.

OK, if there is a blend during a piano run, then this is going to kill whatever levels you had set for other "sensitive" instruments, such as a piano or an upright bass. This is how volume causes the drop, not necessarily a frequency collision. I've experienced digital drop with both.

When the instruments are played alone in the score, they respond excellently to the guidance of the score via the player. In fact, running them regularly through the player is how I determine how I like what I've written for each in the first place. It's just shocking that when mixed with such a small number of other instruments that these "holes" appear. (I've never used this "strummer" plug-in, btw.)

The upshot is that, yes, I am a rookie when it comes to engineering sound - I've worked with engineers, but never did any of this myself. However, this isn't how Sibelius was marketed - that you had to have pro-level mixing skills just to get a good demo. In fact, this is what the notation software was supposed to substitute for. Instead of having to do a bunch of complex mixing, composers were supposed to be able to act as virtual "conductors"of their score, sprinkling articulation and dynamics throughout that was supposed to go a long way toward "mixing", just like a real orchestra would have to do on stage.

Instead, I learned later that the playback is wholly dependent upon the quality of the samples - their ability to read and respond to notation nuances, and these nuances would aid the playback when levels and pan were set. I learned too late that what I bought from Avid doesn't even come close to this.

In reply to by littlereg1

Indeed, the last two bands I was in used PA speakers that were able to be placed behind the band. Without feed back. Everyone could hear how they were fitting into the mix. But it's not for everyone.

Yes, any time multiples of the same instrument play the same pitch, the sound is much, much louder. So don't do that. What? Gasp! No you don't get to write what you want. If you take the song this thread is about and rearrange it for string Quartet, you would write it to fit the sound you are after.

Your altos are panned at the same place. Is one sitting on the other's lap? Your Guitars blast over everything for the same reason.

Have you noticed that there are two sets of versions for many instruments? Probably not. Version 1 is a more up front, in your face sound. Version 2 is more back in the mix. A bit more mellow. You can't just set up a score and go with whatever Sibelius loads. You have many, many choices. All of which sound better than the default load. Why does Sibelius do that? Because it has no idea what you want until you tell it. I always change what the program loads.

No pro level mixing skills required. Notation software has always had one purpose, as you well know. Creating a pro-level demo has never been one of them. Creating a good demo? Sure. But you have to know the tools you are working with. There are a lot of things you can do to get good playback. But you have to learn what they are.

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