Sound in wrong octave

• May 12, 2009 - 17:48

I cannot seem to identify exactely what happens.

It seems that once I have notated a bar in the bass, if I go back to that bar and make additions and alterations, the added notes sound an octave higher than written.

I am able to work around this by reloading. Then everything is in its proper octave.

i have a similar bug with ties. I will describe it in another post.

Regards,


Comments

I just ran into this, and it's been a real pain. I'm attaching one of the segments of the song where (if I recall correctly) the problem existed before I restarted MuseScore. It bears noting that the problem recurs after restarts, but not always (perhaps never?) with the notes I've entered before the restart.

I believe it was either the E or F# notes in the second organ staff that were one octave lower than written. (I realize the OP had the opposite problem, but they sound like the same bug to me.) The other notes were fine.

It's worth noting that I never had this problem before installing the "Harmony Rules" plugin (latest version, as of about 3 weeks ago). I've recently disabled the plugin and deleted all traces of it that I could find in my mcsz file, but it's possible that there are still remnants. This song (perhaps others, too) still has the problem as I add new notes.

I'm using MuseScore 1.2, revision 5470.

I'll try to provide any further information you need.

BTW, a HUGE THANKS to the MuseScore team for making this software! I've been looking for a decent replacement for Cakewalk Scorewriter (which wasn't very good, but it was in my price range) for over 10 years, now. Finally, I've managed to convert my old music to a decent format (Cakewalk > Overture > MuseScore), and--finally--have a means to keep on writing. (Pencil-and-paper is just too darned inefficient for my level of patience.) I'm running it in both Windows and Linux! :-D A special thanks to the whole MuseScore team!

Attachment Size
MuseScore Bug Sample.mscz 2.78 KB

In reply to by dmutters

This score snippet plays completely normally on my system.

You are aware that the organ sample in TimGM6MB plays at 16' pitch (octave down)?? This isn't always apparent on some sound systems as the 16' element of notes below tenor C are below the frequency response range of some speakers.

Incidentally, I too have the plugin you mentioned, and ran it on the score snippet you supplied - it made no difference at all to the sound, nor would I expect it to as the plugin framework doesn't allow the creation of new notes.

In reply to by ChurchOrganist

Thanks for the (speedy) replies, ChurchOrganist and lasconic.

First, the speaker issue: I'm using a set of 2.1 speakers that were fairly expensive when I bought them, and manage better frequency range than most 6" stereo speakers I've seen--so I *think* I'm hearing the organ correctly, but if the score sounds really off, in terms of the pitch being generally grating, I may need to try different speakers to be sure that it's coming out as I think it is. Since I don't personally play the organ, and have little experience writing for one, I'm open to suggestions should you see something I'm obviously screwing-up (such as writing in a nasty octave). :-)

I understand your (apparent) frustration regarding not being able to reproduce this bug from my snippet. In point of fact, neither could I, but it seemed like I should include a segment that gave me problems at some point, in hopes that fiddling might make it happen again. I'm still trying to figure out just what "breaks" things, and I've gotten it narrowed-down to a few candidates, which I'll share here, in case you or another can find the problem before I do.

1) Moving notes around after placing them. My mouse and ear work aren't always accurate, so I find myself dragging notes around a lot. I also use the arrow keys for this.
2) Using CTRL-UP/DOWN to change octaves. I do this semi-frequently, as well, so it's on the list of "suspects."
3) Copying/pasting things--from individual notes to groups of notes to measures, multiple measures, etc. I do this a fair bit, so it's a likely suspect.

As you've said, the HarmonyRules plugin isn't really a note-altering plugin, so I didn't expect it was the problem, but since I recently added it, it seemed prudent to make the suggestion.

Another note: sometimes when I re-start MS, the wrong octaves remain wrong. It seems like whether they get "magically" fixed is kind-of random--to my eyes, thus far, at least.

I'd gladly give you guys some cash to help with the work, but I'm super-broke! Sorry! Maybe when I strike it big with my music... (LOL)

In reply to by dmutters

You haven't said what soundfont you are using. That is the main determinant of what you hear. Also, what happens if you change to a simpler sound, like piano (use the Mixer window)? If you hear the pitches correctly with another sound, then MuseScore is doing it's job correctly, and it's just a matter of you not liking the mix of harmonics present in the particular organ soundfont you are using.

In reply to by Marc Sabatella

Of course. Right now, I'm booted into Windows7 (x64), and am using the default midi soundfont. I'm not sure how to find more specific information on it, but in the MS Preferences window (under I/O), I have this ("x" is a check mark):

x Use internal synthesizer
x Portaudio
Api = MME
Device = Microsoft Sound Mapper Output

I'm using Realtek AC97HD audio hardware/codecs.

I've noticed this problem happening in several different instruments, including "piano," (accoustic/classical) "guitar," "vocals," and others. I'm pretty sure this isn't a problem with the midi soundfont, especially since it also happens in Fedora with several variants of the Freepats soundfont library.

In reply to by dmutters

Are you saying that on *this specific score*, you see a problem even using the default TimGM6mb soundfont and changing the sound to piano? If so, then then sounds like a hardware problem to me, since as noted, it works fine for everyone else. Or if you are not trying to say that *this particular example* sounds wrong on your machine with a piano sound, maybe you could try posting one that does. Because like others, this one sounds fine to me. The fundamental on low organ note is often only barely audible which does fool people into thinking they are hearing a note an octave higher. But try transposing this note an octave higher and you'll hear what that sounds like.

I guess if you're still absolutely convinced you are hearing something different than we are, you could try saving this to an audio file and posting that somewhere. But again, I'm bettering it's just your speakers/ears not fully representing the fundamental, which is normal.

In reply to by Marc Sabatella

That's not what I'm saying.

The score I posted sounds alright, NOW, because after re-loading MuseScore one or more times, or (apparently) after transferring it to another machine, the wrong octaves are corrected. As I said before, what sounds bad right now might not sound bad after re-loading MS a couple of times. This happens on multiple scores, but (in my experience, not having used pre-1.0 versions) it only happens in MS 1.2.

Again, it happens with multiple instruments, on multiple scores. Any given wrong-playing note will "magically" stop being played in the wrong octave if I re-start MS one or more times. As far as I can tell, once a note has been entered and sounds wrong, then MS is restarted and the note sounds right, it won't go back to sounding wrong, again. New notes I type in, however, will sometimes sound in the wrong octave until I restart MS, which "fixes" them permanently (unless I alter them afterward).

I realize that the organ's fundamental is faint at certain frequencies, and I can see how bad speakers could cause problems with hearing it properly. This is not the case with my setup, since any given note has both sounded in the wrong octave and the right one, depending on some pattern I'm still trying to determine. Since even when this happens, it can be corrected without changing the note (by restarting MS), it must therefore NOT be a problem with hardware.

In reply to by dmutters

I don't yet know what I'm doing to cause it, but I've managed to save a .mcsz file and a .ogg of it, which (so far) resproduces the problem each time I open/play them.

Process:

1) Had a song open where the problem was evident.
2) Started a new sheet; added instruments that were active in the section.
3) Without closing original sheet, copy/paste problematic parts.
4) Set tempo to be consistent with original (so it's easier to hear the problem).
5) Save .mcsz of new score and .ogg into new files.

I've re-opened both new files several times--in MuseScore and AIMP3, respectively--and they've shown the problem consistently across several program restarts/re-openings. (Fingers crossed) this should mean that when you download/open the files, they should show the bad behavior. Even if only the .ogg file shows the behavior, you can compare it to the notes, as-written to see the discrepancy. The first note on the organ's treble line, in each measure is one octave too low.

Please note that due to the posting rules of this site, I had to zip the .ogg file in order to attach it. I hope this isn't a problem. (I used 7zip for this, saving as a .zip file with default settings.)

Attachment Size
MuseScore Bug Sample2.mscz 2.4 KB
MuseScore Bug Sample2.zip 66.32 KB

In reply to by dmutters

Sorry, your OGG file really just demonstrates *exactly* what I have been talking about. Listening to the OGG file, we are clearly hearing the same sounds, but I think you are just still not getting how deceptive organ pitches can be because of their overtones. Yes, I hear the *illusion* that the first B of each sixteenth group is an octave lower than written, and maybe the soundfont could stand to not use such obvious sample switching points. Try moving a note with the up/down arrows and you can hear a very clear sample switch between G# and A, and that is what is tricking you into thinking that B is lower than the G#. The B is a little quieter, and has different overtones, such that when you hear them in rapid succession, especially with other sounds in the way, your ear is fooled into hearing the B an octave lower than it is.

But that's all right there in the soundfont; it isn't MuseScore's doing. MuseScore is very clearly playing that soundfont correctly. Try changing to any other sound within that sound, or any other soundfont, and the aural illusion goes away.

In reply to by Marc Sabatella

You could be right about this--particularly by way of the context in which the notes are played.

It's equally possible that we have two problems going on: my ear not being that great for deciphering the soundfont's intricacies, and the font/program actually doing something odd. I mention the second again only because I don't know how else to explain the octaves changing in some instances only after a program restart. If it were just my ear plus context, then restarting the program and listening to the same segment again, from the same starting point, should have no effect. If the program's having a problem, then it could be that my bad ear for the organ soundfont is simply complicating/masking it when I try to show it to others.

I'll pay closer attention to the phenomenon you mentioned, and see if it really is "all in my head," though.

In reply to by dmutters

Now that is interesting!

When I play the mscz file the organ sounds correct.but when I play the ogg file the upper frequencies of the organ sample assigned to B are masked, giving more prominence to the fundamental, and thus the illusion that the sound is an octave lower.

This suggests that it may be an EQ setting on your system which is responsible for this.

Incidentally, MuseScore doesn't use the WIndows internal soundfont. It has it's own called TimGM6MB which has known limitations (not that the WIndows internal one is any better). If you would like better sounds have a look at the SoundFont section of the online manual.

In reply to by ChurchOrganist

Hmmm...I hadn't known that about the internal vs. Windows soundfont. I'll also take a look at my EQ; do you have suggestions on where to find the appropriate settings? Otherwise, if EQ isn't at fault, do you know who maintains TimGM6MB? Is this the sort of thing the maintainer would be interested in hearing about? (I don't mean to be a pest...)

I'll take a look at the online manual. I'd much prefer better sounds, and I can understand why MS doesn't ship with the "good stuff" (size, I assume).

Thanks for the input.

In reply to by dmutters

Such is the nature of illusions - they can appear and disappear according to the angle at which you hold you head, how much coffee you've had, etc. So I'm not surprised if the effect seems to go away depending on how you listen to the audio - whether it comes direct from the playback or whether it has been compressed (OGG is a lossy format) which could easily affect overtones, also whether you listen in combination with other sounds which might mask the overtones, or whether you've taken the time to clear your head by rebooting.

As for the fact that the B has a different overtone structure than the G and G#, I'm not so sure it's really "flaw" in the TimGM6mb soundfont. Organs really do have unusual (by the standards of most other instruments) overtone profiles, and unless you sample each and every note individually - and do it from an instrument devoid of its own quirks - you are going to hear some artifiacts. TimGM6mb's main claim to fame is small size. Lessening the switchovers would mean a larger soundfont.

Anyhow, as for "better" sounds, sure, there are any better soundfonts put there, many of them free. I have tried quite a few soundfonts of varying sizes. While there are individual things I like about a lot of them, I still overall the best for my purposes is the FluidR3 soundfont that the Handbook links to. When I play you example with that soundfont, the illusion is lessened considerably. However, there is still an obvious switchover between that G# and B. not syre if Fluid doesn't sample every note either, of if they samoled an instrument that itself changed pipes or something there.

In reply to by dmutters

"I'd gladly give you guys some cash to help with the work, but I'm super-broke!"

There are other ways of contributing to MuseScore.

Giving support here in the forums to other users is one way, there are also always openings for people who want to get involved with documentation, translation, testing and even programming.

One project which I, personally would be grateful of some help with is the preparation of a MuseScore/MusicXML 3.0 compatible instrument list, which is currently a work in progress as a Google Doc. You can find more information about that here: New cascading instrument definition

Finally there is the production of Demo Scores featuring the new stuff coming in version 2.0

If you feel inclined to help this way, head over to the Development of the website and see what it's all about.

In reply to by ChurchOrganist

Thanks for the suggestions, ChurchOrganist.

Though chronic illness prevents me from working reliably, I do often help on forums related to things I know. As I become more familiar with MuseScore, I hope to be able to contribute more. For now, I'll work on helping the developers solve bugs like this one. :-)

I don't know if it's the same, but I may have encountered this - it might be exclusive to 1.2 and fixed in the trunk.

1. Create score.
2. Enable 'Note Entry'.
3. Enter two crotchets.
4. Disable 'Note Entry'.

5. 'Undo'.
6. 'Play'.
7. 'Redo'.
8. 'Play'.

Result: Only the first note is played.

5. 'Undo' twice.
6. 'Redo' twice.
7. 'Play'.
8. 'Undo'.
9. 'Play'.

Result: Two notes are heard with only one crotchet.

Discussion: I think if you click away or save, it's no longer reproducible.

Using MuseScore 1.2 - Mac 10.7.4.

Update: I've downloaded a more "deluxe" soundfont (FluidR3GM), and while the organ's fundamental note is still a bit quiet at times, it no longer sounds like it's playing in the wrong octave. The previous comments about the tone of the instrument and the quality of the sound font seem to have been correct.

Thanks for your patience.

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