Snail mail bug report

• Mar 12, 2015 - 11:42

This 2.0 bug report came in via snail mail. It was anonymous, so no way to return an answer to sender :(

An mscz file to try reproduce the bug is attached.

16170607904_334292d8f7_c.jpg

Attachment Size
testbugsnailmail.mscz 13 KB

Comments

• Mar 12, 2015 - 13:46

I can't reproduce this on my system.

I wonder if he is using a different soundfont?

MY other thought is that the somewhat heavy vibrato on the violin instrument is making those notes appear out of tune.

Reply

• Mar 16, 2015 - 09:25

Ooer!

I didn't get it until I added an octave A below the other.

The problem is again at sample level.

The samples have so much vibrato on them that it is obscuring the actual pitch.

However, I see that the original author hasn't changed the root key to fit with the calculated version. Doing this may improve things slightly.

Keep an eye out for the changed sf2 version in the Soundfont Forum.

Reply

• Mar 13, 2015 - 02:26

Hello everyone...
I am guessing that it is Violin II in 'testbugsnailmail.mscz' which uses the E string to play those (flat?) notes in question. They also sound ok to me.

As an experiment:
Since 'anonymous' mentions those notes as being almost a half-step flat; to compensate, I raised them almost a half-step (40 cents); and then, half of a half step (25 cents). Listen to the attached file using TimGM6mb vs. FluidR3 GM2-2.SF2. You can even mute/unmute the different instruments for further comparison.
With TimGM6, the tuning really 'zeroes in' at the 'normal' setting.

In addition to the soundfont, and/or heavy vibrato (both already mentioned), this might also be related to musical temperament and tuning practice - especially when talking about 'a couple of cents', as per ZackTheCardshark.
When performing, I often make small adjustments to guitar tunings depending on a score's key, trying to get the 'fretted' perfect octaves to 'sound nice' with the guitar's various 'natural' harmonics. (Not so easy to do - most especially on a cheap guitar with a warped neck.)
Also, I recall reading in these forums that likewise, clarinet players can tweak something or other for the best overall tuning for a particular key.
In real life, though, doing something like this is not possible, for example, on a piano once it is tuned - re: equal temperament.
Similarly, for a software soundfont, I imagine the same holds true. So, perhaps some people discern these tuning compromises which are a part of the nature of the soundfont itself.

Regards.

EDIT: OOPS! 100 cents = a semitone, my math is off - time to call it a day.
The attached example is closer to a quarter tone than semitone. Anyway, it still works when you listen to it.

Attachment Size
Violin II test.mscz 14.26 KB

Reply

• Feb 4, 2017 - 21:51

I've been having the same issues with the high solo violin notes, and it seems like the solo viola does the same thing. A few lower note intervals sound a bit inaccurate too, but I'm not entirely sure about that.

Reply

• Mar 29, 2017 - 19:17

I am always using the latest version of MuseScore. MuseScore 2 comes with the soundfont FluidR3Mono_GM.sf3. If the latest version of this soundfont is not part of any musescore version yet, I probably don't use that, no. Where can I find the latest? Why has there been improvements on this specific problem?

Reply

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