Clarinet in FluidR3 only plays on the left channel and default Zita1 effect reduces the sound level too much
Hello,
I am a fan of MuseScore 1.3, and I have been following the development of 2.0 for several months now. I am using a relativley old HP Pavilion notebook running Windows 7, and in general, MuseScore has been running fine on it. I was happy to see that several of the early bugs of the 2.0 nightlies have been corrected in the 2.0 beta that I have installed lately, and the beta is now a very decent successor of 1.3. In Help->About it says it is revision 1efc609. Nevertheless, here are two problems that I have found trying to migrate scores from 1.3.
The first issue is that the clarinet in FluidR3Mono_GM.sf3 is single-channel (left only). I searched if anyone has already reported this, but did not find much, so I thought about reporting it. Am I the only one to hear this?
The other problem I have found is that in View->Synthesizer, if the Zita1 effect is selected (which I think it is, by default), the sound level is much, much lower than without it. If Zita1 is selected for both Effect A and Effect B, then the sound level almost drops to zero. This may reply to several users' reports about a too-low sound level in MuseScore 2.0.
The above two seem like bugs to me, however I thought I would discuss them before submitting a bug report. I hope this helps!
Angelos
Comments
These issues are somewhat be related, perhaps. The clarinets osunds in both channels for me in the default settings - which includes having Zita *on*. But if I turn zita1 off, then indeed, the clarinet goes to the left channel only.
In reply to These issues are somewhat be by Marc Sabatella
Hi Marc,
I think that the clarinet single-channel is an problem in the soundfount. It would make sense, since the program itselft does not distinguish the clarinet from any other instrument. When Zita1 is "on", it probably mixes the two channels and that's why the problem cannot be heard in that case. Can you also confirm the lower sound level issue?
Regards,
Angelos
In reply to Hi Marc, I think that the by Angelos Varvitsiotis
Hi Angelos.
Well spotted - when I checked the soundfont I had somehow missed setting the clarinet to pan central - it was still panned hard left.
That is now fixed, and will be sent up as a pull request for MuseScore 2.0 beta 2
If, in the meantime you would prefere to use the fixed version of the Soundfont you can download it from my Dropbox account.......
https://www.dropbox.com/s/k38u0yx89pj7d9e/FluidR3Mono%20GM2-9.zip?dl=0
It is the .SF2 version, as yet uncompressed.
In reply to Hi Angelos. Well spotted - by ChurchOrganist
Before making and commiting a new SF3, it would probably be good to review and merge these PR https://github.com/wschweer/sftools/pulls
In reply to Before making and commiting a by [DELETED] 5
Did that but the resulting compile on Windows results in corrupt SF3 files.
Leastways 2 attempts were made to compress the FLuidR3Mono.SF2 soundfont with it and both would not load into MuseScore 2. Beta1
Is this a Linux only thing? I currently don't have Github set up on my Ubuntu Studio VM
In reply to Did that but the resulting by ChurchOrganist
OK I have now got it compiled on my Ubuntu Studio VM - but it won't install.
Typing sudo make install into the terminal just results in the error message: No rule to install STOP
Anyone????
In reply to OK I have now got it compiled by ChurchOrganist
You don't really need to install the sftool.
Steps:
1- Clone sftool github repository;
2- Make sure that cmake and Qt bin folder are in the PATH;
3- "cd" to the sftool folder; "cmake CMakeLists.txt";
4- "make"; the executable "sfconvert" was created;
5- ./sfconvert -z [PATH/TO/Sf2] [name_of_output_sf3]
(also the "-s" option?)
Hope this helps.
Ciao,
ABL
In reply to You don't really need to by ABL
...mmmh...
It appears that also under Linux sfconvert is not generating a loadable sf3 file (at least, in my system).
It generates a loadable sf3 only if I revert the sftool fixes I proposed in my PR #4 (the ones regarding initialization of variables, free of memory allocations, et cetera).
Here is the sf3 generated when these changes are reverted (but keeping the commit from PR #6):
https://drive.google.com/file/d/0BxjayMZiuupObXZnT0FBOEFQNXc/view?usp=s…
I think the file corruption in this case may be related to the same failure under Windows (which was sometimes happening even before my PR). I will try to investigate.
In reply to ...mmmh... It appears that by ABL
I can confirm that sfconvert is not producing a loadable sf3 file.
I have just tried reverting PR#4 - will report back :)
In reply to I can confirm that sfconvert by ChurchOrganist
Yup reverting PR#4 did the trick.
I can now produce SF3 versions of soundfonts.
In reply to ...mmmh... It appears that by ABL
In sfont.h the initialization of QList zones { 0 }; produces a list with a zero Zone*.
Omit the {0} and it will work.
In reply to In sfont.h the initialization by [DELETED] 3
Shame on me, I overlooked the type of "zones" and treated it as a pointer!
Of course, the QList must be initialized as empty QList.
Thank you very much Werner :-)
Now the sfconvert should work properly under Linux (maybe also under Windows? I can't check at the moment)
In reply to Hi Angelos. Well spotted - by ChurchOrganist
A great thanks! I will use the SF2 version until beta2 becomes available. It being uncompressed is not an issue (for me). Thanks for sharing it over dropbox!
In reply to Hi Angelos. Well spotted - by ChurchOrganist
Apparently the new version of that soundfont didn't pass muster, Werner first merged it, but then reverted it again. Any idea why?
Edit: ah, he now reverted the revert so the new font is now in...
In reply to Hi Marc, I think that the by Angelos Varvitsiotis
I hear a slight reduction in overall volume with Zita, yes. That's pretty normal for "real" effects units as well in my experience. Specifics depends on the various settings of course.
In reply to I hear a slight reduction in by Marc Sabatella
Marc,
You can try "EffectA" and "EffectB" combined, and it will reduce the (perceived) sound voulme by at least 50%. Before spotting this as a root cause, I had searched and found several posts discussing why MuseScore 2.0 produces a lower volume than 1.3, but none of these pointed to the zita1 effect as the culprit. Probably some adjustment, either in the volume of the zita1, or on the defaults supplied (it may be that if you tune zita1 otherwise, the volume level increases, I have not tried) would produce a better default.
It is not my intention to insist on this being a bug. It may be the desired behavior. I only believe it is better to have it documented somewhere that the zita1 effect (with its default settings) reduces the output volume.
In reply to Marc, You can try "EffectA" by Angelos Varvitsiotis
Of course the reason for this is that Zita1 is being used as an insert effect across the master output of MuseScore. Consequently mixing the wet and dry signals will reduce volume. The only way you will get full volume is by using a 100% dry mix.
Of course in a studio situation this would be compensated for by either boosting the signal sent to Zita1 or turning up the amplifier at the end of the signal chain.
The way the I think the sound architecture of MuseScore 2 is currently organised means that the 2nd option is not available, so you are reliant on boosting the signal sent to Zita1 to compensate for the reduction in volume.
In reply to Of course the reason for this by ChurchOrganist
So if we keep Zita1 ON by default, I would boost the volume before Zita1 a bit so we have decent sound volume output for the users who know nothing about signal processing or recording techniques but just want the scores to sound "decent" by default.
In reply to So if we keep Zita1 ON by by [DELETED] 5
This seems logical to me, although another solution is to add a final amplification stage to Zita1. I have no idea how feasible that is. But I do vote in favor of keeping Zita1 on by default. I think it sounds good. I guess a few people have said it's too much reverb for them, and we can certainly tweak the specific settings, but I think it sounds better with than without.
FWIW, though, I find the difference in volume within and without Zita1 pretty small, and overall, I still find 2.0 *louder*, not *quieter*, than 1.3 using all default settings. I assume that's because we already have the default volume turned up higher.
In reply to This seems logical to me, by Marc Sabatella
And I also think Zita1 should be left ON by default.
Having no reverb at all would lead to a succession of complaints about sound quality!
In reply to And I also think Zita1 should by ChurchOrganist
I have not read all of this thread in detail. But if it is true that it is best to leave the current default Zita1 (ON), I think as much that the reverb effect is really very pronounced. Too much for some of us into thinking that it is a bug!
Like here: #37261: All Midi output is set to "let it ring"
This is also the case for some systems (on laptops for example). Is that I personally observed on my laptop (Asus, Windows8) with the result of having to disconnect Zita1. On my desktop computer, it's different.
So in the future, would there be a possibility of adding another effect "intermediate" between "No effect" and "Zita1", so with a more reasonable/soft level of reverb?
In reply to I have not read all of this by cadiz1
I don't understand!
The current level of Reverb on Zita1 is almost non existent.
Is it possible that your laptop has one of the soundcard "Enhancements" turned on? This is very often the case on Netbooks etc - first thing I always do with a new laptop is check that the soundcard is delivering a dry signal!
This would account for the fact that there is too much reverb.
Given that it is easy enough to adjust the wet/dry control on Zita1 to produce the amount of Reverb required, I don't see the problem.
In reply to I don't understand! The by ChurchOrganist
Where do you adjust the wet/dry control of Reverb on Zita1? It is not in Synthesizer, agree with that or not?
NB: the problem is that apparently I'm not alone in considering this question, through ignorance, it is possible.
For that, take the time to read and "listen" (in file on comment #3) the following thread:
#37261: All Midi output is set to "let it ring"
In reply to Where do you adjust the by cadiz1
The wet/dry control is adjustable from the Master effects page but see this bug #28776: Master effects interface disappears (when no plugins loaded) - Windows
The workaround is described in the issue tracker thread.
On listening to the file you mention, I do not find that to be an excessive amount of Reverb, unless you are looking for a completely dry sound, which is not waht most people would want, and shows up every imperfection in the soundfont.
If you are looking for better quality sound output then you should be thinking in terms of installing and configuring JACK and using external VST instruments and effects.
In reply to The wet/dry control is by ChurchOrganist
Thanks. I never don't use configuration with JACK and external VSB instruments. Not my "cup of tea".
In my opinion, my taste and my ears, in the previous file, the bass is drowned in reverb. No matter, the issue of reverb is a "chestnut" in all forums whatsoever.