Layout altered when score viewed online or downloaded as PDF
Hi there, I've got a little request: the layout of one of my recent scores is altered when viewed online or downloaded as PDF. https://musescore.com/user/409631/scores/1380671
I double-checked my mscz file and the whole score fits on an A4 page. This is not the case online, the score has two pages.
Your help will be appreciated.
Comments
Takes two pages on my system as well. Looks like you are using 2.0.1. What OS? There were some changes made between 2.0.1 and 2.0.2 to try to make font handling - which is a big part of these sort of occasional differences - more consistent between OS's.
So in theory, if you werre running 2.0.2, you'd see the same thing no matter which OS you were on, and it would match what you see on musescore.com (which I believe is a Linux system). In practice, there will probably still be occasional differences. But downloading 2.0.2 and making things look the way you like - reducing stretch, adding line breaks, etc - should help.
In reply to Takes two pages on my system by Marc Sabatella
Hi Marc, I'm using MuseScore on Mac OS X. I just downloaded version 2.0.2 and the score fits correctly on one page. I re-uploaded the file but it still takes two pages online and in PDF.
Is this a bug in the software?
In reply to Hi Marc, I'm using MuseScore by jbq
It does look the same on MuseScore.com and in MuseScore 2.0.2 and the latest development builds on Windows 7 (Ebterprise, 64bit).
Decreasing stretch of the measures 61 till the end by two notches makes it fit one page, see attached.
In reply to Hi Marc, I'm using MuseScore by jbq
As I said, even though 2.0.2 is much better than previous releases at identical layout between different OS's, no doubt occasional differences might remain. Which is why, as I said, you should add breaks and reduce stretch to "lock in" that layout to be safe. Reducing stretch won't change anything visibly on your system,but gives MuseScore more wiggle room should some font difference or whatever cause a system to require slightly more space on another person's computer (as opposed to pushing a measure to the next system which is what mught happen if it's just barely fitting currently).
In reply to As I said, even though 2.0.2 by Marc Sabatella
Hmm, pushing out 2 measures to the next page is more then just barely fitting?
In reply to Hmm, pushing out 2 measures by Jojo-Schmitz
Well, I can't see how it looks on his system, but I am assuming it didn't push two measures from the same system. Probably two *different* systems were just barely on the edge of fitting, and in both cases something about the rendering on musescore.com caused both of those systems to no longer fit, and so each system pushed one extra measure to the next system, or something like that. Or only one measure got pushed, but it got pushed in such a way that it displaced two measures from the *next* system. Or maybe the last system is supposed to have only two measures, but a line break elsewhere caused a si gle displacement to result in an extra system. Impossible to say for sure since I can't see how it looked originally.
As for whether it is a bug, it could be, or it could simply be an unavoidable consequence of a number of factors like:
- maybe some text font being used isn't available on musescore.com so another is substituted and it happens to be just enough larger to cause a problem
- maybe it is so close that it really comes down to different roundoff behaviors in floating point calculations on different hardware
-
In reply to Well, I can't see how it by Marc Sabatella
Playing around a little bit, it seems likely the issue is with the seventh system as shown on musescore.com - measures 40-45. This system is *just barely* not abke to fit one more measure. Selecting any one of those measures and reducing stretch one notch will result in one more measure fitting on that system, which then has a ripple effect that results in the last two measures (the last of which has already been reduced in stretch) fitting onto the last system.
It appears all measures have been stretched three notches already, presumably in an attempt to control where the line breaks occur. The better way to do this if desired is to simply add line breaks.
As for *why* that seventh system seems so sensitive, I do not believe it is floating point roundoff differences, as the numbers are close but not *that* close on my Linux system. We are testing to see if measure 46 will fit, and are finding the current minimum width of the system is 892.44320690543725, the minimum width of the measure we are testing is 193.03421370154166, and in order to fit, these need to add up to less than or equal to 1078.9119555263162. The actual sum is 1085.4774206069787. So it is too wide by about 7 units, which in this context is somewhat less than 1sp. It doesn't take much to change this calculation such that the next measure fits, but I don't think floating point roundoff explains it. More likely something involving a font calculation where the width of the notes themselves is slightly different, or something like that, is my guess.
Hi JB, just a short notice that we are investigating the issue. We'll come back as soon as we have an update.
In reply to Hi JB, just a short notice by Thomas
As we use Linux servers on musescore.com, your scores are rendered with Linux and not with Mac OS X. Normally the score rendering should be consistent and when it doesn't do this as in your case, we can consider this a bug. We are searching for the root cause of this issue.
In reply to Hi JB, just a short notice by Thomas
Hi JB, Werner found the root cause if the different layout rendering. It has to do with a different Freetype version on Linux vs Mac OS X. Werner is working on a fix now. Keeping you posted.
In reply to Hi JB, Werner found the root by Thomas
More info https://musescore.org/en/node/87861