#131 SVG in LibreOffice under MacOS

[maaaaaaaartin]

Dear all:

First of all, many thanks for this wonderfull piece of software, which is really critical for my academic life—the only reason I keep up with LibreOffice actually.

I noticed that SVG equations now look horrible in presentation mode (they look fine in normal mode) in LibreOffice (6.4) under MacOS (10.15). I tried with and without OpenGL. Is there anything I can do to make it work? Thanks!

Martin

[Roland Baudin]

The LibreOffice team has probably changed the SVG API in version 6.4... I suggest you stick on the 6.3.x version.

[maaaaaaaartin]

Thank you! Unfortunately it also occurs with LO 6.3.4… Any idea where it
could come from? Thanks again!

[Roland Baudin]

TexMaths 0.48 should work as expected with LibreOffice 6.3.4.2 in MacOS X.

Could you should try to rename your LibreOffice profile and then reinstall TexMaths?

[maaaaaaaartin]

I think TeXMath works perfectly. It is just that the generated SVG, even
though they look perfect under Normal View, look ugly in Presentation
mode. Just created a new profile, installed TeXMath again, and am
getting the same result. Any idea where could it come from? Thanks anyway.

[Roland Baudin]

That's strange... Perhaps you could try to download again LibreOffice and reinstall it. Or try to use the previous version (6.2.x). What is your system version?

[maaaaaaaartin]

Very strange indeed.

Even stranger: on the presenter screen (i.e., the main screen, not the
external), SVG equations are rendered perfectly on the "next slide"
panel, but not on the "current slide" panel.

I use MacOS 10.15. I tried with LO 6.3 & 6.2. If I activate OpenGL it is
slightly better, but definitely not quite perfect.

Thanks anyway for your help.

[Roland Baudin]

Could you retest with TexMaths 0.49 in MacOS?

[maaaaaaaartin]

Unfortunately it is still the same w/ LO 7.1.1 & TeXMath 0.49, cf. attached screenshots. Upper equation is PNG 1200, lower is SVG. First screenshot is in view mode, second is in presenter mode. So weird. Thanks for your help!!

[Roland Baudin]

I did some tests in Linux, Windows and MacOS and seems there is some distortion of SVG images in presentation mode in Windows and MacOS. I don't see any distortion in Linux.

So perhaps, it's time to make a bug report for LibreOffice.

Could you send me a document example where the distortion is apparent?
Thanks.

[maaaaaaaartin]

Sorry for the delay. Here is an example with twice the same equation, the upper one being rendered in PNG (1200 DPI) while the lower one is rendered in SVG. Both equations look nice in normal view, but the lower one is distorted in presentation mode. Many thanks for your help!

[Roland Baudin]

Thanks for the document. However, when I compile the equations, I don't get the same result as you. Did you changed the preamble ?
Which TexMaths version do you use?

[maaaaaaaartin]

Using 0.49.
Here is my preamble:

\usepackage{amsmath}
\usepackage{amssymb}
\usepackage{mathtools}
\usepackage{bbm}
\usepackage{stmaryrd}
\usepackage{blkarray}
\usepackage{mathdots}

%%% sans serif
\renewcommand\familydefault{\sfdefault}
\everymath{\mathsf{\xdef\mysf{\mathgroup\the\mathgroup\relax}}\mysf}
\usepackage{mathastext}

%%% sans serif bb
\usepackage{bbold}

%%% upright greek letters
\usepackage{upgreek}
\renewcommand{\alpha}{\upalpha}
\renewcommand{\beta}{\upbeta}
\renewcommand{\gamma}{\upgamma}
\renewcommand{\delta}{\updelta}
\renewcommand{\epsilon}{\upepsilon}
\renewcommand{\zeta}{\upzeta}
\renewcommand{\eta}{\upeta}
\renewcommand{\theta}{\uptheta}
\renewcommand{\kappa}{\upkappa}
\renewcommand{\lambda}{\uplambda}
\renewcommand{\mu}{\upmu}
\renewcommand{\nu}{\upnu}
\renewcommand{\xi}{\upxi}
\renewcommand{\pi}{\uppi}
\renewcommand{\rho}{\uprho}
\renewcommand{\sigma}{\upsigma}
\renewcommand{\tau}{\uptau}
\renewcommand{\chi}{\upchi}
\renewcommand{\psi}{\uppsi}
\renewcommand{\omega}{\upomega}

%%% color
\usepackage[usenames]{xcolor}
\definecolor{Xblue}{RGB}{2,59,86}
\color{Xblue}

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

\newcommand{\ds}{\displaystyle}

\newcommand{\Chi}{\mathlarger{\mathlarger{\chi}}}

\newcommand{\p}[1]{\left( #1 \right)}
\newcommand{\rb}[1]{\left( #1 \right)}
\renewcommand{\sb}[1]{\left[ #1 \right]}
\newcommand{\cb}[1]{\left\{ #1 \right\}}
\newcommand{\norm}[1]{\left\| #1 \right\|}
\newcommand{\tI}[1]{\underline{ #1 }}
\newcommand{\tII}[1]{\underline{\underline{ #1 }}}
\newcommand{\tIII}[1]{\underline{\underline{\underline{ #1 }}}}
\newcommand{\tIIII}[1]{\underline{\underline{\underline{\underline{ #1 }}}}}
\renewcommand{\u}[1]{\underline{ #1 }}
\newcommand{\uu}[1]{\underline{\underline{ #1 }}}
\newcommand{\uuu}[1]{\underline{\underline{\underline{ #1 }}}}
\newcommand{\uuuu}[1]{\underline{\underline{\underline{\underline{ #1 }}}}}
\newcommand{\dint}{\displaystyle\int}
\newcommand{\jump}[1]{\left[\!\left[ #1 \right]\!\right]}
\newcommand{\ljump}[1]{\left[\!\left[ #1 \right.\right.}
\newcommand{\rjump}[1]{\left.\left. #1 \right]\!\right]}
\newcommand{\diff}[2]{\frac{\partial #1}{\partial #2}}
\newcommand{\diffI}[2]{\frac{\partial   #1}{\partial #2   }}
\newcommand{\diffII}[2]{\frac{\partial^2 #1}{\partial #2 ^2}}
\newcommand{\diffIII}[2]{\frac{\partial^3 #1}{\partial #2 ^3}}
\newcommand{\ddiff}[2]{\dfrac{\partial #1}{\partial #2}}
\newcommand{\ddiffI}[2]{\dfrac{\partial   #1}{\partial #2   }}
\newcommand{\ddiffII}[2]{\dfrac{\partial^2 #1}{\partial #2 ^2}}
\newcommand{\ddiffIII}[2]{\dfrac{\partial^3 #1}{\partial #2 ^3}}

\renewcommand{\det}[1]{\text{det}\p{ #1 }}
\newcommand{\Det}[1]{\text{Det}\p{ #1 }}
\newcommand{\tr}[1]{\text{tr}\p{ #1 }}
\newcommand{\Tr}[1]{\text{Tr}\p{ #1 }}
\newcommand{\ugrad}[1]{\u{\text{grad}}\p{ #1 }}
\newcommand{\uGrad}[1]{\u{\text{Grad}}\p{ #1 }}
\newcommand{\uugrad}[1]{\uu{\text{grad}}\p{ #1 }}
\newcommand{\uuGrad}[1]{\uu{\text{Grad}}\p{ #1 }}
\renewcommand{\div}[1]{\text{div}\p{ #1 }}
\newcommand{\Div}[1]{\text{Div}\p{ #1 }}
\newcommand{\udiv}[1]{\u{\text{div}}\p{ #1 }}
\newcommand{\uDiv}[1]{\u{\text{Div}}\p{ #1 }}
\newcommand{\urot}[1]{\u{\text{rot}}\p{ #1 }}
\newcommand{\uRot}[1]{\u{\text{Rot}}\p{ #1 }}

\newcommand{\ofx}{\p{x}}
\newcommand{\ofX}{\p{X}}
\newcommand{\ofvx}{\p{\tI{x}}}
\newcommand{\ofvX}{\p{\tI{X}}}

\DeclareMathOperator*{\argmax}{argmax}
\DeclareMathOperator*{\argmin}{argmin}

[Roland Baudin]

OK, thanks, now I can recompile the equation.

The problem is a known LibreOffice bug (https://bugs.documentfoundation.org/show_bug.cgi?id=150531).

I've submitted your example to the LO debug team.

As a workaround, in Windows and Mac OS X, you can disable Skia rendering. For that, in Writer go to Tools / Options / View and uncheck Skia rendering (but keep Hardware Acceleration enabled).