MTPRO2 with Y&Y tex - my results

For feedback and discussion for testers of the MathTime Professional II fonts.

Moderators: PTIForAdmin, WaS, Michael Spivak

gvokalek
Posts: 3
Joined: Thu Mar 09, 2006 1:05 am
Location: Australia
Contact:

MTPRO2 with Y&Y tex - my results

Post by gvokalek »

Following in murray's footsteps, here are my results with Y&Y.

Here is my platform:
Y&Y TeX 2.2.3 2001 May 22
DVIWindo (32) Previewer 2.2.6
DVIPSONE 2.2.6 2001 Jun 23
Platform: Win2k SP4 on P4 with 512MB

I installed the fonts manually after unzipping. The stylefile, fontdefs
and tfms went into the right places, and the mtpro2 pfb files into c:\psfonts
where other tex .pfb's reside.

I then copied

\texmf\doc\fonts\mtpro2\Test_latex.tex

to a working directory and latex'd it with:

yandytex.exe +latex -v -L test_latex

y&y yielded usual tex verbosity with no warnings.

I then tried the y&y previewer DVIWindo which yielded
garbled results, presumably because win2k does not know
about the mathtime fonts. I've seen this behaviour in the
past for fonts which lack .pfm files. There were no .pfm's
in the MTPRO2 zip.

I then generated PS using:

dvipsone -l=letter Test_latex.dvi

which gave the following warnings:

** No /.notdef line* No /.notdef line* No /.notdef line* No /.notdef line* No /.
notdef line*** No /.notdef line* No /.notdef line* No /.notdef line* No /.notdef
line* No /.notdef line* No /.notdef line** No /.notdef line...

These warnings were generated in response to the mtpro2 fonts, but
I don't have enough knowledge to diagnose further. If I switch back to default fonts, the warnings disappear. I have also seen similar warnings
caused by other fonts when using a RIP, so I am sure the warnings are
caused by the fonts.

I then distilled the PS file with acrobat 5 distiller. No problem there.

I then compared my PDF to the PDF in the MTPRO2 zip file.

The files compared well. All characters present and accounted for.

HOWEVER : I am surprised by the small variations in character positions
between the two PDF files. I would expect tex/dvips to produce identical
positioning, but there are subtle shifts in some characters but not in
others. Fortunately for me, my version looks better :-)

Conclusion:

1) good result

2) need PFM files to make the previewer work correctly.

3) something in the fonts needs to be fixed to prevent the .notdef
warnings.

George.
Michael Spivak
Posts: 52
Joined: Mon Oct 10, 2005 2:10 pm

Re: MTPRO2 with Y&Y tex - my results

Post by Michael Spivak »

gvokalek wrote:
HOWEVER : I am surprised by the small variations in character positions
between the two PDF files. I would expect tex/dvips to produce identical
positioning, but there are subtle shifts in some characters but not in
others. Fortunately for me, my version looks better :-)

3) something in the fonts needs to be fixed to prevent the .notdef
warnings.

George.
I am separately emailing you new pfbs, new tfms, also pfms, and slightly
revised style file mtpro2.sty. This should fix (almost) all problems. [The .notdef warnings was not because something was wrong with the
fonts but because something is silly with the previewer]

The only thing worrying me is the thing about small variations in character positions in the pdf files. If that persists, please let me know.
gvokalek
Posts: 3
Joined: Thu Mar 09, 2006 1:05 am
Location: Australia
Contact:

Re: MTPRO2 with Y&Y tex - my results

Post by gvokalek »

Michael Spivak wrote: The .notdef warnings was not because something was wrong with the
fonts but because something is silly with the previewer
Michael, please tell me the cause of the .notdef warning, because I
see it when other fonts pass through a RIP which I use occasionally.
I'd like to know the cause.
Michael Spivak wrote: The only thing worrying me is the thing about small variations in character positions in the pdf files. If that persists, please let me know.
These variations are still present. I have prepared test files which you can download from:

http://www.causalproductions.com/JOBS/ ... ution.pdf
This is the original document from the distribution zip.

http://www.causalproductions.com/JOBS/ ... y_tex.pdf
This is a PDF I prepared using Y&Y and Acrobat 5

http://www.causalproductions.com/JOBS/ ... rison.gif
This is a 2-image animated GIF of the top left hand area of page 1. Use a GIF viewer which plays the animation, and you will see the text slightly move as the images alternate.

Another way to see the variations is to open both files in full-window mode in Acrobat Reader, then use CTRL-TAB to flip between the two files, and let your eye spot the slight movements in characters.

I don't know enough about tex internals to know if these variations are a problem or not.

Let me know which other files you need to follow up this issue (eg .dvi file ?)
Michael Spivak
Posts: 52
Joined: Mon Oct 10, 2005 2:10 pm

Re: MTPRO2 with Y&Y tex - my results

Post by Michael Spivak »

gvokalek wrote:
Michael Spivak wrote: The .notdef warnings was not because something was wrong with the
fonts but because something is silly with the previewer
Michael, please tell me the cause of the .notdef warning, because I
see it when other fonts pass through a RIP which I use occasionally.
I'd like to know the cause.
The previewer wants to make sure that every position that doesn't have an actual character will instead have the character /.notdef listed. So it insists
on looking for an initial line that puts /.notdef in all positions before the
characters are inserted. This line

0 1 255 {1 index exch /.notdef put} for

should thus occur right after the line

/Encoding 256 array

Fontographer (and probably other programs) don't put in this line, they
simply make sure to put /.notdef everywhere it should be. (Of course, if you're getting the problem with other RIPs and other fonts, it might even be that those fonts don't make sure to have /.notdef everywhere, and so the RIP is giving a valid gripe.)

The only way to get around this, as far as I know, is to use pfbtopfa
to get a .pfa file from your .pfb file, then add the necessary line with
an editor, and then use pfatopfb to remake the pfb file. Probably one can
make a simple program, or a script file, to do this automatically, but I
just did it by hand.
Michael Spivak
Posts: 52
Joined: Mon Oct 10, 2005 2:10 pm

Re: MTPRO2 with Y&Y tex - my results

Post by Michael Spivak »

gvokalek wrote:
Michael Spivak wrote: The only thing worrying me is the thing about small variations in character positions in the pdf files. If that persists, please let me know.
These variations are still present.


http://www.causalproductions.com/JOBS/M ... arison.gif

This is a 2-image animated GIF of the top left hand area of page 1. Use a GIF viewer which plays the animation, and you will see the text slightly move as the images alternate.

Another way to see the variations is to open both files in full-window mode in Acrobat Reader, then use CTRL-TAB to flip between the two files, and let your eye spot the slight movements in characters.

I don't know enough about tex internals to know if these variations are a problem or not.
I suspect that this is due to the differences in the way dvips and Y&Y's
divpsone work. At one time, and perhaps still now, dvips would put down several characters from a word at a time, using the information from the tfm file to figure out where they should go, and then at various times, adjusting things to make sure the final position agreed with what it should be based on the widths in the PostScript font file (or something like that!),
whereas Y&Y uses only the information in the PostScript font file (thereby
avoiding the potentially devastating possibility that a typesetter printing
from the ps file might have a tfm file for one of the fonts differing from the
tfm file that the user had---of course, this is pretty irrelevant now, since
people just send .pdf files, not .ps files, to a printer).

If this is what's happening, then I guess there's really nothing to be done about it. And it's not really a problem, so long as line breaks and page breaks are the same. What really worried me is that murray said there was even a different page break somewhere.
Michael Spivak
Posts: 52
Joined: Mon Oct 10, 2005 2:10 pm

Re: MTPRO2 with Y&Y tex - my results

Post by Michael Spivak »

Michael Spivak wrote:
gvokalek wrote:
Michael Spivak wrote: The only thing worrying me is the thing about small variations in character positions in the pdf files. If that persists, please let me know.
These variations are still present.


http://www.causalproductions.com/JOBS/M ... arison.gif

This is a 2-image animated GIF of the top left hand area of page 1. Use a GIF viewer which plays the animation, and you will see the text slightly move as the images alternate.

Another way to see the variations is to open both files in full-window mode in Acrobat Reader, then use CTRL-TAB to flip between the two files, and let your eye spot the slight movements in characters.

I don't know enough about tex internals to know if these variations are a problem or not.
I suspect that this is due to the differences in the way dvips and Y&Y's
divpsone work. At one time, and perhaps still now, dvips would put down several characters from a word at a time, using the information from the tfm file to figure out where they should go, and then at various times, adjusting things to make sure the final position agreed with what it should be based on the widths in the PostScript font file (or something like that!),
whereas Y&Y uses only the information in the PostScript font file (thereby
avoiding the potentially devastating possibility that a typesetter printing
from the ps file might have a tfm file for one of the fonts differing from the
tfm file that the user had---of course, this is pretty irrelevant now, since
people just send .pdf files, not .ps files, to a printer).

If this is what's happening, then I guess there's really nothing to be done about it. And it's not really a problem, so long as line breaks and page breaks are the same. What really worried me is that murray said there was even a different page break somewhere.
ADDED: Note that the differences occur even a line that is completely in Computer Modern, so it has nothing to do with the MTPRoII fonts, it is
just something about the different RIPS.
Locked