Typos in the command line option page for V 6

[ronsavage]

I don't wish to come across as too pedantic here. After all, you don't have to make any of these hoped-for changes.

Does '-deskew threshold' take an optional %?

The text uses both 'gray-scale' and 'grayscale'.

Under '-distort method arguments', the text says '-define distort:scale={scale_factor}'. Is it correct to assume the {} can be dropped from that line?

The text for '-displace horizontal-scalexvertical-scale' mentions % and !, but neither of the 2 summary lines does.

For '-dissolve src_percent[xdst_percent]', the text reads as though % is not accepted as part of the input. Is that correct?

For '-modulate brightness[,saturation,hue]', the same comment applies as for -dissolve.

Under '-normalize', 'came amount' should read 'same amount'.

For '-sepia-tone threshold', is the % optional?

For '-sigmoidal-contrast contrastxmid-point', also, is the % optional?

Likewise for '-solarize threshold'?

For each of the following, it'd be nice to see another summary line using the '+' prefix, since '+' is discussed in the text (although there may be too many here to bother with):

-adjoin
-antialias
-append
-channel type
-clip
-clip-mask
-clip-path id
-clone index(s)
-compress type
-contrast
etc, etc. Just scan for + to see the rest.

[magick]

Fixed most of your suggestions, update available sometime tomorrow. Thanks. Note, you can issue a GIT pull request on the documentation and make the fixes yourself.

[fmw42]

The use of gray-scale should only be there for descriptions. I expect there may be places that use grayscale and gray-scale for descriptions. The function -grayscale should not have and hyphen separating the two words. If you see any, let us know.

The only function that I know that requires a % symbol is the -deskew value. It does not seem to work without it. I believe that -sigmoidal-contrast typically uses the percent. If it is left off, then it uses raw graylevel values in Q range of the compile. Likewise -solarize and -sepia-tone. It is explained for -sepia-tone in the description.

% are optional for only those functions that refer to grayscale value, which can be either raw values (in the Q range of the compile) or percent values. If the latter, the % is needed to so indicate that. These are functions such as -level, -contrast-stretch, -normalize, -sigmoidal-contrast, -solarize, -sepia-tone, etc). Note that -colorize will use a value between 0 and 100 with or without the %.

-dissolve and -modulate do not expect % symbols and may fail if they are included. -dissolve and -blend are old syntax used with composite. The newer syntax uses convert ... -compose dissolve/blend without the minus. In either case, they are percent concepts for mixing images and thus do not need %. They may still work with them, but I have not tested that.

Many of the +formats simply turn off the -formats (antialias, clip, etc. I am unaware of what +contrast might do. Most of the + formats are discussed in the - format section and so do not need repeating. I would hope that any + forms have some description. If not, let us know. But I doubt this will be changed to include a separate description for the +forms.

If Magick tells me to work on the options page for the spelling and bracket/braces inconsistency, I will try to clarify as many of the above as I can.

EDIT: I see he just posted a note that he fixed most of these. So I will hold off and see what changes are made and what is missing.

Let us know if you find any other inconsistencies or typos or errors.

[ronsavage]

Thanx for the quick response.

Many years (decades?) ago I was allowed to log on to IM and made some doc changes, but they were obliterated a couple of versions later, so I've been deeply reluctant to make any since but I do appreciate the availability of it on github, since I use github myself (https://github.com/ronsavage?tab=repositories).

[ronsavage]

Re the table for '-ordered-dither threshold_map{,level...}', do I read it correctly that columns 1 and 3 only are valid inputs?

[fmw42]

Re the table for '-ordered-dither threshold_map{,level...}', do I read it correctly that columns 1 and 3 only are valid inputs?

As far as I know only column 1 is used. The other columns are descriptions as far as I understand. I do not know about the optional argument, level. I am not an expert on dithering and the only explanation I know is at http://www.imagemagick.org/Usage/quanti ... red-dither.

glennrp would probably be the one who can explain this best.

[ronsavage]

OK. Thanx.

[snibgo]

I'll add: IM doesn't seem to have a function that can parse and execute what we might call a "single-image operation", such as:

Code: Select all

-resize "10x200" -blur 0x0.5 -background White -extent 500x500

There is a function that will take all the "convert" arguments and execute them, but that needs input and output files.

[fmw42]

snibgo: Does't IM 7 scripting do what you suggest. See viewtopic.php?f=1&t=28884

[magick]

And a reminder, IMv7 will have a general availability release on April 30th of this year. Its currently in Beta.

[snibgo]

I don't know if V7 has a function that we can call from software that does as I suggest. V6 doesn't seem to have such a function.

[fmw42]

IM 7 has -script. See example at viewtopic.php?f=1&t=28884

Not sure it is exactly what you are asking about.

[snibgo]

That's for the command-line interface. Perhaps V7 also provides a function that can takes strings like "-blur 0x0.5 -background White -extent 500x500" and apply them to an image.

[fmw42]

If not the command line, what are you interested in?

[snibgo]

I've written a Windows GUI (called "Swim") for IM. It's a standalone program, but also a process module so it can be run from inside a "convert" command, and operates like any other IM process on all the images in the list.

It operates on images in memory, and doesn't run other programs or do any file I/O so it is very fast with smooth updates to images when sliders are moved.

Currently, all Swim operations are hard-coded, calling IM at the ImageCore level. This works fine, but it could be more general-purpose if a function that I describe was available. It could be like ConvertImageCommand() but without the need to name input or output images.

This would simplify my life, and anyone who is writing high-level interface to IM.

Of course, I could write it myself, but it would be a maintenance nightmare keeping up with changes to IM.

EDIT: I should add: it might help the OP ronsavage who is "writing a Perl module which parses IM command lines" for users who prototype using "convert".