Legacy ImageMagick Discussions Archive

Can you dig out the equivalent of +map's operations using actually documented features so I can switch to those?
Ideally I don't want to have the next person come along and have no way to find out what this command is doing.

Me trying to do that is what this whole thread's about, but changing CMYK to sRGB changes the appearance!

There's the problem - I have no idea what the uploader intends, and am trying to not change it from what they're likely to think it looks like.

Right, here are the ones starting with the CMYK original: http://postimage.org/image/dvwkfrxz5/ And here is the opposite problem, starting from an sRGB original (iPhone screenshot in Safari): http://postimage.org/image/5ybr18ltz/ As an aside, you can see Safari has done a horrible job, which is why ...

Hardly suprising that there is a massive difference in the pixel values. Then why did you mention it in the first place? Do you see the same difference in all viewers? If every image appears the same in all viewers, then yes, the differences are also all the same. Perhaps you can post a screenshot ...

Each image on its own appears consistently across all viewers (Microsoft Photo Viewer, Paint, Paint.NET, Explorer and Nautilus thumbnails, eog). Here's the ouput using the sRGB.icc from 6.7.6: $ compare -metric RMSE Channel_digital_image_CMYK_color.jpg out.jpg null: 88.7424 (0.348009) Using ...

I can see a significant difference on a variety of both Windows and Linux viewers. Additionally, using -colorspace RGB on that image gives output that does look exactly the same. Are you using the same version as me? I know colorspace conversions have been pretty broken previously. Have just checked ...

It certainly looks nicer, but the output still does not exactly match the colours of the input.
I don't want to affect the appearance at all - just standardise the representation for maximum compatibility.

the URL of a sample file

http://upload.wikimedia.org/wikipedia/e ... _color.jpg

That still ends up lightening CMYK input, whether there's a -strip or not.

I have input images from various sources in various colorspaces and need to standardise them, removing all embedded profile stuff, but leaving the resulting images looking the same. Basically I want to achieve what happens when you open them in MSPaint and re-save as a bitmap (then convert back to ...

Not in documentation but in my working production code (6.7.9-6) convert -dispose Background in-%02d.png[1-50] -set delay 10 -loop 0 -dither FloydSteinberg +map out.gif No warnings, and different results from +remap or omitting it. The high speed of +map vs. +remap is desirable, and I was after some ...

The two commands clearly do different things to an image sequence: +map is much faster and +remap gives more accurate colours in the output. However, there doesn't seem to be any documentation for +map (but it appears in lots of code around the place). What is different in the way these commands ...

For example, suppose I want to apply a convolution matrix (sharpening in this case) to every frame. In such a case you should really use ffmpeg's own filters - saving potentially massive amounts of time and space. ffmpeg -i input.mpg -filter:v "unsharp" output.mpg The "unsharp" filter has various ...

No, being able to redirect the stderr stream to arbitrary places does not help with a non-zero exit code.