Convert DDS to PNG using linux command line
2013-08
I need to convert thousands of DDS images to PNG format in Linux, preferably in command line.
Is there any program available for such task?
ImageMagick reads but doesn't write DDS. And of course it reads and writes PNG.
From identify -list format
:
...
DDS* DDS r-- Microsoft DirectDraw Surface
...
PNG* PNG rw- Portable Network Graphics (libpng 1.2.37)
...
To convert a file (leaving the original intact):
convert test.dds test.png
To convert a directory full:
for file in *.dds
do
convert "$file" "$(basename "$file" .dds).png"
done
Does anyone know a good way to batch-convert a bunch of PNGs into JPGs in linux? (I'm using Ubuntu).
A png2jpg binary that I could just drop into a shell script would be ideal.
Your best bet would be to use Imagemagick
I am not an expert in the actual usage, but I know you can pretty much do anything image related with this!
An example is:
convert image.png image.jpg
and it will keep the original as well as creating the converted image. As for batch. I think you need to use the Mogrify tool (from the same command line when in imagemagick). Keep in mind that this overwrites the old images.
The command is:
mogrify -format jpg *.png
The convert
command found on many Linux distributions is installed as part of the ImageMagick suite. Here's the bash code to run convert
on all PNG files in a directory and avoid that double extension problem:
for img in *.png; do
filename=${img%.*}
convert "$filename.png" "$filename.jpg"
done
The actual "png2jpg
" command you are looking for is in reality split into two commands called pngtopnm
and cjpeg
, and they are part of the netpbm
and libjpeg-progs
packages, respectively.
png2pnm foo.png | cjpeg > foo.jpeg
I have a couple more solutions.
The simplest solution is like most already posted. A simple bash for loop.
for i in *.png ; do convert "$i" "${i%.*}.jpg" ; done
For some reason I tend to avoid loops in bash so here is a more unixy xargs approach, using bash for the name-mangling.
ls -1 *.png | xargs -n 1 bash -c 'echo "$0" "${0%.*}.jpg"'
The one I use. It uses GNU Parallel to run multiple jobs at once, giving you a performance boost. It is installed by default on many systems and is almost definitely in your repo (it is a good program to have around).
parallel convert '{}' '{.}.jpg' ::: *.png
The number of jobs defaults to the number of processes you have. I found better CPU usage using 3 jobs on my dual-core system.
parallel -j 3 convert '{}' '{.}.jpg' ::: *.png
And if you want some stats (an ETA, jobs completed, average time per job...)
parallel --eta convert '{}' '{.}.jpg' ::: *.png
The final command that I use looks like this for refrence. I use ls
rather than parallel's syntax for who-knows-what reason.
ls -1 *.png | parallel -j 3 --eta convert '{}' '{.}.jpg'
For batch processing:
for img in *.png; do
convert "$img" "$img.jpg"
done
You will end up with file names like image1.png.jpg though.
This will work in bash, and maybe bourne. I don't know about other shells, but the only difference would likely be the loop syntax.
my quick solution
for i in $(ls | grep .png); do convert $i $(echo $i.jpg | sed s/.png//g); done