Re: keeping file sizes down -colors -depth
Posted: 2009-05-18T16:56:08-07:00
More than likely the PNG created by montage has 16 bit values in it and prosibly an alpha channel. Of course this is only a guess on my part. You "identify" to check.
Filtering the image through TIFF probebly trunctated the color values to 8 bit per colro value and removed the unneeded alpha channel, thus causing the new PNG to be 3/8 th's the size of the previous TIFF (reduced to 3/4 by lossing alpha, and then halving that by reducing depth).
The problem with "montage" is that when the end of the command is reached (basically equivelent to a "convert" with a few modifications) "montage" continues to process the in memory images, to produce the "montage". Unfortunately their is currently no way to modify the final montage image such as setting depth, or removing the alpha channel, before it saves the result.
I would like to eventually see a "-montage" or "-layers montage" type option that replicates the existing montage operation.
But before that is done many of the existing intermediate montage image processing steps would also need to be moved from the program merged into the ImageMagick core library. That is create operations such, performing the 'montage framing', and 'grid layout' in the core library. It is however not straight forward as metadata needed for constructing 'HTML image maps' also needs to be preserved as part of the 'layout' image processing.
Filtering the image through TIFF probebly trunctated the color values to 8 bit per colro value and removed the unneeded alpha channel, thus causing the new PNG to be 3/8 th's the size of the previous TIFF (reduced to 3/4 by lossing alpha, and then halving that by reducing depth).
The problem with "montage" is that when the end of the command is reached (basically equivelent to a "convert" with a few modifications) "montage" continues to process the in memory images, to produce the "montage". Unfortunately their is currently no way to modify the final montage image such as setting depth, or removing the alpha channel, before it saves the result.
I would like to eventually see a "-montage" or "-layers montage" type option that replicates the existing montage operation.
But before that is done many of the existing intermediate montage image processing steps would also need to be moved from the program merged into the ImageMagick core library. That is create operations such, performing the 'montage framing', and 'grid layout' in the core library. It is however not straight forward as metadata needed for constructing 'HTML image maps' also needs to be preserved as part of the 'layout' image processing.