I noticed that NegPy exported TIFFs often end up quite a bit larger than my original linear TIFFs, so I ran a few tests.
Results (same image):
Original TIFF (no compression): 128.6 MiB
NegPy LZW: 163 MiB
ImageMagick LZW: 140 MiB
ImageMagick ZIP: 108 MiB
Photoshop LZW: 161 MiB
Photoshop ZIP: 121 MiB
It seems like LZW performs quite poorly on this type of 16-bit scan data compared to ZIP compression, or even no compression, with file sizes growing larger than the original uncompressed tiff.
I also noticed that ZIP results vary quite a bit between tools, which is likely related to different TIFF predictor settings (horizontal differencing) and internal encoding/layout choices.
Maybe it would make sense to switch the default TIFF export to ZIP (with the most efficient predictor settings)?
(Support for lossless JEPG XL might also be a cool feature down the road, as it compresses quite a bit better still..)
I noticed that NegPy exported TIFFs often end up quite a bit larger than my original linear TIFFs, so I ran a few tests.
Results (same image):
Original TIFF (no compression): 128.6 MiB
NegPy LZW: 163 MiB
ImageMagick LZW: 140 MiB
ImageMagick ZIP: 108 MiB
Photoshop LZW: 161 MiB
Photoshop ZIP: 121 MiB
It seems like LZW performs quite poorly on this type of 16-bit scan data compared to ZIP compression, or even no compression, with file sizes growing larger than the original uncompressed tiff.
I also noticed that ZIP results vary quite a bit between tools, which is likely related to different TIFF predictor settings (horizontal differencing) and internal encoding/layout choices.
Maybe it would make sense to switch the default TIFF export to ZIP (with the most efficient predictor settings)?
(Support for lossless JEPG XL might also be a cool feature down the road, as it compresses quite a bit better still..)