TIF to JPEG sizing query?

If you would like to post, you'll need to register. Note that if you have a BCG store account, you'll need a new, separate account here (we keep the two sites separate for security purposes).

ruley74

Well-known member
Hi there,

Curious on a situation i have atm, in the image below I wanted to extend the canvas size above and in front of the Egret so I saved the TIF and exported to PS for that task. given the size of the TIF once completed I saved as 100% quality JPEG but the file size is only 3.6M. I also saved from PS as a TIF and saved as JPEG in Topaz Denoise once I cleaned it up in there....

Is it sheerly the fact that it's got a lot of black that the file is so small?

Reason I ask is I'm going to print this one.

Regards,
D50_8530insta.jpg
You can only see EXIF info for this image if you are logged in.
 
Is it sheerly the fact that it's got a lot of black that the file is so small?
Yes, jpeg compression works by identifying very similar pixels and not saving full RGB representations (8 bits for each color) when pixels are not noticeably different from one another. In the case of an all black or all white background with no detail many pixels are literally identical to one another so the jpeg compression savings are large. IOW for perhaps half the pixels in that image the jpeg compression algorithm doesn't have to save 8 bits of red, 8 bits of green and 8 bits of blue data but just saves a pointer to the previous totally black pixel which is a pretty substantial file size reduction in that kind of image.

The literal process of compression is a bit more complicated and involves breaking the colors into a different color space to separate lightness data from chroma data and then breaking the image into small blocks (8x8 pixels) that are individually compressed but in the case of an image with large completely black or completely white areas the result is the same, you get a ton of compression.

That said, I save my files destined for printing as TIFFs which can be compressed with no data loss (LZW or ZIP compression) instead of jpegs that use lossy compression algorithms (the original data cannot be completely reconstructed from the compressed files). In the case of the black areas of your photo above it really doesn't make any difference but each successive jpeg save robs a bit of detail from the portions of the image that actually have detail. In many cases it really doesn't matter and probably won't make any difference in the image above especially if you only did one (or a few) jpeg saves from a photo editor but it's usually just as easy to save it as a TIFF with lossless compression and most print houses will accept compressed TIFFs.

Looking at the pixel dimensions of an image is a much better way to assess how large you can print it compared to just looking at file size for any kind of compressed files including LZW or ZIP compressed TIF files but also for jpegs that are always compressed.
 
Last edited:
Yes, jpeg compression works by identifying very similar pixels and not saving full RGB representations (8 bits for each color) when pixels are not noticeably different from one another. In the case of an all black or all white background with no detail many pixels are literally identical to one another so the jpeg compression savings are large. IOW for perhaps half the pixels in that image the jpeg compression algorithm doesn't have to save 8 bits of red, 8 bits of green and 8 bits of blue data but just saves a pointer to the previous totally black pixel which is a pretty substantial file size reduction in that kind of image.

The literal process of compression is a bit more complicated and involves breaking the colors into a different color space to separate lightness data from chroma data and then breaking the image into small blocks (8x8 pixels) that are individually compressed but in the case of an image with large completely black or completely white areas the result is the same, you get a ton of compression.

That said, I save my files destined for printing as TIFFs which can be compressed with no data loss (LZW or ZIP compression) instead of jpegs that use lossy compression algorithms (the original data cannot be completely reconstructed from the compressed files). In the case of the black areas of your photo above it really doesn't make any difference but each successive jpeg save robs a bit of detail from the portions of the image that actually have detail. In many cases it really doesn't matter and probably won't make any difference in the image above especially if you only did one (or a few) jpeg saves from a photo editor but it's usually just as easy to save it as a TIFF with lossless compression and most print houses will accept compressed TIFFs.

Looking at the pixel dimensions of an image is a much better way to assess how large you can print it compared to just looking at file size for any kind of compressed files including LZW or ZIP compressed TIF files but also for jpegs that are always compressed.
Thank you for taking the time to explain, makes sense.

Yeah I only did it once just to print as the TIF was 100MB and the print place only took up to 25MB files online. The pixel dimensions were correct, I did check that to see if something funny was going on.

Unsure if the software makes a difference regards size of TIF, I'm using DXO PL3 for the RAW conversion and saving as TIF for PS work.
 
Unsure if the software makes a difference regards size of TIF, I'm using DXO PL3 for the RAW conversion and saving as TIF for PS work.
I don't use DXO PL3 so I don't know their save dialog box, but programs like PS give you several options for saving TIFs: Uncompressed, Lossless LZW compressed and Lossless ZIP compressed. Uncompressed TIFs can be huge for big files especially if they're saved in 16 bit format but either LZW or ZIP compression can dramatically reduce the file size and similar to jpeg compression those large totally black areas will mean the compressed files get a lot smaller.

The other things that can result in overly large TIF files is maintaining your Photoshop layer stack instead of flattening the image or saving the TIF in 16 bit vs 8 bit mode. If I'm saving a file for future editing I'll maintain both the layer structure and the high bit data representation but when I send a file out to a print house it will be flattened and converted to 8 bit mode. The big advantage of high bit representation is during the editing where there's much less chance of banding and posterization when working with higher bit depths when doing things like: contrast, saturation and color adjustments. But 8 bit representation is plenty for printing the final image.

Unless you're printing a really large image I'd be surprised if your flattened, 8 bit, LZW compressed TIF was over 25 MB.
 
Not a problem with the white egret but with other subjects going to 8-bit TIFF will compress the tonal range and details will be lost. I do wish I could save as 12-bit TIFF instead of 16-bit and have much smaller files to manage. But if I am uploading to a lab's FTP server the file size does not really matter.
 
Back
Top