Lossless Compression for several pictures

Hi there,

This might sound very simple but it seems I’m still a beginner with RT, and even i try to find my way by myself, I’d like to know if there is a good way for this :

I need to compress 25 pictures in a .zip folder of less than 10 Mo.

I treated my Raws in RT, and if i want them to fit in less than 10Mo, the quality has to be reduced to 10 in the export settings. It looks really bad in jpg (pink artefacts and pixels everywhere)…

What would you recomand for better quality and lower file size ?

Thank you !

Hello @Little_Navy
In that case, when I want to send a significant quantity of photos or upload them on some site I usualyy do that
1- in RT, save photos with high quality jpg
2- I use xnview photo viewer ( but all other viewer are ok for that)
3- I make a first trial on a few photos in xnview batch function
3.1 reduce the definition (resize function) to expected display definition (that can be done in RT but you have to process twice to keep high quality)
3.2 reduce progressivly jpeg quality until the foreseen space is ok

  1. then you can batch process all your photos with above parameters.
1 Like

Anything below compression level ~80 to me is unacceptable. I use 92 for normal photos with chroma subsampling, and 98 without chroma subsampling if there’s any pixel-wide text or lines. You invest money in your gear, take the time to go somewhere and take a photo at the right moment, then spend more time processing it… don’t screw it up at the last step - export.

I would recommend you find a different way of sharing it, and/or drastically decrease the resolution.

1 Like

Assuming Mo. means Megabytes this should be possible if you down scale (and possibly denoise) them.

To understand the Problem:
10 mb / 25 images = 400kb / image
A 24 mpix image is ~72mb (uncompressed), so you are asking for almost 200x compression, that’s quite a lot.
Something will have to give. I’d suggest you give up resolution by downsampling (just like @Morgan_Hardwood mentioned) to something like 1024 pixels wide (or heigh). Then you should be able to use a JPEG quality of ~85-90 assuming you strip the exif data and color profiles.
The you’ll be able to fit 25 small but nice photos in your zip. Beats 25 large but ugly ones in my opinion.

Jonas

1 Like

Thank you so much !
It seems like it’s something I should have known for a while,
but thanks, I definitely agree with you @Morgan_Hardwood… It do want to honor the gear and time it took for these pictures .

:smiley:

Hey hey @Little_Navy,

Here’s my two cents. And I hope the RawTherapee buffs won’t jump me for my potential lack-of-quality heresy that I am end up spouting in the next few lines. :stuck_out_tongue:

Here’s my workflow:
Step 1: I edit in RT and export the raws as Highest Quality sampling JPG at full resolution compression quality 90%. Now, let me explain why: I find this setting provides great quality even viewed at 100% and I find the loss between 90% quality and 100% quality is negligible. _However, if I were to send the images out for a large print: I would likely save 100%. Anyway.
Step 2: I do whatever else needs to be done to them, if anything, in GIMP and save at highest quality.
Step 3: In order to export for web I resize them using RawTherapee, with bounding box (I think it’s called) option selected, both values 1200, this way whichever the larger side is becomes 1200, usually yielding something like 1200x790px, which is alightly larger than my blog size. I also enable the post-process sharpening option, which I consider does a great job! Noise removal options are plenty and are definitely useful for noisy images to reduce chances of RT thinking the noise needs to be made obvious. I then export them also as Best Sampling and 90% compression quality, yielding 300-400kb files.

I’ll post some samples for you :slight_smile:

From my blog - uploaded there directly:

Sent to a friend via Facebook, so these may have some Facebook witchcraft added to them:

![](upload://c9HV61AXr6HH6KyCMJ1rIzdTkYD.jpg)

Hope it helps! :slight_smile: