I am super super pleased with my Topaz Studio. You guys rock! I am considering my workflow and looking for technical improvements.
Where would ‘JPG to Raw AI’ fit in?
Do I get better results if I convert the jpg crop and expand the Tiff file?
Right now I Gigapixel the whole file and then crop. What are the technical advantages of the extra plugin?
Bo
Jpeg2Raw is most probably best used on photos from phone cameras. Phone cameras do not do Raw out of the box, and the extra bit-depth in Raw enhance processing no end.
But what advantage is it over Sharpen AI for a Topaz Studio user. I’m trying both and Sharpen AI makes lossless pngs or Tiffs.
I use lightroom and if I need metadata I’m ok with the database and sidecar files.
Is there a performance value, if you already have the other software?
Hi Bo,
It depends on what type of images your source images are.
JPEG to RAW only makes sense if your originals are JPEGs. Because JPEGs are lossy compressed, there is less information in the file, which is why they can be compressed so much. J2R then attempts/allows to recreate some of that lost information and outputs a more robust file for postprocessing.
If you start with Raw or TIFF images, then J2R has nothing to offer you.
Cheers,
Bart
Sharpen is for sharpening - Jpeg2Raw is a format changer - little to do with sharpening. As I say, Jpeg2Raw gives extra bit-depth and makes things like sharpening and denoising more comprehensive.
Using a Pub Domain image Jurij
Here is a head cropped directly from Gigapixel Ai and the same head cropped from a converted DNG (using Jpg to Raw) and then Expanded in Gigapixel Ai.
I don’t know what this proves. Its very hard to second guess the AI process to know if either program is going to have advantages. In this case I would say there are no appreciable differences.
That looks like a scan to me and not a JPEG image … here is a 2400 x 1600 image cropped from a EOS 5DsR image and then sent through JPEG to RAW AI and you can see the blur reduction and the improved dynamic range:
Original:
JPEG to RAW opened in Studio:
And you would use Jpeg to Raw in that situation over another AI product, eg Denoise, because?
Dynamic range? Preserving Data?
What is it about the algorithms in Jpeg to Raw AI that makes it worth it on top of the rest of the Ai Series. What does it do better than any other AI modifier?
I love the product line. I’m looking for the rational(s) to add Jpeg to Raw to my collection.
It tries to undo the lossy compression that’s specific for JPEGs. That’s basically all. It’s a specific technical tool. The other AI applications address other aspects of image-processing, such as noise, or sharpening, or scaling.
I guess the purpose is that JPEG to RAW AI is highly effective at removing compression artifacts and recovering color detail. There’s also enhanced dynamic range as shown by deeper shadows and enhanced highlights.
Where this is most effective is for highly compressed images typically produced by lower quality Phone images or indeed images produced by older cameras.
Typically JPEG to RAW AI is exceptional at removing visible compression artifacts in your image and preserving natural image features.
JPEG to RAW AI itself tries to recover dynamic range, i.e. highlight and shadow detail, and then improving the color space and color depth from the original sRGB image by using AI to convert to ProPhotoRGB because an 8-bit JPEG only has 256 values per channel but JPEG to RAW AI will expand it to 65,532 values per channel. Then it will try to recover fine detail removed during the conversion process and try to remove compression artifacts in your image, typically caused by JPEG compression algorithms.
I would try running old JPEG images through to see the difference, those from older cameras or point and shoot cameras or indeed early sports action cameras such as GoPro’s which had very ‘smooth’ images and small dynamic range.
I have tried J2R on a number of Jpg pictures including some taken with an older Konica Minolta 10MP digital camera and some taken with a newer iPhone. I have not seen even one improvement in any, There is a small amount of sharpening but no artifact removal in faces (teeth are a good are to look).
I don’t think J2R is intended to make improvements. Its intended to be a deep and faithful copy rendered in a more precise and flexible format. You are supposed to fix artifacts and make improvements in the better applications designed for RAW files.
From the Topaz website: Use machine learning to convert JPEG to high-quality RAW for better editing. Prevent banding, remove compression artifacts, recover detail, and enhance dynamic range.
Some do, Jack. Mine does…
JPEG to RAW AI is the successor to its predecessor the DeJPEG plugin
JPEG 2 RAW claims to:
“Use machine learning to convert JPEG to high-quality RAW for better editing. Prevent banding, remove compression artifacts, recover detail, and enhance dynamic range.”
For Noise reduction there have been, and still are, better dedicated solutions unless the ‘noise’ was caused by lossy compression artifacts from lower quality JPEGs.
Sorry it is not. DeJPEG is similar to DeNoise insofar as it removes JPEG artifacts and enhances image clarity.
Know what you are saying, but there is a really big difference if you run the same image through both, especially in dynamic range recovery.
I agree that JPEG to RAW AI does a much better job !
I have a licence for DeJPEG, but am not being offered either a free or reduced price upgrade to JPEG to RAW AI. How does one go about sorting that?
Usually if you qualify for a free upgrade to any Topaz product there should be a coupon code noting the freebie on your acct page. If you don’t see anything there when you check then you can open a cust svc tkt and ask there (I don’t think tech spt is the proper place for that).
