Many users, including myself in communication with customer support via email, have noted that JPEG to Raw AI produces vertical stripes in images. This in effect makes the software unusable.
The suggestion offered to try disabling GPU mode doesn’t work, as others have noted.
Even currently the product page for JPEG to Raw AI states,
We’re continually training our JPEG to RAW conversion model with more data, which makes it better the more time passes. You’ll get the latest and greatest when the model improves.
Yet, there has been no update in well over a year. I believe I paid for this product as part of a bundle of four programs (the Utility bundle), but seeing as it has never been fixed, I suspect that it’s abandoned. Topaz – is a partial refund available if that is the case? If it’s not an abandoned project, will those who purchased it be allowed to upgrade for free even if the upgrade license has expired?
There is no upgrade license for JPEG to RAW and there has been no announcement regarding abandoning the product. Also you are not going to get an answer here as it is a user to user forum.
Note that not all users get vertical stripes in their images after being processed in JPEG to RAW. The details under the Help-> Graphics info. can help in finding a cause as the logs can. It is recommended that you press the Copy button on that dialog and post the results when raising a thread like this.
You also have the option to save as TIF not DNG on the output panel. Also, in the meantime, you can use Sharpen, DeNoise and GigaPixel to produce DNG files if you are finding JPEG to RAW produces compression or vertical artifacts. Artifacting is also a condition caused by the quality of the input image.
Ah, I was thinking of the affinity photo forums where the developers and tech support interact on the forums.
I’m aware I can feed jpegs to denoise/sharpen and produce a tiff or dng as output, but from what I understand JPEG to RAW AI has features geared towrds highlight & shadow recovery, and purports to increase the color from the sRGB gamut to ProPhoto as well as expand to 16 bit. The former feature I imagine is just re-encoding via an ICC profile conversion, but I’m not sure if the 16 bit is just re-encoding 8-bit data by padding or if values are interpolated. So maybe there’s no benefit here, but at the very least the features targetting highlight recovery, shadow detail recovery, and smoothing compression artifacts are of interest to me.