Topaz JPEG to RAW AI

I’m a Topaz fan, but Topaz are over-hyping their latest product, “JPEG to RAW AI”. It’s a deceptive title which has a lot to do with marketing and very little to do with reality. The program does try to make the best of a file in the JPG format, but does it create a RAW file? Emphatically not. Not even close.

This is a shame because, after forcing myself to ignore the hype and stay with the testing, I did find a program which is useful, even though it doesn’t do what it says on the tin. Here’s a look at what it can do.

Sometimes, all you have is a JPG file and no amount of wishful thinking or regret will get you the RAW file that you wish for. Some of us develop strategies to get the best out of these JPEG files, some just give up and go take another picture. But now, JPEG to RAW AI does bring something extra to the party. Whether it is extra enough to be worth paying nearly US$100 for is another question.

I did tests; lots of tests. Here is one test that is fairly indicative of them all. The original JPG file was well exposed and taken at a low ISO. Nevertheless, the results are not that pretty - this is a screenshot of a portion of the picture at 200% magnification. As you can see, it’s fairly ‘gritty’ with both JPEG artefacts and noise. On an arbitrary quality scale of 0-5, I’m going to score this at the bottom = ‘0’ (the best sample will score a ‘5’). (You may need to view the pictures larger to see the differences.)

In the second screen-shot, I applied just enough noise reduction to remove the ‘grit’. I used Topaz DeNoise 6 on a fairly low setting and then added a bit of sharpening. It’s better, the grit has gone, but now the grass is turning to mush. This is the typical noise reduction tradeoff and, depending on your personal preference, may raise the quality score to 1 or maybe a 2.

The third screenshot is what I usually do with JPEG files, I use Topaz AI Gigapixel to reduce the file size by 50% which removes most of the grit, and then enlarge by 200% which restores some clean detail using Topaz’s AI algorithms. Finally, a touch of DeNoise 6 deals with any small amount of remaining ‘grit’ while also taking the edge off of AI Gigapixel’s aggressive sharpening. The result is much better all around but, at 200% magnification, some ‘staircasing’ is evident and the very finest of detail has been lost - this is almost invisible at normal viewing magnification. For me, this jumps the quality score up to about a 4 and produces a usable file

Finally, I run JPEG to RAW AI and again, because this deals with JPEG artefacts but not actual noise, I have to add a touch of DeNoise 6, just as I did with the other samples. This deals with all the ‘grit’ in the original file, shows no ‘staircasing’ and retains the finest of detail. This is the best of all the four tests and therefore sets the upper bar of the score table at 5.

Topaz also claim an improvement in highlight and shadow recovery. Quite frankly, I could find no evidence of this beyond that which anyone can get from converting a JPEG file to a 16 bit TIFF before attempting the recovery - and you don’t need expensive software to do that.

All this creates a dilemma; JPEG to RAW AI does do a very good job but I already have AI Gigapixel and can get most of the way there without spending another US$100 on a new piece of software (which, by the way, takes several minutes to process a 20Mb JPEG file). If I can get ⅘ of the improvement without spending a cent more, then is JPEG to RAW AI worth its rather high price for that last little step? Probably not.

And, Topaz, please drop this JPEG to RAW nonsense - it’s just embarrassing and makes you look silly.

Topaz AI Gigapixel - tested

When Topaz came out with their AI Gigapixel (AIG), image resizing product, I was sceptical. Firstly, I had tried many resizing products in the past and had settled on one that worked quite well (within the accepted constraints of resizing). Could AIG, do better than what I was already using? I doubted it.

My first attempt was not good. AIG crashed on first use and clearly didn't want to work with my graphics card. I walked away disappointed but not altogether surprised. I told myself that, anyway, it probably wouldn't have been any better than what I was already using.

I must be a glutton for punishment because when the email arrived announcing an improved 1.1 version, I downloaded it and tried again. This version included a CPU mode, which bypassed the GPU on my graphics card. That should, at least, allow it to run.

It did run, and the first output looked surprisingly good. So I devised a stricter test - I took a 24MP DSLR image and downsized it, in my normal editing program, to 6MP (half the width and height). I then put the 6MP image through AI Gigapixel at 200% to bring it back to its original size. I could then compare the original 24MP image with the enlarged 24MP AIG image and see how different they were.

The AIG image was sharper than the original. AIG obviously adds some sharpening during the resizing process. To compensate, I added some sharpening to the original file so that they were more directly comparable. What I was really looking for was a loss of detail and fat 'jaggy' edges in the enlargement. At 100% I could not see any difference between the two files. Definitely, not the result I was expecting.

Screenshot - Original 24MP file on left - 6MP file enlarged to 24MP on right (click to see full size) 

The screen-shot above is taken at 200% magnification and contains both in focus and out of focus items. The original image is on the left and the AIG resized image is on the right. I am still struggling to see any significant difference between the two images. Not quite believing what I was seeing, I repeated the same test on other images all with a similar result.

As a final test, I took an image from a different camera - a compact with a 1/2.3", 16Mp sized sensor. I enlarged the image by 200% in my image editor, using the bicubic sampling method and then enlarged the same image in AIG and compared the results. The screenshot below, is at 200% magnification and the results are quite surprising - the AIG image has much finer detail (particularly at the edges), texture seems to have been replicated rather than enlarged, and image grain has been suppressed to a far greater extent, leading to a much cleaner image that would be difficult to pick as an enlargement.

Screenshot - Bicubic enlargement on left, AIG enlargement on right (click to see full size)

All my tests with AIG have been at 200% enlargement because that is what I would normally use. But AIG can enlarge up to 600% (though the results may not be so acceptable). The shrink-enlarge experiment indicates that AIG could be an effective substitute for a camera sensor with four times as many pixels (16MP becomes 64MP) or for a longer focal length lens. AIG produces significantly better results when compared with either a standard bicubic enlargement or previous industry-leading enlargers.

US$99 does seem a lot for a single task piece of software. But when compared with the cost of a camera with a larger sensor or a lens with a longer reach, AI Gigapixel may well be a reasonably priced alternative.