I believe DxO Photolab2 does make some image adjustments by default that do alter the appearance of the photo. Sharpen AI can’t adjust colors or basic sliders, so I would think it would leave the RAW image looking flat and bland as they are. I’m guessing if you don’t have the problem with a TIF, it’s because adjustments were made to the image before outputting to the TIF.
@ScottO PhotoLab has a “No Adjustments” setting, which is what I used when exporting the first jpg. It should have just converted the raw to a jpg without changing the basic values. I will look into this a bit more.
OK. That makes sense. Fortunately opening raw files in Sharpen AI is not my default action. I normally adjust them in PhotoLab or CaptureOne and export tiffs to Topaz stuff. I guess I will continue to do that.
But what exactly does it mean that RAW can be opened but is not supported?
Does it mean the software can’t/won’t do any default adjustments? Or that they can’t save any adjustments back out to the same RAW format? I can open NEF, make an adjustment, and save the result to various non-RAW formats.
In my experience, most software that opens RAW images make some default adjustments automatically. The primary reasons are to make their software look good and to help give the user a better starting point. Completely unadjusted RAW images are quite bland, like Mike’s second image above. They’re also fairly soft. I assume the second image isn’t soft because SharpenAI sharpened it. Basically, RAW images look pretty crappy, especially if your camera settings turn off as much in-camera processing as possible. Most users never see that because their software never shows the unadulterated RAW image.
What I was getting at is that the look of Mike’s second image doesn’t seem like a bug to me.
Topaz Products do no adjustments to RAW images and simply use a linear conversion. Note all products start with linear conversions and add a tone curve, some the add pre-sharpening and noise reduction.
The fact that it is a Linear conversion is not the real issue because it is a better starting point, the issue comes about because each and every sensor has different attributes and the best overall camera settings (ISO, shutter speed, f-stop) and image quality attributes (dynamic range, noise, bit-depth, sensor size) need to be selected to make the best image. And, obviously the quality of the lens makes a big difference during the capture process.
But as you can see from this image that, especially shooting in natural light, there needs to be some way to create a more even starting point. This is a ‘harsh’ example because you can see the reflections caused but heat, makeup and sunscreen … they are not clippings:
So the Linear Conversion (Curve) is a better starting point but, where Linear really becomes important is in application to a landscape scene where dynamic range is really important.
But, and it is a big “BUT”, you need the tools to craft the image to your basics as AI is not sufficiently advanced to recognise the overall scene and adjust it for you. The basic tools to adjust things such as Exposure, Contrast, Mid-tones, Saturation, Curves and Localized Adjustments are still needed IMHO.
And you can see that simply applying a curve to brighten the midtones, in this case of a linear conversion, provides a really good starting point, albeit a little flatter than the standard curve that is normally applied to the RAW conversion you see above: