RAW Converter Comparison / Challenge

Setting the Scene

With the New Years celebrations I shot some nice photos of the fireworks and wanted to share some of them.

One of them needed a bit of straightening and perhaps cropping before sharing -- so I opened the RAW file in RAW Power (which uses Apples RAW converter that used to power Aperture).

I always shoot RAW + JPEG so I knew what the file should look like out-of-camera ... and I was unpleasantly surprised with the initial results of RAW conversion.

So I set out to compare different RAW converters in earnest. Most of them I had already toyed around with earlier, but not as serious as now. Now I had a file that could really show some weaknesses...

TL;DR: Affinity, C1, darktable do the best job of the tools I've tried. Exposure X6 loses some detail but does quite a good job still.

Topaz Denoise also does a good job; Topaz Studio 2 (without Denoise or other plugins) does OK but seems unsharp.

RAW Power can be made to look passable with a fair bit of tweaking.

RawTherapee and ACDSee produced output I consider unusable (mind you, with this specific photo and the ones I took immediately before and after. They are OK on other photos I've tried before).

Did not try: Lightroom, Iridient X-Transformer, Luminar, SilkyPix.

I will not purchase either of them for various reasons that have nothing to do with their quality and are probably best discussed in other threads, but I would be very much interested in seeing the results of how these tools handle the particular files I used.

SilkyPix is too darn slow for me.

If you know of any other RAW conversion tools for Fuji files, I'd be very interested to see how they do.

Links to my files, for anyone who wants to try out for themselves what results they can get with any of their favourite tools (I'm particularly interested in seeing the output of Lightroom and Iridient X-Transformer).

Google Drive: RAW / OoC JPEG

Dropbox: RAW / OoC JPEG

(I shared via 2 ways in case one may become unavailable due to unforeseen circumstances)

The Out-of-Camera JPEG!

Straight out of the camera, I can see that the image has a lot of fine detail.

These are the fireworks, straight out of camera JPEG

Screenshots from various RAW conversion tools

To best get an idea of what various RAW conversion tools did with the file, I decided to attach a number of screenshots here, taken from these tools, on 2 different computers (one a Mac, the other Windows).

RAWPower

Here is a screenshot taken from RAWPower (using Apples Aperture engine). Issues:

• Loss of detail
• Colour Fringing

I could to some degree resolve these issues by reducing exposure, applying a lot of sharpening.

RAWPower loses detail

Exposure X6

I quite like the results of Exposure X6 but I feel that it loses some fine detail. Good, but not the best. (Overal quite a good tool for working on RAW files btw).

Exposure X6 does well, I think

Topaz Denoise

Topaz Denoise is not a RAW conversion tool, but it does open RAW files to process them. So I gave it a try.

• Good retention of detail.
• Changes the colours -- it does that with all my RAW files it seems, even from Canon.
• While it generates a preview you can see a "rough" rendering  that has a lot of colour fringing, which it resolves very nicely in it's final output.

If it didn't have the colour-cast issues, you could use this as a first step in your workflow: demosaicing, noise-reduction, save as TIFF and move to next tool.

Topaz Denoise output

Moving now to the tools I used on my Windows machine.

ACDSee

Biggest issue: Colour fringing. Also a loss of detail.

First it looked like this:

ACDSee / wish to Unsee

With  some tweaking I got this:

ACDSee / passable

Neither versions is as good to me as the JPEG out of camera.

Next up:

Topaz Studio

Topaz Studio is Topaz's attempt at a RAW conversion tool / image editor. No asset management whatsoever and very limited in its workflow. I don't think I could even save my changes without exporting.

The other Topaz tools can plug in to Topaz Studio.

Anyway, the image seemed very fuzzy when opened in this tool. Which surprised me.

I did not try to add any filters or corrections on the image now.

Topaz Studio

Affinity

Affinity did quite well in my opinion, retaining detail and good colour, very little to no fringing. It's not the most capable RAW editor, I think their idea is that you first "develop" the RAW files and then use the more extensive tools of the photo editor to complete your work. So it's more of an import-tool to their Photoshop clone.

But it does do a pretty good job of RAW conversion (especially considering what I've seen come out of some other tools!).

Affinity

Affinity

darktable

At first I didn't try Darktable, since I thought it would give the same results as RawTherapee (implementing the same algorithms). This morning however I decided I should still try it and not dismiss it just for using the same open-source algorithms as RawTherapee, glad I did as it produces much better output.

After applying lens corrections, the output is up there with the best, I think.

It's too bad that personally I've always struggled with usability of darktable and therefore I'm not likely to actually be using it.

darktable

CaptureOne 21

CaptureOne produced good results. No complaints. In fact, I found that I got on rather well with C1 21 and might just buy a license today before they raise their prices!

The output looks quite different from the screenshot  of darktable right above, a lot of the smoke visible seems missing from darktable output. I think it's a matter of how they choose their exposure  by default.

CaptureOne 21

RawTherapee

I saved this one for the last. It's the first tool I tried after RAWPower disappointed me, and RawTherapee disappointed me even more. What should have been there amongst the best (according to various threads I've read in the past), was the worst: 3-pass Markesteijn demosaicing, in RawTherapee.

I tried all the different demosaicing engines, and found 1-pass (fast) to produce least unacceptable results.

Using 3-pass Markesteijn demosaicing. What the actual freak happened there?

1-pass fast. Looks better, still a lot of colour fringing.

So. There you have it.

My full comparison of RAW conversion tools for Fuji X-Trans cameras.

And-roid wrote:

Tim van der Leeuw wrote:

In my testing the best xtrans converter, technically, is Lr + Ed tool Capture 1 runs it a close second and might be preferable in some circumstances, also the x-transformer an interesting option for Lr too but not for me. Once you've converted you can do what you want in the plug-ins, Topaz/Exposure/Luminar etc but generally speaking they are compromised for raw conversion and slow too.

The licensing of Adobe is seriously an issue for me (personal thing that everyone should resolve for themselves of course), but I'd be quite interested in seeing what you  could produce with the same RAW file!

You’ve been very busy Tim! Thanks for taking the time to do this, I’ll study your results when I get a minute. 

Capture one 20  : Velvia base

Interestingly enough, I have just finished doing some testing myself. A couple of my observations are fairly similar to yours, even though I chose a different method.

I didn't want to jack your thread, so I didn't post my images. I could, if you want.

I used Capture 1, ACDSee, X RAW Studio, and Affinity.

What I did differently, was I intentionally underexposed my test raw by a full stop so that I had a controlled example of being able to push my conversions one stop each. Other than that, I just added sharpening. No white balance adjustments; no highlight/shadow adjustments, etc.

To me, this gave me a better base for just RAW conversions, as opposed to conversion+ adjustments. I figured that any adjustments beyond what I did would be up to the users taste.

My results?

Capture 1 and X RAW Studio were almost identical. ACDSee seemed a little 'flat,' compared to the first two, and Affinity was a little more saturated.

As far as I'm concerned, with a little effort, all could be made to appear almost identical with a little tweaking -but as I said, this would be editing, and not just converting.

Honestly, unless fireworks are your primary focus, this is really a poor choice for evaluating RAW converters. All three color channels are blown to some degree and there isn't really any fine detail or nuanced color to work with. This really only shows how well a converter might handle irretrievably blown highlights and fringing. Some of your examples that exhibit less attractive color here may actually reveal more nuanced color detail with a "normal' image.

This is screenshot from RawDigger that shows irretrievable RAW clipping (in red).

RawDigger screenshot

Here's one from Lightroom (X-Transformer is no help with this)

Lightroom

William Loney wrote:

Interestingly enough, I have just finished doing some testing myself. A couple of my observations are fairly similar to yours, even though I chose a different method.

I didn't want to jack your thread, so I didn't post my images. I could, if you want.

I used Capture 1, ACDSee, X RAW Studio, and Affinity.

What I did differently, was I intentionally underexposed my test raw by a full stop so that I had a controlled example of being able to push my conversions one stop each. Other than that, I just added sharpening. No white balance adjustments; no highlight/shadow adjustments, etc.

To me, this gave me a better base for just RAW conversions, as opposed to conversion+ adjustments. I figured that any adjustments beyond what I did would be up to the users taste.

Comparing without making highlight/shadow adjustments etc. is really only comparing whoever came up with the different RAW processors' default import settings tastes. RAW files aren't images, and every RAW converter will apply it's own unique default processing with color profiles and tone curves which may produce a significantly different base conversion compared to other converters, but which can be easily changed to be similar to other converters (and more to your liking).

To get a good idea of which RAW editor will serve you best, you really ought to (IMO) compare them after you've tweaked each to get the best output.

My results?

Capture 1 and X RAW Studio were almost identical. ACDSee seemed a little 'flat,' compared to the first two, and Affinity was a little more saturated.

As far as I'm concerned, with a little effort, all could be made to appear almost identical with a little tweaking -but as I said, this would be editing, and not just converting.

William Loney wrote:

Interestingly enough, I have just finished doing some testing myself. A couple of my observations are fairly similar to yours, even though I chose a different method.

I didn't want to jack your thread, so I didn't post my images. I could, if you want.

I don't mind if you add your images to this thread as well!

I used Capture 1, ACDSee, X RAW Studio, and Affinity.

What I did differently, was I intentionally underexposed my test raw by a full stop so that I had a controlled example of being able to push my conversions one stop each. Other than that, I just added sharpening. No white balance adjustments; no highlight/shadow adjustments, etc.

To me, this gave me a better base for just RAW conversions, as opposed to conversion+ adjustments. I figured that any adjustments beyond what I did would be up to the users taste.

I didn't shoot these images with the purpose of testing RAW converters -- but it proved a very good test-bed for them. A lot of detail and sharp lines. Some tools really messed that up, more than I had expected.

My results?

Capture 1 and X RAW Studio were almost identical. ACDSee seemed a little 'flat,' compared to the first two, and Affinity was a little more saturated.

As far as I'm concerned, with a little effort, all could be made to appear almost identical with a little tweaking -but as I said, this would be editing, and not just converting.

Interesting observations!

I could not get acceptable results out of ACDSee, nor out of RawTherapee (although I had previously tried them with different images and they performed OK on those easier images).

So I wouldn't say that they all can be made to appear almost  identical!

BTW I mostly didn't do any editing either, looking mostly at just RAW file as it was opened, but when the converter clearly screwed up I tried to see if I could fix the file by enabling or disabling lens corrections etc.

Erik Baumgartner wrote:

Honestly, unless fireworks are your primary focus, this is really a poor choice for evaluating RAW converters. All three color channels are blown to some degree and there isn't really any fine detail or nuanced color to work with. This really only shows how well a converter might handle irretrievably blown highlights and fringing. Some of your examples that exhibit less attractive color here may actually reveal more nuanced color detail with a "normal' image.

Thanks for your insights!

It does indeed heavily judge them on  fringing and highlights, that's true. And fireworks is not going to be my main focus of course.

But blown highlights will happen with some subjects.

To me it suddenly and unexpectedly highlighted how big differences can be between different converters, sometimes you need extreme situations to bring them out.

Erik Baumgartner wrote:

William Loney wrote:

Interestingly enough, I have just finished doing some testing myself. A couple of my observations are fairly similar to yours, even though I chose a different method.

I didn't want to jack your thread, so I didn't post my images. I could, if you want.

I used Capture 1, ACDSee, X RAW Studio, and Affinity.

What I did differently, was I intentionally underexposed my test raw by a full stop so that I had a controlled example of being able to push my conversions one stop each. Other than that, I just added sharpening. No white balance adjustments; no highlight/shadow adjustments, etc.

To me, this gave me a better base for just RAW conversions, as opposed to conversion+ adjustments. I figured that any adjustments beyond what I did would be up to the users taste.

Comparing without making highlight/shadow adjustments etc. is really only comparing whoever came up with the different RAW processors' default import settings tastes.

That was the whole point of my test: Does the converter give one a reasonable basis with which to apply adjustments as they see fit.

RAW files aren't images, and every RAW converter will apply it's own unique default processing with color profiles and tone curves which may produce a significantly different base conversion compared to other converters, but which can be easily changed to be similar to other converters (and more to your liking).

To get a good idea of which RAW editor will serve you best, you really ought to (IMO) compare them after you've tweaked each to get the best output.

Yes, I understand, but using a baseline test for a simple CONVERSION only will give the user a fair idea of the potential AFTER adjusting their images. Furthermore, some converters have fairly limited adjustments. Take X RAW Studio for instance. IMHO, it's among the best for Fuji raw files. It has no curves adjustments like the other three that I tested.

My results?

Capture 1 and X RAW Studio were almost identical. ACDSee seemed a little 'flat,' compared to the first two, and Affinity was a little more saturated.

As far as I'm concerned, with a little effort, all could be made to appear almost identical with a little tweaking -but as I said, this would be editing, and not just converting.

I used a two flash setup, intentionally underexposing by a stop so that I could push the images one stop with their respective converters. I chose a high contrast side lighting to show the differences from light to dark. X-T3, 90 @ f5.6, 1/250. Pro Neg Hi.

This is the original image converted with X RAW Studio. No adjustments:

Here's X RAW Studio pushed one stop, and minimally sharpened:

Capture 1.

Affinity:

ACDSee:

All images can be downloaded/viewed full size.

Tim van der Leeuw wrote:

Erik Baumgartner wrote:

Honestly, unless fireworks are your primary focus, this is really a poor choice for evaluating RAW converters. All three color channels are blown to some degree and there isn't really any fine detail or nuanced color to work with. This really only shows how well a converter might handle irretrievably blown highlights and fringing. Some of your examples that exhibit less attractive color here may actually reveal more nuanced color detail with a "normal' image.

Thanks for your insights!

It does indeed heavily judge them on fringing and highlights, that's true. And fireworks is not going to be my main focus of course.

But blown highlights will happen with some subjects.

Blown highlights won't happen if you expose correctly, not important highlights anyway.

To me it suddenly and unexpectedly highlighted how big differences can be between different converters, sometimes you need extreme situations to bring them out.

True, and none of the converters are always the best at everything, they all come up short in one respect or another. I don't have a lot of experience with with some of the lesser known players, but every time I try one, each will usually have some significant failing that I can't live with. To date, Capture One and Lightroom are the only two options that tick enough of the necessary boxes for me.

With some adjustments under the exposure tab, mainly Tone Equalizer.

used also Dynamic range compression which darkens the (blown out) highlights but retains a little better the relative intensities.

It's always interesting to see these types of comparisons. ACR was configured with sharpening and noise reduction off. As my site resizes the image, I'm including the OOC as well as my conversions

-1 OOC

Out Of Camera

-2 Photoshop ACR

Photoshop ACR

-3 Photoshop ACR + Topaz Labs Denoise

Photoshop ACR + Topaz Labs Denoise

I feel I could get more out of this file by adjusting the exposure controls in ACR yet this is about the raw converters and not processing skills.

Morris

I don't know this tool, ART, but it seems to me when looking at 100% that I see some spots that I didn't see in the results from other files, like dark halos around some of the sparks of  the fireworks.

Do you see what I mean?

Morris0 wrote:

It's always interesting to see these types of comparisons. ACR was configured with sharpening and noise reduction off. As my site resizes the image, I'm including the OOC as well as my conversions

-1 OOC

Out Of Camera

-2 Photoshop ACR

Photoshop ACR

-3 Photoshop ACR + Topaz Labs Denoise

Photoshop ACR + Topaz Labs Denoise

I feel I could get more out of this file by adjusting the exposure controls in ACR yet this is about the raw converters and not processing skills.

Morris

Thanks for the other conversions Morris. Indeed, I knowyou could get more out of these images, if that was your intention -I've seen your birding pictures!