Raw vs. JPeg

[Aswald]

Has anyone of you been able to edit and convert a raw file to look exactly like the camera's jpeg?

 

With Fujifilm's Raw File Converter EX by Silkypix?

[aceflibble]

With Silkypix, no; with Lightroom, yes for most film simulations, almost for the others.

 

Lightroom's film simulations match Fuji's film simulations except for the purest of green tones, and Lightroom's H/S/L sliders can compensate for that perfectly. (I have now set Lightroom to automatically shift the green slider to +5 for both luminance and saturation for Fuji raw files and that seems to make it match 99% of the time.) The exception is Classic Chrome, which Lightroom has nailed so closely already I've yet to see any difference with real-world photos and only a minor difference in technical testing. The sharpening and noise reduction controls of Lightroom aren't as robust as I'd like, but they are good enough* that they can replicate the look of Fuji's sharpening and noise reduction if you really want that, or they can simply be better than Fuji's options if you just knock noise reduction right down to 0 for everything below ISO 800.

 

Silkypix is trickier to get close to the Fuji in-camera simulations. It doesn't mishandle green as much as Lightroom does but all colours are evenly 'off' and contrast is a little different to Fuji's. I tend to notice softer contrast at the extreme shadows and highlights with Silkypix and harder contrast throughout the midtones.

I expect making Silkypix duplicate the look of Fuji's files exactly would be possible, but considering how Lightroom is only one colour off and Silkypix has every colour off and some differences in contrast, I've yet to bother trying to get that kind of result out of Silkypix. It'll be more effort and time for a result another program already gives me automatically.

 

If I shot landscapes I'd probably try to get more out of Silkypix, but as someone who is mostly shooting indoors, Lightroom makes more sense for me. When it comes to matching the Fuji in-camera files exactly, Lightroom blows Silkypix out of the water. Hit up Lightroom and you'll get the results you want, unless you're doing a lot of traditional landscapes.

*I say "good enough", but I think it may be more apt to say that Lightroom's noise reduction is bad enough to replicate Fuji's noise reduction.

[Aswald]

Thanks for the detailed run down of Lightroom. I've heard a lot about how good Lightroom is but I have not had the chance to try it out yet. I can considering Lightroom if it helps with getting better results out of Fujifilm's raw files.

 

So far, I'm not getting very good results editing raw files with raw file converter. Pictures come out muted, dull and noisy.

[flysurfer]

RFC EX 2 has decent Fuji film simulations. I wrote a column comparing them. Since the DR function is automatically emulated, the look is mostly okay.

 

Lightroom doesn't emulate the DR function, so results may be way off, and there are differences with shadow contrast and orange/yellow tones that turn red. Also, the blue sky can be off.

 

The best Fuji simulations I know are the current v3.3 simulations for Iridient Developer. Plus, Iridient perfectly emulates the DR function, so the results will look quite similar. Colors may be a bit different, but that simply means that they may even be nicer. 

 

The Fuji sRGB simulations for Capture One Pro 8 are usable, too. 

[Aswald]

fly, can you post a link to your column please? And thanks for chipping in. Really helpful.

[Aswald]

Weird, why would RFC 2 keep a copy of the older RFC after installation. I've been using RFC 1 without knowing.....    

[aceflibble]

 

 

Lightroom doesn't emulate the DR function, so results may be way off, and there are differences with shadow contrast and orange/yellow tones that turn red. Also, the blue sky can be off.

 

So, I've done some testing of the dynamic range before, and in terms of visual results—I don't know the exact way around Fuji is doing it, but the final image looks the same—simply setting the shadow slider to +25 gets you in the ballpark of 200% and +50 for 400%. For some images you don't get so lucky and a more precise adjustment using the tone curve is required, but nine times out of then, the shadow slider has you covered. After all, all Fuji are doing is overstating their ISO so the image is inherently underexposed, and then they have the camera push the low tones up by a certain amount depending on the dynamic range setting.

Given the dynamic range option produces a less and less authentic result as you increase the ISO, you can be safe in the knowledge that leaving it at 100% and pushing the shadows later will give you a sharper and cleaner image with more accurate colours, if that's your thing, or if you do want to just replicate the in-camera .jpg files then hey, just use the shadow slider, it's not absolutely perfect but it's close enough.

 

Shadow contrast is the same thing. For colour, blue used to be a problem but that does seem fixed now, as my previous colour testing showed; blue is barely different in 'lab' testing and indistinguishable in real world examples. Yellow and orange tones do show some differences in lab-esque tests but it's not that they shift towards red (in fact the Classic Chrome simulation actually shifts yellow very slightly more toward green), it's actually just that they come out a touch lighter. In real world examples, however, no such difference can be observed. My testing suggests that it is only primary yellow which ever shifts.

Red itself can shift a little more on a couple of film simulations, being quite obvious in controlled tests, but it's still barely noticable in actual photographs and again, very easy to fix with a small slider adjustment which can then be set as the default for all of your Fuji raw files; change it once and then you never have to worry about it ever again.

 

Point is, replicating the look of the in-camera .jpgs is not only possible with Lightroom, but easier and quicker with Lightrom than any other program. The few differences there are are very, very simply 'fixed' with a couple of sliders, and once you've done that once you can simply set Lightroom to always apply those adjustments to every Fuji raw file on import. From that point on, all of your Fuji raw files will match the in-camera .jpg results without a single additional click; they're literally matched the moment they're imported from the memory card. I've given Iridient and Silkypix a go and they bother require a little more effort to get looking right and they're not as efficient in terms of the necessary adjustments being made immediately. I have heard Photo Ninja can do a better job, but for whatever reason I can't get that bastard program to run on this Mac properly so I've not been able to test that as fully as I'd have liked. 

 

And unlike Rico, I don't have any kind of personal connection to any particular software or books to flog, so I don't have an agenda to push particular programs, I just advise what works most effectively. 

[flysurfer]

 

And unlike Rico, I don't have any kind of personal connection to any particular software or books to flog, so I don't have an agenda to push particular programs, I just advise what works most effectively. 

 

Well, since this is already a lie, I certainly wouldn't trust your judgement.

[flysurfer]

So, I've done some testing of the dynamic range before, and in terms of visual results—I don't know the exact way around Fuji is doing it, but the final image looks the same—simply setting the shadow slider to +25 gets you in the ballpark of 200% and +50 for 400%.

 

The shadow slider has nothing to do with the DR function. DR function is about adding highlight DR, not shadow DR.

 

The difference between DR100% and DR400% is like this in the JPEG:

 

 

It's very hard (basically impossible) to get the right-hand image (colors, contrast, brightness levels) with LR from the RAW.

It's much easier with Iridient Developer.

[graflex]

Thanks for the detailed run down of Lightroom. I've heard a lot about how good Lightroom is but I have not had the chance to try it out yet. I can considering Lightroom if it helps with getting better results out of Fujifilm's raw files.

 

So far, I'm not getting very good results editing raw files with raw file converter. Pictures come out muted, dull and noisy.

 

The goal shouldn't be to duplicate the camera generated JPEG. If that were the case just use the camera JPEG. SilkyPix isn't the most full-featured raw converter available but it does a decent job and you should be able to process a raw file with SilkyPix to an end result that's consistently better than the camera JPEG. If you'd like some assistance you could post some examples, or I'd be happy to post some examples.

[Aswald]

........ and you should be able to process a raw file with SilkyPix to an end result that's consistently better than the camera JPEG.

 

Thanks for the offer.   That's what I'm trying to achieve. In scenarios where the jpegs aren't what I'm looking for, I should be able to edit raw to maybe a slightly better end result. I'm finally using the RFC EX2 and I can say that it's far better than the version 1. I'll post examples of the two in a bit.

[graflex]

Thanks for the offer.   That's what I'm trying to achieve. In scenarios where the jpegs aren't what I'm looking for, I should be able to edit raw to maybe a slightly better end result. I'm finally using the RFC EX2 and I can say that it's far better than the version 1. I'll post examples of the two in a bit.

 

I'll check back tomorrow. I've used SilkyPix for a number of years and I'm pretty familiar with what it can do.

[aceflibble]

Well, since this is already a lie, I certainly wouldn't trust your judgement.

Please explain to me the personal or business connections I have with any piece of software. I'd love to know how you know more about my career than I do. You've already stated you know and did some form of work with the development of the Iridient processing, and you've a book to flog, so we know the bias you have. My connections in the photography world are strictly in hardware and marketing, so I'd love to know why and where you think I have connections and reason to push any software agenda.

 

The shadow slider has nothing to do with the DR function. DR function is about adding highlight DR, not shadow DR.

 

That's the result, but not how it works as far as the actual process is concerned, though. DR underexposes the image using Fuji's misstated ISO numbers and then pushes the shadows up to match what the ISO should be. I.e., DR at 400% gives you the highlights of ISO 200 with the shadow exposure of ISO 800. In many cameras this would be somewhat hard to duplicate with software, but with Fuji it is easy since Fuji's sensor is essentially an ISO-less sensor and all exposures are pushed to simulate the ISO even in-camera, let alone in software; ISO 200 is being pushed up from the actual ISO of 140, ISO 400 is simply doubling that push again, etc. It's why the noise holds so consistently across ISO values.

What this means is that when it comes to replicating the dynamic range function, you just repeat the steps the Fuji system sets up: underexpose by 1-2 stops and push the shadows accordingly, bearing in mind that even at ISO 200, Fuji is pushing the shadows a little to fake a wider natural dynamic range.

 

 

For topic relevancy, I was walking past this old boat being used as a planter and took a quick picture with RAW+L. Classic Chrome, NR -2, sharpness -1, DR at 200% and colour, highlight and shadow all at 0. White balance set to 5600K with no tint bias. Lens is the 35mm f/1.4 at f/2.8 and I use a Tiffen Black Mist 1/4 filter always, so that's on there adding a little diffusion to the look.

Processing the raw image just now in Lightroom, all I did was simply switch the film simulation from Pro Neg. S (my LR default) to Classic Chrome. NR is at 0, sharpening at 15/0.7/25/20. WB, exposure, contrast, saturation, etc, all left at defaults/0. Highlight, white and black all left at 0, shadows at +25 to match the DR setting of the camera's file.

 

If you can tell a difference between the camera file and the raw file, your must have Superman's x-ray vision.

Dead centre

 

Corner, which vaguely shows a very, very slight difference in the shadow contrast if you really squint at it hard

 

And here is the full image, ONE and TWO. You can try and work out which is the raw and which is the in-camera .jpg yourself.

 

 

Point is, you're not going to get a better match any quicker with any other software. If you want to spend more time on it then you could match the top corner shadow exposure slightly better (though I won't say in which direction; that woudl give away which file is which) but that's a fringe case and in 99% of situations, hey look, all I had to do is select the matching film simulation and my LR defaults had everything else matched already. Import, one click, export, job done. You can't match the in-camera look any faster or easier than that, which is the point of this thread.

[Sapphie]

 

It's very hard (basically impossible) to get the right-hand image (colors, contrast, brightness levels) with LR from the RAW.

It's much easier with Iridient Developer.

 

Hi Rico

 

Would be able to share your technique in ID for recovering highlight detail? It's one area of ID that I struggle with. For example I have a shot taken through a mostly woody path with some sky in the background. With LR I can reduce highlights by -100% and clearly see some details in the clouds and even some blue sky. I can't see to get that with ID - it remains as an overexposed section. Does anyone know of any tutorials that would show worked examples?

 

Lee

[flysurfer]

Iridient offers the Extreme Highlight Recovery slider.

As mentioned before, it works very much like the DR function of the camera, using a similar tone-mapping scheme.

Of course, you need a RAW with DR200%/DR400% to retrieve the highlights.

 

LR is pretty good at recovering highlights that aren't in the RAW (but Photo Ninja is even better most of the time). The thing with LR is its use of adaptive tone-mapping, which may deliver nice results (especially for landscapes), but simply doesn't match the mechanism used bei Fuji's built-in RAW converter.

[owlcavestudios]

Of course, you need a RAW with DR200%/DR400% to retrieve the highlights.

 

So I was under the impression that the DR settings had no effect on the raw file (aside from the bump in ISO). Speaking for raw processing only, couldn't I achieve similar highlight recovery by under exposing 1 or 2 stops (for DR200 and DR400 respectively) with DR set to 100%, or are you saying that Iridient (or other raw processor) is looking for some DR tag in the metadata to do this?

 

The reason I ask, is I'm shooting raw with my X100T, and instead of turning DR on, I keep it off and rely on the Natural Live View histogram to not blow highlights or block shadows. I think I'm essentially doing the same thing, but with more resolution (not just whole stop steps). I'm sure Fuji's behind the scenes process for DR is more elaborate, but I have gotten pretty impressive (only to me probably ) results with regards to highlight recovery and dynamic range with just monkeying with the exposure, highlights, and shadows sliders in LR. I wouldn't say it is "basically impossible" to achieve similar results to your example above in LR. I'm far from a LR expert, and I don't have a great deal of experience at all with Iridient Developer though FWIW.

 

As a mainly raw shooter, I just like to know which in-camera settings directly affect the raw capture, so I can understand what my camera is doing. So if the DR setting is only affecting the ISO, then I will keep it off and manage that under exposure myself. If there's something more going on, it would be great to know!

[graflex]

So I was under the impression that the DR settings had no effect on the raw file (aside from the bump in ISO).

 

The DR setting effects the raw file by NOT applying an analog gain (ISO) that would otherwise be applied prior to A/D conversion. To use DR 200% or DR 400% you must raise the ISO to 400 or 800 respectively. Normally, with either ISO increase, the sensor signal receives an anlog amplification which is applied when the A/D converter creates the raw file. The DR 200% and 400% settings cause that analog gain to be skipped in the raw file. If you're using the camera metering system as Fuji expects, the result is an underexposed raw file.

 

Use of the DR 200% or 400% functions is of course noted in the raw file metadata and can be read and acted upon by raw conversion software -- some do and some don't.

[flysurfer]

Speaking for raw processing only, couldn't I achieve similar highlight recovery by under exposing 1 or 2 stops (for DR200 and DR400 respectively) with DR set to 100%?

 

Since the RAW data would be identical, you could do the same processing in a RAW converter. Of course, you'd have to lower or raise exposure after import to normalize the brightness between the two different (but actually identical) RAWs, but apart from that, the processing would be identical. We are doing this all the time to demonstrate how the DR function works.

[owlcavestudios]

Thanks graflex and flysurfer!

 

I think this all makes sense to me now.