Wednesday, February 27, 2013

Posted in Arrangement, Art, Business

More presets to come....

Profiles are built on a 900+ colour calibration chart and compared against the in-camera JPG profile simulations.

This is a little update to my last post of my Velvia 100 RVP Film Simulation with a link to the preset for your own use. It has been really well received and I thank all those who have sent me feedback and also donated to the cause. Every little bit is appreciated.

Above you can see a little sample of my work flow in matching profiles to the colours it presents. This specific example is matching the in camera simulations made by my X-Pro1 and adjusting the RAF file to match each of the different simulations.

I started this project back in November but wasn't completely happy with the results. Matching across 900 different colours was the easy part (as in the example above), but there were other factors that were alluding me at how at least in this case, how FujiFilm was creating their profiles. When I applied the adjusted profiles from the colour calibration chart onto a live image, it didn't work.

I made some recent discoveries from my own generic Velvia 100 RVP film simulation development that helped bring me back to working on this profile.  Below is a sample of my results which is very close. Bear in mind that JPG images are only 8 bit, and the RAF files are 12bit>16bit images which of course makes some of the matching a little challenging.

For X-Trans users who have been looking for presets that match the in-camera JPG simulations, I'm hoping to have those simulations out very soon.

The bonus of these presets is that although I'm working on making a specific preset for the X-Trans sensors, these presets can be used for other cameras as well. But my primary objective is first to match the camera output as much as possible then to make it available to the public.

Thanks to all my supporters as I continue to develop this.

You are always welcome to download my presets for free, but I also welcome donations to my efforts:

Select an amount you'd like to donate:

Once a profile has been built, it is compared against live samples. Some adjustments made, and thrown back at the calibration chart. It's a painstaking process of doing this manually back and forth.

Side-note: Although it appears that all my settings are along the side of my examples, it's a little more complicated than that. There is some other calibration that is not seen in these examples that are essential in making a perfect match to the camera output.

Monday, February 25, 2013

Posted in Arrangement, Art, Business

Lightroom 4.4RC with new improved X-Trans Sensor Support

Lightroom 4.4RC vs Lightroom 4.3 (micro details look natural vs watercolour effect)
Today Adobe Labs announced the latest version of Adobe Lightroom, bringing it to version 4.4. This specific version addresses some of the issues that Lightroom 4.3 had with the X-Trans files from the X-Pro1 and the X-E1.

Click here to download Lightroom 4.4RC from Adobe Labs

I tested this against two reference files that really pushes the limits of almost a dozen RAW processors I've thrown at the FujiFilm RAF files. I'm please to inform everyone that the dreaded watercolour, or colour smearing issue has been abated in this version which is very welcome. The muddied colours really caused odd colour casts (you can see a slight green tinge in the LR 4.3 version), and obviously the details were also creating the watercolour effect appearance.

New Lightroom 4.4RC vs Capture One 7.02
Although details are still not as good as some of the other RAW processors that I've thrown at it, it is improved over the previous versions. One of the more notable details is in the shadow details. LR4.3 lost a lot of micro details in the darker or shadier areas. LR4.4RC improves that dramatically. But as you can see, the Capture One details are still surpassing what both default settings are putting out in this case. One could argue that Capture One is over sharpening, which I might agree, but I found that it's much easier to deal with over sharpening in Capture One, while in LR4.4RC, it's difficult to sharpen the images without introducing a moire issue (a maze like pattern in some textures).

The good news is that LR4.4RC responds better with the Radius setting in sharpness. By reducing the size of the pixel to something like 1/2 pixel, it dramatically reduced the moire issue (I wonder if Adobe was listening in on some of my theories on 1/2 median filtering). So like Capture One, default settings might not be the most ideal setting to use. If you're using RAW processing, you're likely not going to use the default settings anyhow.

Lightroom 4.4RC vs in camera JPG

Lastly, one of the best improvements that I can compare and finally say that if you were shooting only in JPG because you felt the in camera results were the best, Lightroom 4.4RC actually looks much better than the JPG rendering. This is a big improvement over the previous version. Details in the logs above is dramatically improved especially in the subtle shades. There are some softening of details which is likely a result of a different approach, but it is welcome considering how much of an improvement it has over the previous version's issues and even in camera JPG.

Personally I am comfortable to say that Lightroom 4.4 has at least finally matured as a good choice for RAW processing of FujiFilm X-Trans files. Is it perfect? No it isn't, but its improvements are dramatic, and it's apparent that Adobe and FujiFilm did indeed trade information to correct for some of the glaring issues that has kept most FujiFilm X-Trans system owners from shooting in RAW.

This is a pretty quick review and I'm still exploring the potential of Lightroom 4.4, but I'm very pleased by the results at this stage.

Two things I'd like to update, first the question of what settings to use in Lightroom for sharpening. I found the best balance seems to be the following:

Amount: 70 (adjust this setting to suit your preference)
Radius: 0.5 (almost like half median filtering that I did in early workflows)
Detail: 30
Masking: 35

This gave me the best balance of sharpening without too much moire artifacts, specifically the maze like effect that happens on ultra high frequency textures (usually only visible at 200%).

The second update is that something an improvement that most people wouldn't know about is something I call Zipper Aliasing (something I mentioned several posts back). 

This isn't something that Lightroom 4.3 had a problem with (because it never resolved the details high enough), but with the increase of details, it's a fine balance between this unique zipper alias that happens only with the X-Trans files.

It is as far as I can tell by looking at just a RAW demosaic file, that it's the Green sensors that cause a unique zipper like pattern to form between certain details (only visible at 200-400%). This was present in processors like Silkypix and DCRAW especially if you over-sharpened. Lightroom's increase in detail did not result in any zipper aliasing when over-sharpening which is good. The trade off of slightly softer interpolation gives better control over that phenomena.

Saturday, February 23, 2013

Posted in Arrangement, Art, Business

FujiFilm Velvia 100 RVP Simulation

Early version of my simulations. Started with simulating the dynamic range.
One of the things I like to do in my spare time is trying to capture a specific film look.

In the past couple of weeks I've been using Capture One 7.02 to do some panachromatic film simulations which is really spectacular for doing that. I found that Lightroom, or at least at this time hasn't quite got there for me (yet), but I have been spending more time in Capture One for doing black and white conversions.

In Lightroom, I've been trying to simulate popular films, with some success.

I had been using 900+ colour reference charts, and comparing them to how others have done their presets and also how companies like FujiFilm has been trying to simulate them in their own cameras (such as the X-Pro1).

Reds and Greens. Signature trademarks of
FujiFilm Velvia RVP 100
Even FujiFilm's simulations are still not quite where I'd expect them.

Regardless, I decided to try to attack one film that few or none have tried to simulate, which is Velvia 100 Reversal Positive film or slide film. I'm a huge fan of the long retired Kodachrome 64 film which just has a character that's hard to copy (and some of the simulations I've seen out there come close but still not quite there).

I find that most people doing simulations are trying to copy something they are remembering from memory, fortunately for me, FujiFilm sent me a case of Velvia RVP 100 for free and I've been shooting hundreds of frames for the past couple of years. Paired with my own 20+ years of imaging, I decided to tackle this task again.

There's two known facts about FujiFilm Velvia. It's horrible for portraits, but awesome for landscapes. This is based on the fact that the formula for Velvia RVP makes it more sensitive to reds and greens, something that FujiFilms has always been famous for. In fact while I shot many landscape images with Kodachrome, I always knew it was much better suited for portrait photography (as made famous by Steve McCurray and his famous 'Afgan Girl'.

Pulling the dynamic range closer to film.
4 or 5 years ago, I had created some Velvia F simulations that were well received by the public when I shared my results, but I abandoned the project because I was frustrated by the software that I was using at the time (this was called Picture Styles Editor and was specific to Canon equipment). Taking some of that experience and applying it here was logical, but Lightroom is a little different (and potentially frustrating as well) in it's approach to being able to simulate films.

Strong pinks, a signature of Velvia RVP 100
I've downloaded hundreds of different simulations to see how each person has approached their simulations, and it occurred to me that most of them were trying to simulate colour and contrast, and not the dMAX range of film.

What makes film special is what it reproduces on paper or what our eyes can see. I've often heard (maybe too much) that digital cannot simulate what film produces. Part of that problem is that film actually doesn't hold back what it can show you (when looking at RVP on a light table, it just looks 3D for some reason). Digital technology has tremendous range, but the problem is in display technology. 8 bit technology just can't show what is found in the digital file.

Digital Slide Copier
This isn't going to be a discussion of high dynamic range, but part of the problem with film simulations that I've seen, does not take into account that with our eyes, what we see on film appears to have more subtle shadow and highlight details. Something that Lightroom 4 in the past year has given us more control over. Also something that many that do create simulations have not taken to task to simulate.

The argument that film has more dynamic range, is more rooted in the fact that the technology to present the dynamic range of digital isn't up to task. Frankly, I've seen the same problems of the film's dynamic range during the scanning of film into a digital format. No adjustments done during scanning results in a very flat dynamic range in the digital file.

Balancing the Rich Reds and Oranges
Last couple of years I had experimented and made into my workflow a digital slide copier from my iPad and my 5Dmk3. The range and image quality I'm getting from that kind of process far surpasses any flat bed scanner I've used, and also rivals or comes close to what Crosfield drum scans. The RVP's when photographed actually look like the images that I'm seeing with my eyes, and as a result when I take it into post processing, reveals a greater range of tones that I can manipulate or adjust.

I don't claim to be an expert with film by any stretch of the imagination, but I will agree that it seems that film has a 'look' that digital has a hard time simulating. For years I've been trying to figure out why? I make some assumptions, look for some theories behind it, and also experiment on my own, ways to simulate them. So my own epiphany came as a result of years of looking into this and looking at the various manufacturer charts for various films. The spectral sensitivity of each crystal had me intrigued, but I really didn't understand what it all meant. It came down to the fact that what I many if not all were trying to do was simulate not just the colour, 'feel', contrast, and clarity of a film, but it's dynamic range.

Preserving shadows without it looking like HDR
Whether or not you like the idea of tone compression or not, I didn't want to just compress the tones and call it a day. I took it much further than that. I wanted to balanced the exposure, but also return some of the lost shadow details back. Another signature of Velvia is 'richer blacks'. What the hell does that even mean? Blacker black? Well, as it turns out, after shooting all the Velvia that I've shot, I've noticed that blacks do look blacker, but it's more than just black is black, it's the tonal range that's found in the shadows. So just pushing the shadows up actually really makes things look HDR like. I had to preserve blacks when I could, give some tonal compression when needed in the shadows, and also reduce some of the contrast to give a greater range of tone. Its a very careful balancing act but in the end I got results that looked very close and pleasing.

One of the other challenges I found in this exercise was balancing the colours after I adjusted for dynamic range. The most challenging of colours was reds and oranges. Another signature of Velvia is Reds can look Orange, Orange look Red, and Yellows look Green (this at least in Velvia F). Velvia RVP 100 is actually pretty accurate with those colours, but balancing Reds often mean that it affects other colours like pinks or even skin tones.

Now normally this film isn't a great film to use for portraits, but RVP does actually respond well enough to skin tone. So I also as a point of reference applied my preset to skin tones to see how it would react. After balancing that parameter and getting a satisfactory result, I returned to landscape images that were rich in red tones to see how it would work. By balancing between the two extremes I found a nice balance of tone compression to pleasing visible range of reds.

Once I got the range I was happy with. I applied it across 16 various images to see if one preset would work across all of them. I know, like different films, that there's no guarantee that it will all respond well to different shooting conditions. Velvia 100 for one would normally be shot under decent amount of light, not low light. But at the end, my presets only ever needed a couple of small tweaks after it was applied in mostly just exposure. In my 16, only about 2 or 3 really needed a little bit of an adjustment (they ended up slightly underexposed).

Last thing I did want to test was range of cameras. I think I have about 4 different camera images that I tested this preset on. They all responded as expected. Now the key is that they are all properly exposed to begin with but even if it is under or overexposed, you should be able to apply this preset with good results (you might need to apply some more noise reduction to your images).

For those of you who use Lightroom 4, you may download my preset here.

Also completely no obligation, if you feel like contributing a little donation to my efforts, I'd be happy to spend more time in producing other simulations in the future. You can use this paypal button to contribute for my efforts.

Select an amount you'd like to donate:

Friday, February 15, 2013

Posted in Arrangement, Art, Business

Just say no to LaCie

With 25 years of imaging and I've had my fair share of various electronics and high tech equipment. Some have been absolute winners while others have been disappointments. But one company that I have to share my displeasure with is LaCie.

Simply put, they make their products just good enough to last either at the end of warranty or just after.

No, my big beef isn't so much about the fact that it fails after warranty, it's more the fact that they have no after warranty program that is reasonable.

Certainly feel that if your product is being sold in Costco (that's not where I bought mine though), there's a little bit of an expectation that it's mass produced and may not be the best product out there. However, regardless of that, I've had more than my fair share of LaCie products fail on me within the end of their warranty or just after.

How they deal with it is that they don't offer an after warranty repair option other than a discounted purchase for a new upgraded product. What is offered is downright offensive which is barely 15% off retail of a new item.

For me, the past 3 years I've been using a LaCie 324 monitor for my workflow. What gave me a 92% AdobeRGB accuracy in a 30bit monitor was definitely a plus for calibrating my photography and fine art prints. With the SpyderPro4, I was able to calibrate it against my other large monitor and compare it to my printed examples which were very good.

Unfortunately the monitor at about 1 or 2 months out of 3 year warranty stopped working one day.

I checked online (click here to see how many people who have had issues) for a few tips and found no real solutions but a bunch of other 324 owners with the exact same problem. Sadly, all of them also indicated that it was almost at the same time they failed. Everyone also indicated that they tried to get LaCie to acknowledge an issue with the model to no avail. They were all out of luck with their reasonably expensive door stops.

I spent a better part of 3 months trying to get answers and acknowledgement from them. Ironically they keep sending me reminder messages to close my trouble ticket on their system. I've left it open as a constant reminder of this poorly handled situation.

Being that I've been marketing multi-million dollar companies with customer service departments for many years, it troubles me that they have products that fail as much as they do (I've had three myself and up until recently kept buying them) with no decent after warranty policy. If I was to consult for that company, they clearly need to outline a service map that deals with customer issues, complaints and out of warranty issues.

Let alone the fact that their products seem to fail within such a short time frame, I'm warning my readers that if you do buy their products, to be aware that you may be faced with similar issues like mine. Get any issues dealt with immediately within your warranty period. However I do have technology that lasts double/triple the warranty period and LaCie's products don't ever seem to go further than that time period (very troubling in my experience). With that in mind, I advise to look elsewhere but if you do have a LaCie product, here's my advice to you. Contact them with your concern that the product is going to fail at the end of the warranty as there are many owners that have indicated failure of their products just after that warranty period. Get them to do an evaluation of your product and insist on any repairs to ensure that it does not fail soon. Many repair services even warranty the repairs.

Insist for a change in their policy if they don't or at least a guarantee that if your equipment fails after the warranty, that there is a repair policy that is reasonable. Had it been even 40% off retail to me that might have been more reasonable to me, but frankly they just don't seem to care that they don't make products that last even longer than the warranty period.

Every company has their service issue challenges. I'm no stranger to that and certainly appreciate some of the efforts made by individuals at LaCie for trying. However clearly not only is their products broken, but a system on how to work with situations like mine in a reasonable and respectful manner. Unfortunately I will no longer be a customer of theirs.

Sunday, February 10, 2013

Posted in Arrangement, Art, Business

Sugar Coated Landscape (Week7)

This week's latest series that features frost on the gardens in Gorge Park in Victoria, BC.

Processing was done in Capture One 7.0.2 which I'm find more and more a part of my workflow. The fidelity and the tonal control that I learned during a personalized visit from a rep from Phase One really opened my eyes. Lightroom and Aperture looks like toys in comparison to what I can do with the advanced features of Capture One.

Granted this software was originally designed for medium format digital, the approach is beyond just using absolute values found in RGB colour models. There's a lot of intelligent algorithms that are behind Capture One that gives it a more analogue like response when doing adjustments vs the more absolute digital approach of most if not all modern RAW processors out there.

I don't even completely understand it, but one such simple principle is when adjusting clarity or contrast, generally speaking it affects saturation (which I've grown accustomed to adjusting). Capture One's algorithms do not affect the saturation levels like other programs, which is both a blessing and maybe a little bit of a curse (because I'm used to that kind of response). The net result is that it preserves colour information when it needs to, and allows for far more advanced colour adjustments and effects.

One of my objectives is to make my images more 3D in appearance on a 2D medium. Capture One certainly gives me that flexibility to simulate that better.

I hope to make some time in the coming weeks to share my experiences with Capture One. I don't want to just review it, but really get into the heart of the program and talk about some of the more advanced techniques in it. From these days out, I will be only using Capture One for digital black and white conversions for sure.

Sunday, February 3, 2013

Posted in Arrangement, Art, Business

Latest Upgrades from FujiFilm (including new RAW Processor)

FujiFilm has been very busy updating all their software and making big announcements. Several of these updates included firmware updates and a couple of lens updates.

One of those updates was supposed to come out on January 18th but was delayed after a bug was later identified. However at CP+ they were able to correct for that bug and release it then.

One other note I'd like to add and hopefully dispel those conspiracy theories out there. FujiFilm has actually shared it's algorithms with all major third party RAW processing software companies. This was actually reported by a senior VP from FujiFilm to the media. The issues with having it incorporated into the software is far more complex than just adding the algorithms, but requires some major re-engineering of their respective software packages. The economy of doing that for a small user base, as you might imagine, is not very cost effective. While 99% of the world is on a traditional Bayer mosaic sensors, the remaining 1% on the X-Trans sensor will have to be satisfied with the software companies modifying their Bayer demosaic algorithms to work with the X-Trans sensor. I've been testing Capture One for some time now and very impressed with the results. SilkPix Developer is still one of the best ones out there, but still a very kludgy and unfriendly interface.

The Updates:

X-Pro1 was updated from version 2.0.1 to 2.0.3 (click here for update). Some highlights to the changes:
  1. Added the new XF 14mm F/2.8 R lens (this lens comes with the buggy 2.0.2 firmware on an SD card.
  2. Improvements in AF performances under various conditions, with the 35mm lens, better contrast detection and improvements to the AE-L/AF-L locking button (more on this later).
X-E1 was updated from version 1.01 to 1.04 (click here for update). Some highlights of the changes:
  1. Added the new XF 14mm F/2.8 R lens (this lens comes with the buggy 2.0.2 firmware on an SD card.
  2. Shutter release now works with external mic or remote.
  3. Improvements in AF performances under various conditions, with the 35mm lens, better contrast detection and improvements to the AE-L/AF-L locking button.
Fujinon XF 35 F/1.4 has been updated from version 2.01 to 2.02 (click here for update). Highlights of changes:
  1. Improvement of AF focus
  2. Improvement of AE-LAF-L button focus in manual focus mode.
Raw File Converter. Lastly, and kind of under the radar is a new update to the RAW processor. Windows 8/7/Vista/XP version and Mac OS X version - ( click here for update). This is a third party processor is made specifically for FujiFilm by SilkyPix. There are no significant improvements from the previous version to report. Some compatibility issues and minor bug fixes. Still the same interface. SilkyPix Developer is the latest version that is vastly superior to the one they made for FujiFilm.

Firmware Before and After Performance:
I tested the firmware by using a stopwatch, close to me and an object far. I started on the far object, and pressed all the way through the shutter and upon click I switched to the near stopwatch and photographed that in the same way. The results of 20 shots was then differenced and averaged. This was set up in EVF and with the Macro setting on to maximize the range of the focus. My far object was 12 feet away with my near being 1 foot away.

Old 2.0.1 Firmware gave me an average of 4.4 seconds for focus from near to far and back to near.  New 2.0.3 Firmware also gave me the same average of 4.4 seconds.

However contrast detection was not the expected improvement in this firmware update, what I will say is improved is the straight press through visual feedback. In the old update, if you pressed the shutter through, you never knew if you got sharp focus until you reviewed the image (in fact you can only assume that it captured a blurry image because that appears to be the last thing you see). In the update, the press through shutter results in a sharp image as the shutter clicks through. So from a functional point of view, now you can trust that the camera is doing what it is supposed to do.

Again tested with Far and Near objects, this time with a 2 foot stop watch, and 15 feet black and white object. Average Focus time from near to far back to near was 2.9 seconds (faster than the last test, however I was shooting in Macro mode in the previous test which explains the quicker average time).

With the AF-L button on manual focus mode, this was a slightly different result. This was the area that FujiFilm claims to improve but hard to be sure with my testing method. I did about 30 shots and averaged the results to see the best numbers I could get. The average worked out slightly slower than the click through method at 4.7 seconds. The .2 seconds difference from click-through is likely my reaction time. The older firmware is slightly slower with a lot of seeking AF errors. Results were mixed from the AF speed acquisition with a similar average around the 4.7 mark, but was certainly unreliable. I would say that an improvement was found in making it more consistent.

What I did notice is that it appears the focusing to far objects is much faster than bringing it back to near. This is something I observed in the update and mostly in the AF-L button press. It might be possible that AF has been slightly improved at the distance where it will matter the most. The visual feedback from click-through shutter is certainly also a more reassuring improvement and most certainly the reliability of of the AF-L lock is definitely improved. Now that my camera is updated I can't test it further, but I tried to think of ways to test it and keep it consistent to see if there was any difference.

No real difference in speed, but big improvements in visual feedback and in reliability in the AF-L/Manual focus lock. High frequency (contrast detection), appears also to be improved, but that might be more subjective than anything. I feel that the visual feedback might account for slight improvements from this update, and will make a difference when using a smaller focusing point.