Using the S2 and the M9 together ...LR presets?

[Roger]

Last week I got to use both the S2 and the M9 at the US Open. My goal was to shoot the event as one of America s iconic sporting events (verse trying to get the cover of Sports Illustrated) . I used the S2 for 2 days and my 2 m9 s for the other three.

I know from past experience that the images from both cameras can be rendered to look very similar in LR . But seeing the files shot in essentially the same light and then using Josh s presets for the S2 files brought out some differences .

The S2 files seem to be exposed to the right situations ..the S2 really got most of the exposures right without much correction. A normal file looked too light when compared to the M9 files .

The I looked at Josh s presets and I see that he really pulls down the shadows and darks compared to what I would do with an M9 file. Once I apply the preset ....the files start to look similar.

So two questions :

1. Do you have a similar set of presets for the M9 ?

2. Have you evaluated how clarity works . In the past clarity affected the mid tone contrast but my understanding is that it also brings in the deconvilution algorithms. Some of the LR gurus are now recommending 70-100 setting for images with a lot of fine detail in the midtones . I have always used around 20 but on a .NEF files I use over 50 always .

It would be nice to create presets that produced a standard rendering on import so that the S2 and M9 files would look similar when starting the edit.

[Josh Lehrer]

Roger:

I tend to add more shadow contrast to the S2 because the files tend to be a bit flatter out of camera due to the increased dynamic range that the files have compared to the M9, which I find has excellent contrast straight out of the camera. I always prefer slight overexposure when shooting with the S2, yet another reason for me to add a bit more shadow contrast.

I do not have a similar preset for the M9, but I am sure that I could make one for you. I will post it here once it is created.

As for clarity, I tend to keep it under 10. Its primary function is to add mid tone contrast, and while I have not experimented to see if it adds any sharpening, my personal taste has always been minimal clarity and then separately adjusting my sharpness. I also find that too much clarity slightly desaturates the mid tone areas of the images as well.

If you create an S2 preset with added clarity that creates a look that you like, I would certainly suggest posting it in this thread for myself and other users to experiment with.

[Roger]

Thanks Josh

I know as you build an archive overtime that there is an advantage to having some consistency in rendering. For example when the pros look back at their life s work ..its nice that everything they shot on kodachrome or tri x has a similar look. For example look at the Leica video on Steve McCurry ....the images that were taken years apart hang together .

So my goal is to have the S2 and M9 files render in a similar way .

My question on clarity really comes from more recent insights on LR3 and how Adobe has worked in improvements in the raw conversions. This may be more related to Nikon files which suffer from the AA filter and benefit greatly from deconvolution algorithms . But the logic was similar above 50% the algorithms start to kick in. I will ask on the adobe forums. (most leica users developed their presets during prior versions of LR ) .

[fotografz]

Interesting thread. I currently do not use any presets on S2 images or M9 shots for that matter.

I almost always shoot the S2 and M9 together at weddings using the metadata filters to work on each camera group separately, and rely on visual adjustments to maximize the file quality ... depending on the ambient conditions.

While I do use some Clarity with the S2 files ... for me, it is a dangerous slider because of what it can do to skin captured at that level of resolution, so I more often than not use localized Clarity over a wholesale adjustment.

Where the midtone contrast shows itself most is when doing B&W conversions. When shooting both cameras in the same ambient conditions with-in seconds of one another, I actually find the M9 and S2 fairly similar (a bit flatter) compared say to the M8.

When doing B&W conversions in Nik Silver Efex I often start with the same preset for the B&W look I want ... Push Process-1 ... then finely adjust from there by hardening the grain and increasing the grain per pixel, often to the max. Again, localized exposure adjustments are used to refine a file depending on the ambient ... Brightness, Contrast and Clarity. Or, I scroll through the film types to see which one visually hits the mark for that specific file then make refinements.

-Marc

[David Farkas]

Thanks Josh

I know as you build an archive overtime that there is an advantage to having some consistency in rendering. For example when the pros look back at their life s work ..its nice that everything they shot on kodachrome or tri x has a similar look. For example look at the Leica video on Steve McCurry ....the images that were taken years apart hang together .

So my goal is to have the S2 and M9 files render in a similar way .

My question on clarity really comes from more recent insights on LR3 and how Adobe has worked in improvements in the raw conversions. This may be more related to Nikon files which suffer from the AA filter and benefit greatly from deconvolution algorithms . But the logic was similar above 50% the algorithms start to kick in. I will ask on the adobe forums. (most leica users developed their presets during prior versions of LR ) .

Roger,

I think you may be confusing clarity with detail. The detail slider under the sharpening settings acts in this way:

0 = 100% USM
50 = 50/50 USM + deconvolution
100 = 100% deconvolution

I prefer to set the slider to 75, which emphasizes the deconvolution method which I prefer for textural detail with a bit of USM which enhances edge contrast. Here's the sharpening settings from my S2 preset (David's LR 3.4.1 Preset - S2.lrtemplate) which I use as a starting point for all my S2 files:



My clarity is pegged at +7 and never moves. I have, on occasion, brushed in some negative clarity using the local adjustment brush. Just a little goes a long way in a portrait without softening the image overall.

[ka7197]

I think you may be confusing Clarity with Detail.

Yes, I think that's what he did. His question about Clarity doesn't make any sense—until we swap Clarity for Detail.

The Detail slider under the sharpening settings acts in this way:

0 = 100 % USM
50 = 50/50 USM + deconvolution
100 = 100 % deconvolution

Oops—David, I'm afraid now it's you who's being confused. In fact, the Detail slider in ACR's and Lightroom's sharpening group of controls acts like this:
0 = 100 % deconvolution
50 = 50/50 USM + deconvolution
100 = 100 % USM

This slider is named "Detail" for a reason. In general, it makes sense to push it up to 50 - 100 for high-frequency images (i. e. images containg lots of fine and sharply rendered detail, like landscapes) which have low or no noise. For this kind of images, you'd also use low or no Masking. For low-frequency images (e. g. portraits), you'd pull Detail down to 0 - 30 or thereabouts (and use more Masking).

Things will get tricky when you have lots of detail and lots of noise—then you'll need to find a good combination of Detail and Masking (and noise reduction, too) that works. In this situation it might be tempting to push both Detail and Masking up ... but this may end up looking unnatural because it may lead to plasticy areas with over-sharpened edges.

Furthermore, I think that "100 % USM" is a bit of a simplification. In fact it's similar to USM but not exactly the same. Still, for practical intents and purposes it's close enough to think of it as USM.

To find to best settings for capture sharpening, I often apply this workflow in Camera Raw or Lightroom:

• Disable the noise reduction settings, i. e. set both the amounts for luminance smoothing and colour noise reduction to zero. Set the display view size to 200 %.
• In the sharpening group of controls, put the Detail slider at medium setting, Masking at zero, and then boldly push Amount up to 100. You'll see a vastly over-sharpened image preview.
• Grab the Radius slider and move it all the way from 0.0 to 3.0 and back again a few times. The trick is to move the slider at the right speed. It should be as fast as possible but still slow enough for the preview display to follow the slider's motion. Having a fast machine with a powerful graphics adapter helps. This way you will see the sweet spot for the Radius setting; it will really pop out at you.
• Push the Detail slider to 100 and then slowly back until the sharpening halos look reasonable. Do not try to get rid of them altogether but find a good compromise between halos and sharpening effect.
• Now reset Amount to zero and then slowly push it up until the sharpening effect looks good overall. From there, back up again a bit or two, as you usually will overshoot the Amount. Remember, this is capture sharpening, not output sharpening.
• Adjust Masking to taste and apply some fine-adjustment to Detail if required.
• Finally, adjust noise reduction. At maximum view size (400 %), I always pull Color and Color Detail down to zero, then push up Color step by step (in increments of 5 for low-noise images or increments of 10 for high-noise images) until I don't see any further improvement, then adjust Color Detail to where some of the colour noise is just on the brink of re-appearing—which means some serious pixel-peeping.

[Stuart Richardson]

Regardless of how the sharpening in lightroom works, I have found that for the M9 at least, the detail slider is best used sparingly. I know this runs counter to most people here. This is primarily based on my experience as an exhibition printer. When printing M9 images large -- i.e. 20x24"/50x60cm and over, you want to be very careful with sharpening. There is a very fine balance between creating a print that appears sharp, and a print that appears very sharp but also has visible digital and sharpening artifacts. When you are printing really big, a slight softness at the 100% detail is more natural and desirable than ramping up the sharpening, which makes those digital artifacts that much clearer. This is also one of the reasons it is possible to enlarge film so much and still get great results -- there is a graceful roll off in resolution which is more natural to the eye than digital's abrupt transition from high-contrast detail to unresolved detail. Managing that transition is one of the most important parts of exhibition printing.

I also think the extreme sharpness of the M9 and S2 (and other non-AA filter cameras with sharp lenses) demands an entirely different sharpening technique than other AA filtered digital cameras. In practice, I have found that mid levels of sharpening (40-80), low radius (.5), low or no detail and medium masking (40-80) does the best job of accentuating the sharpness of what needs to be sharp (such as obvious edges), while not sharpening noise or accentuating digital artifacts.

On the other hand, for smaller prints a bit more sharpening can be a nice addition, as its impact on the print will not be as clearly resolved by the printer, allowing the global effect to be more to the fore than the significant manipulation you are doing at the actual pixel level.

[Stuart Richardson]

And just another note -- Eric Chan confirmed that the detail slider towards the higher side brings into play deconvolution sharpening -- so that would mean that David has it right...
http://www.luminous-landscape.com/fo...?topic=45038.0

Also, it would seem to make sense to me anyway as to why the detail slider is best used sparingly with the M9 and S2 -- since these cameras do not have AA filters and have exceptionally sharp lenses, there is less convolution (blur) than in traditional cameras. So when you try to deconvolute that which is not convoluted in the first place, it starts to look weird! That's my take anyway...

[David K]

I'm far from a LR guru and still prefer a round trip to CS5 for most adjustments. But I'd be willing to bet a couple of bucks that Roger is NOT confused about anything in LR.

[David Farkas]

And just another note -- Eric Chan confirmed that the detail slider towards the higher side brings into play deconvolution sharpening -- so that would mean that David has it right...
http://www.luminous-landscape.com/fo...?topic=45038.0

Also, it would seem to make sense to me anyway as to why the detail slider is best used sparingly with the M9 and S2 -- since these cameras do not have AA filters and have exceptionally sharp lenses, there is less convolution (blur) than in traditional cameras. So when you try to deconvolute that which is not convoluted in the first place, it starts to look weird! That's my take anyway...

Stuart,

Here's what Eric Chan wrote over on LL:

Yes, the Detail slider in CR 6 & LR 3 is a blend of sharpening/deblur methods and if you want the deconv-based method then you crank up the Detail slider (even up to 100 if you want the pure deconv-based method). I do this for most landscapes and images with a lot of texture (rocks, bark, twigs, etc.) and I find it's not bad for that.

I've emailed back and forth with Eric and we had a nice chat last year at Photo Plus. He's a wealth of information and a lot of my LR settings are based off his advice, as well as that of Stephan Shulz. Eric is a senior developer for Adobe and actually rewrote much of the sharpening algorithms for LR3.

[Stuart Richardson]

Thanks David,
I read that bit. I still think his advice probably applies more to standard cameras. Did he mention specifically that he was recommending high levels of detail for non-AA filtered cameras? Because if he is, I would still disagree with him! I certainly am not saying you cannot get decent results by using high levels on the detail slider, but if you do want to go that way, you need tone the sharpening way back down into the lower levels (teens to 30s) otherwise things start looking freaky. Again, that's just my taste, and I am sure it differs. But, here is an example, just so I can demonstrate what I am talking about. I am including 3 screen shots. The first sharpened the way I normally would, using 0 detail in the sharpening, and then two with 100% detail. One adjusted to eye again, and the other at the same sharpening level as it was with the detail at 0. I think the deconvolution sharpening can work well, but at least with the M9, it seems to be really aggressive and prone to artifacts. Again, I think this makes a lot of sense in that the M9 sensor type is very different from the vast majority of cameras, which is clearly what Lightroom's deconvolution algorithms are designed to deal with.

[David Farkas]

Thanks David,
I read that bit. I still think his advice probably applies more to standard cameras. Did he mention specifically that he was recommending high levels of detail for non-AA filtered cameras? Because if he is, I would still disagree with him! I certainly am not saying you cannot get decent results by using high levels on the detail slider, but if you do want to go that way, you need tone the sharpening way back down into the lower levels (teens to 30s) otherwise things start looking freaky. Again, that's just my taste, and I am sure it differs.

A while back, Eric wrote me this by email:

Regarding the sharpening settings: One of the things that has changed in PV 2010 is the effective Radius setting. Previously it was very difficult to extract the finest details (pixel-level details, like texture in bark, grass blades, fur, hair, etc.) even when the radius was set to its smallest value of 0.5. So with PV 2010 we’ve internally adjusted the Radius scale, so that in the [0.5, 1.0] range, it’s thinner than it was previously. In other words, an edge halo with PV 2010 at Radius = 0.5 is thinner than it was when using the same radius setting and PV 2003. The thinner halo is somewhat less visible, which can give the impression that the image is less sharp. Please try increasing the Radius to a value in the 0.8 to 1.2 range and see if that helps.

If you have a low-noise image (e.g., ISO 80 or 160, or a well-exposed ISO 320 image) that has a lot of natural image texture (e.g., landscapes), you can also try increasing Detail to very high levels — even 75 or 100! This will extract a lot of texture and small details, but of course can make noise more visible. This is why it’s more effective on low-noise images.

And, yes, he is referring to S2 processing specifically. Eric was responsible for working with Leica's digital team in Germany to tweak the processing for the S2.

I don't ever turn my sharpening amount past 45, but rarely do I ever turn it down, either.

Here's a 100% crop using my standard sharpening settings (taken from my recent S120 vs P105 article):

[Stuart Richardson]

Thanks for the information David! That photo looks very well sharpened too. I will experiment with higher levels and see if I can't be won over. I will have to see how it works in big prints too, as that is my main criteria...I usually use photokit's output sharpener in tandem, so it may act differently if the capture sharpening technique is different...

[Roger]

Not confused but maybe misinformed. I did a LR class with Kevin Ames ...the photoshop guru that writes about beauty retouching. He was all over the use of the clarity slider and was recommending very high settings (for landscapes) . I know on the Leica files that clarity around 20 works for me but on the Nikon files I go up to 50 . Marc is correct that most recommend using the local adjustment brush when working with skin tones and that clarity can often go slightly negative.

Not really all that important as my primary objective is to create files that are similar when using both the S2 and the M9. The two cameras work well together as they should .

[ka7197]

The detail slider under the sharpening settings acts in this way:

0 = 100 % USM
50 = 50/50 USM + deconvolution
100 = 100 % deconvolution

Oops—David, I'm afraid now it's you who's being confused. In fact, the Detail slider in ACR's and Lightroom's sharpening group of controls acts like this:
0 = 100 % deconvolution
50 = 50/50 USM + deconvolution
100 = 100 % USM

It seems both of us, David and me, were wrong here. In fact, in Camera Raw's/Lightroom's capture sharpening there is no USM involved at all. David is right as far as the Detail slider's right end (labeled "100") is concerned—that's 100 % of a deconvolution-based sharpening method. But at the left end there's no USM but another sharpening method applied. I don't know which method exactly but it's smoother and creates less haloes than the deconv-based method at "100" ... and definitely is much smoother than USM would be. In comparison, pure USM is a rather coarse sharpening method—that's why I thought it must be the method used at the scale's "100" end.

[fotografz]

A while back, Eric wrote me this by email:

Regarding the sharpening settings: One of the things that has changed in PV 2010 is the effective Radius setting. Previously it was very difficult to extract the finest details (pixel-level details, like texture in bark, grass blades, fur, hair, etc.) even when the radius was set to its smallest value of 0.5. So with PV 2010 we’ve internally adjusted the Radius scale, so that in the [0.5, 1.0] range, it’s thinner than it was previously. In other words, an edge halo with PV 2010 at Radius = 0.5 is thinner than it was when using the same radius setting and PV 2003. The thinner halo is somewhat less visible, which can give the impression that the image is less sharp. Please try increasing the Radius to a value in the 0.8 to 1.2 range and see if that helps.

If you have a low-noise image (e.g., ISO 80 or 160, or a well-exposed ISO 320 image) that has a lot of natural image texture (e.g., landscapes), you can also try increasing Detail to very high levels — even 75 or 100! This will extract a lot of texture and small details, but of course can make noise more visible. This is why it’s more effective on low-noise images.

And, yes, he is referring to S2 processing specifically. Eric was responsible for working with Leica's digital team in Germany to tweak the processing for the S2.

I don't ever turn my sharpening amount past 45, but rarely do I ever turn it down, either.

Here's a 100% crop using my standard sharpening settings (taken from my recent S120 vs P105 article):

Attachment 583

David, this sample I find to have an artificial sharp feel to it ... similar to clipping out the subject and dropping it into another background. Especially evident camera right on the pipe bend. It may have to do with the contrast of the lighting and contrasting colors, but the edge sharpening is heightening that as opposed to mitigating it.

The sense of edge sharpness has always been one reason I could never quite take to many Japanese optics which often produced this type of effect with no help from post at all. I find it to be unnatural, and have felt that way for a long time. The DMR and R optics rarely produced this feel, where my Nikon D3/D3X produced it frequently.

Leica has always had the reputation of producing a sense of sharpness via "Micro Contrast" ... which often worked so well that even slightly mis-focused images still looked sharp printed to normal viewing size.

Despite howls of protest from die-hards, the newer ASPH M lenses still have the Micro-Contrast thing going for them IMO, and IF you avoid over sharpening can produce natural looking, sharp images. It is there that the M ASPH optics and S lenses seem to be most similar. Thus Rogers request comes naturally IMO.

As I mentioned, I don't use a sharpening preset because each lighting scenario is so different, and I find adjusting the combination of contrast, clarity and sharpness to be subjectively different with each image. Where I found that to be very evident is when working with Silver Efex and their localized adjustment tool which has those three sliders attached to the brush. The adjustments show exactly how these three effect one another.

This is very interesting subject to me because, the more "perfection on paper" we strive for, the more likely the results can get overly clinical looking and the natural feel that has always been the Leica look starts looking like something made in Japan. No disrespect meant here, just a different perspective as to look and feel that I freely admit is fully subjective.

A perfect example of this was the claim some years ago, backed up with charts and such, that Mamiya made about their optical performance compared to Contax 645 lenses ... touting Mamiya as being sharper. While the Zeiss lenses for the Contax 645 were made in Japan, they were designed in Germany and over-sight done by the Carl Ziess Institute in Japan. IMO, and the opinion of many others, the Zeiss lenses where far more more desirable for the look they produced ... also likely due to Micro-Contrast, and very smooth focus fall-off as opposed to edge sharpness and a more abrupt focus fall-off

On a separate note: I just watched a Video of studio work using the Freemask transmitter available with Hensel lighting. It fires strobes in a burst that lights the subject on one shot and then only the background on the next shot ... effectively creating a perfect mask for stripping. What struck me was how natural the edges were when using this. Unfortunately the S2 nor any MFD camera that I know of fires fast enough to use this.

-Marc

[madmanchan]

You can use high Detail amounts (even 100) with S2 files, but please do so carefully. High Detail means ACR/Lr will do its best to extract every ounce of detail & texture available in the file. This is almost always undesirable for portraits or perhaps even general-purpose use where a smoother result is preferred. But for extremely fine detail (pixel-sized features like blades of grass, twigs, leaves, etc.) and texture + low ISO cases, a high Detail setting can be appropriate and highly effective. For such scenarios I also suggest reducing the Radius setting from its default value of 1, to something like 0.5 or 0.6.

Eric Chan
Camera Raw Engineer

[David Farkas]

You can use high Detail amounts (even 100) with S2 files, but please do so carefully. High Detail means ACR/Lr will do its best to extract every ounce of detail & texture available in the file. This is almost always undesirable for portraits or perhaps even general-purpose use where a smoother result is preferred. But for extremely fine detail (pixel-sized features like blades of grass, twigs, leaves, etc.) and texture + low ISO cases, a high Detail setting can be appropriate and highly effective. For such scenarios I also suggest reducing the Radius setting from its default value of 1, to something like 0.5 or 0.6.

Eric Chan
Camera Raw Engineer

Eric,

Welcome to Red Dot Forum and thanks so much for posting!

I am curious about the relationship between radius and sharpening amount. For high detail scenarios, where you suggest reducing the radius to 0.5 or 0.6 (I use 0.8 by default), would you turn up the sharpening amount as well?

Thanks again for chiming in.

[madmanchan]

Eric,
I am curious about the relationship between radius and sharpening amount. For high detail scenarios, where you suggest reducing the radius to 0.5 or 0.6 (I use 0.8 by default), would you turn up the sharpening amount as well?

Not necessarily. I think of Amount as a "volume" control and hence it's for seasoning to taste. In contrast, Radius and Detail are more for adapting the type of sharpening to the scene content.