Excessive contrast in D810 slide copying

[rodeo_joe1]

<p>Edward, short exposure, or high intensity reciprocity failure only affects film. It's due to the finite splitting/recombination time of electrons and ions within the Agx crystals. And besides it acts to lower contrast rather than increase it (long exposure, low intensity RF increases contrast). Digital sensors suffer no such effect at any shutter speed offered on current cameras, as can be easily seen from doing a set of consecutive exposures at various shutter-aperture combinations, or from cutting flash power while opening the aperture to compensate.</p>

[Ed_Ingold]

<p>Rodeo, it's a matter of degree, rather than kind.</p>

<p>The real point is, you don't need nor want to use a flash. The usual means by which exposure is set automatically with flash don't seem to be operable in this application. There is a difference, and it's avoidable.</p>

[Ian Rance]

<p>I wish Nikon had made a high quality version of the old slide copiers that used daylight through a diffuser - they are much easier to use.</p>

[rodeo_joe1]

<p>Here's proof - like anyone should need it - that a JPEG is incapable of showing the contrast range of a slide.</p>

<p>Left is the JPEG version straight from a D800 of a Fujichrome 50 slide. It's not Velvia, but pretty close. Velvia has a bit more saturation and an even higher Dmax. As you can see, the JPEG shadows are completely blocked up, and the highlights are tending to overexposure and lack detail. Mid tones appear darker than seen by directly viewing the slide.</p>

<p>Right is the simultaneous RAW capture adjusted in NX-2. The Highlight and Shadow protection sliders were used to compress the contrast range so that all of the slide detail could be seen. I also made a slight "S" shape in the tone curve to give back some of the punch that was lost by compressing the contrast, and a slight saturation boost to visually match the original slide. There's still a lack of deep shadow detail, but there was none to be seen on the slide either.</p>

<p>I think you'll agree that the RH version is much more satisfactory, and closer to what might have been got by shooting the original scene with a DSLR. Although I strongly believe the shadows would have been better rendered with the D800 directly. (Film - pah! Glad to see the back of the stuff.)</p>

<p>Methodology: I used a cheapish 'zoom' T2-mount slide copier attachment on a D800. Illumination was from a hotshoe mounted flash pointed forward at a white wall in front of the camera, about 2 ft away. The flash was set to 1/4 power manual. Shutter speed was 1/60th. Effective aperture of copier is (at a guess) around f/16. WB was set to flash.<br /> I first tried an I-TTL auto exposure, which was nowhere near correct and about 2 stops underexposed. I deliberately didn't crop the white plastic slide frame out to show that reflected light from the front of the slide is no issue.</p>

<p>Edit: The original was shot on a Nikon FE, lens unknown, but probably 24mm f/2.8 Ai-S Nikkor.</p><div></div>

[Ed_Ingold]

<p>I agree that the RAW version is better. However the JPEG version is subject to processing in the camera, which may be less than ideal. Secondly, both images were uploaded as JPEG files to PNET, which indicates JPEG is capable of rendering the dynamic range properly with the correct processing. Using the flash and exposure in manual mode appears to be the correct solution.</p>

<blockquote>

<p>I wish Nikon had made a high quality version of the old slide copiers that used daylight through a diffuser - they are much easier to use.</p>

</blockquote>

<p>The ES-1 seems to be made well enough. It is all metal, except for the diffuser, and the diffuser could be used with daylight. However "daylight" is not the same in blue sky or clouds. Artificial light is more consistent, and IMO, more convenient, especially with the hours I seem to keep these days.</p>

<p>What the ES-1 lacks is a reasonable means to hold a strip of up to 6 negatives. When I have time, I plan to use an EH-3 holder from my Nikon scanner. It was designed to fit into a single-slide holder on the scanner, so the borders are the same width as in a slide mount. Nothing would touch the negatives except the plastic insert of the EH-3. Like everything else related to Nikon scanners, the EH-3 is like gold these days.</p>

[rodeo_joe1]

<p>Update: I just copied the same slide using daylight reflected off the same white wall and the results are indistinguishable from the flash exposure, apart from a very slight shift in colour balance. There's certainly no noticeable difference in contrast. Exposure needed using reflected daylight (through a window) was around 1.5 seconds.</p>

<p>Edward, I agree that I should have qualified my statement about a JPEG being unable to cope with the contrast range of a slide. Indeed, with the processing available to a 14bit RAW file the contrast of a slide <em>can</em> be tamed enough to fit into a JPEG's limited colour space. But that's not what happens inside the camera, since Nikon don't provide a Picture Control designed to cope with slide copying. And I can confirm that the review image on the rear LCD screen looks extremely contrasty, with almost no shadow detail visible. However I'm not seeing that much difference in the Live View image either. Maybe the D810 is different in this respect.</p>

<p>I found the camera's metering was also very poor at measuring exposure for slide copying. I was consistently getting around 2 stops underexposure when the camera meter was allowed to control exposure, whether I-TTL flash or continuous light was used. Changing the metering mode to spot helped, as did lowering the eyepiece blind for continuous light exposures, but the camera still needed a lot of positive compensation applied before 'automated' exposure came anywhere near to being right.</p>

<p>Another update: This slide-copying malarkey gets weirder and weirder. The T-mount slide copier acts like a manual lens, so the camera defaults to the last non-CPU lens data. It was set to my 50mm f/1.2 lens and gave underexposure as stated. I changed the non-CPU lens data to 55mm f/3.5 and immediately the exposure doubled from 1.3 seconds to 2.5. Changed back and the exposure halved, back again and it doubled - WTF?!</p>

[Ed_Ingold]

<p>Using a flash does not afford an opportunity for the camera to adjust the exposure from slide to slide. Otherwise, the camera will try to keep the slide from overexposing, which permanently loses those highlights. A RAW image offers much more ability to retrieve the shadows, although a JPEG file is probably not far behind in that regard. If you find the camera consistently over or under exposes slide copies, even in continuous light, you can set exposure compensation accordingly.</p>

<p>I find my copy images tend to be somewhat flat and sometimes off color. It is not hard to clean them up in Lightroom. I start using the AUTO settings, then the slider for exposure and sometimes white balance or tint. The same adjustments can usually be applied to blocks of slides taken under similar conditions.</p>

<p>If you use the auto-diaphragm feature, the D810 needs information regarding the lens in order to set the exposure, particularly the maximum aperture. Otherwise, it only knows how many stops from the maximum opening to use. I think the camera can read a manual aperture setting too, but you'd need to check the manual. The D3 is pretty flexible with manual lenses, but the Sony A7 always reads the light with the diaphragm closed to its setting.</p>

<p>It sounds like your copier mounts directly to the camera, using a built-in lens, and has no diaphragm setting. The D810 probably doesn't know what to do, and defaults to the programmed aperture (f/3.5?) you entered. Exposure compensation may be your only reliable option.</p>

<p>I'm using a regular lens (55/2.8 AIS Micro-Nikkor) with an extension tube. The ES-1 is a passive device which screws into the filter ring (52mm) and has a telescoping tube to adjust the working distance. The entire assembly is nearly 10" long, but light and reasonably easy to handle. I prop a WB card at an angle and shine a desk lamp on it from above. I insert and center the slide, and point the whole rig at the card. I check the focus every so often, but it doesn't change unless I bump something. Unlike the D810, the Sony is live view all the time, on the back panel or viewfinder, so it's a natural for this application. I focus at 12x digital magnification, which makes the grain clearly visible.</p>

<p><img src="http://d6d2h4gfvy8t8.cloudfront.net/18079912-lg.jpg" alt="" width="600" height="450" /></p>

[Ed_Ingold]

<p>I'm still copying old slides, and haven't gotten to any as hot as Velvia yet. Most were taken on Kodachrome 25, which is supposed to be hard to scan. Here are a couple of JPEG examples from 1987, converted from a RAW file for PNET, and the other converted in the camera (Sony A7ii). The RAW file has a lot more shadow detail, which could be further revealed with Lightroom processing. Basically, it's the same thing you observe.</p>

<p>I have many options for picture profiles, some pre-programmed (portraits, landscape, sunset, etc.) an others user-programmed, including super-flat "S Log2" gamma, which will squeeze 16 stops into something manageable (mostly for video).</p>

<p>Camera conversion<br /> <img src="http://d6d2h4gfvy8t8.cloudfront.net/18079933-lg.jpg" alt="" width="467" height="700" /></p>

<p>RAW Conversion<br /> <img src="http://d6d2h4gfvy8t8.cloudfront.net/18079932-lg.jpg" alt="" width="467" height="700" /></p>

[pete_s.]

<p>How do you guys expect to get good results using desklamps and what not? CRI anybody? Discontinued spectrum of light?</p>

<p>@Rodeo Joe<br /> Nikon has a utility and you can define your own custom picture control in it and download to your camera. With it you can make custom tone curves. The image you posted that was converted from raw outside the camera, looks like you could get in-camera with a custom curve and the right amount of ADR.<br>

PS. Look here: http://imaging.nikon.com/lineup/microsite/picturecontrol/adjustment/custom.htm </p>

 

[Ed_Ingold]

<p>I checked the output of the LED bulb, and it is very uniform across the spectrum. For $15 ($8 with discount), I have a 13W bulb with a 75W (equivalent) output at 5500K. It barely gets warm to the touch. The photos I attached in the previous post are as-is, no adjustments for exposure or color. If I set a custom white balance, it comes out 5500K just as advertised.</p>

<p>I was skeptical going into this, but the daylight LED bulb meets all my expectations. Even a warm white (3400K) bulb did well. Check the histogram on these images - there's nothing off scale, white, black, or RGB. If you see a problem, speak up.</p>

<p>Slide film, including Kodachrome and Velvia, render color nicely in daylight, especially if underexposed by about1/3rd stop. Let the sun go under a cloud or a few drops fall, all bets are off. It's not just blue, but purple to my eye. Room light is not just warm, but orange. Using a camera to "scan" with AWB seems to rectify some of these color issues, not perfectly, but better overall. Best of all, the exposure stays within bounds.</p>

[keith_b1]

<p>The slide film's density range, when diffusely illuminated(the ES-1 has a diffuser), does not challenge the D810. The slides are not "too contrasty". <br>

Flash does not add contrast; when used with an auto exposure flash sync cord(Nikon SC-29 or equivalent), flash is an excellent and easy-to-use light source. <br>

Most other light sources have a poorer color spectrum(I'm looking at you, LEDs), and therefore will demand more tinkering in post production to get the color where you want it.<br>

Are you on auto WB? Lose that.</p>

<p>Expose via the histogram. The rear LCD is, as is usually the case, of limited value.<br>

Shooting JPEG? Don't bother. JPEGs, in this application, are a way of asking for disappointment. Each different slide film, and sometimes different batches of the same film, and sometimes different labs or different runs at the same lab.... looks different.<br>

<strong>What processing/viewing app are you using to view the raw files? </strong>This probably is the most important issue here.<br>

What display screen? Cheapo laptop, or cheapo LCD display might not be a good choice...depends on how bad they are.<br>

I have shot slides with the D2X, D3, and D800/E and the finished files look pretty good, with a moderate amount of color and contrast adjustment, which I do in Lightroom.<br>

You should be able to get good results also.</p>

<p> </p>

[Ed_Ingold]

<p>Daylight LED bulbs have a reasonably smooth spectrum with a spike about 400 nm (deep blue). The blue LED component is used to excite a broad band phosphor, together with red and amber LEDs, give smooth results. The spike seems to have little or o effect on the results, since it is on the lower edge of the sensor range. Daylight is ideal, and electronic flash is close to daylight in content and smoothness. However 5500K LED bulbs work very well, stay cool and make exposure much simpler. Most of my "scans" can be used with little or no adjustment, mostly to repair balance or exposure of the original to something more appealing. With Lightroom, all adjustments are non-destructive.</p>

<p>https://images.search.yahoo.com/search/images;_ylt=AwrBT4eW_OhV1VkAeCpXNyoA;_ylu=X3oDMTE0NWd1cHM4BGNvbG8DYmYxBHBvcwMxBHZ0aWQDQjAzNjdfMQRzZWMDcGl2cw--?p=led+bulb+spectrum&fr=sfp&fr2=piv-web#id=6&iurl=http%3A%2F%2Fhousecraft.ca%2Fwp-content%2Fuploads%2F2012%2F09%2Fspectral_responses2.png&action=click</p>

[roland_vink]

<p>If the slide shows a full range of tones from deepest shadows to bright highlights, and the dark parts of the slide are dense enough to block light from the flash, then as you increase the flash power you <em>will</em> increase the contrast - the shadows remain dark but the highlights continue to get brighter. To lower the contrast and record the full range of tones, I'd try turning down the flash and diffusing the light (which also tends to reduce grain). Experiment and see what works.</p>

[pete_s.]

<p>Edward, thanks for you info on the LED bulbs. Seems like they are better than I thought.</p>

<p> </p>

[john_hinkey]

<blockquote>

<p>The slide film's density range, when diffusely illuminated(the ES-1 has a diffuser), does not challenge the D810. The slides are not "too contrasty".</p>

</blockquote>

<p>As others and I have stated, the density range of the <em><strong>recorded</strong></em> image on slide film can have a huge dynamic range and, depending on the image, can indeed challenge a D810.<br>

When on my slide table I can clearly see details in the deep deep shadows that cannot be captured in one exposure on a D8XX camera while also getting the bright highlights. The only way I've found to get all there is out of a slide that has highlights and deep shadows is to combine multiple exposures.</p>

[Ed_Ingold]

Contrast increases with exposure? A digital image cannot get whiter than white. The shadows can open up.

 

It's easy to program a bracketed exposure on one burst. I haven't needed to do that yet, but it's another tool in the box.

[rodeo_joe1]

<p>Edward, the issue with the non-CPU lens data is that it <em>shouldn't</em> alter the camera's metering in the slightest. The camera meters whatever light is entering its viewfinder - regardless of the lens aperture. With the copier in place it's as if any non-CPU MF lens was fitted and set at wide open aperture. Therefore, telling the camera it's f/1.2 or f/3.5, or even f/11, should make no difference to the amount of light the meter sees and consequently to the shutter speed it sets. Except obviously it does; and it's this aspect that I find worrying and in need of further investigation. It really doesn't matter that I'm using a slide copier, it could just as easily be the non-auto bellows I have, and that will bug me if the exposure depends on what lens I tell the camera is attached.<br>

What happens on a Sony Alpha 7 really is of no interest or relevance here.</p>

<p>WRT light source CRI - that's not terribly important either. A slide film image consists of Cyan, Yellow and Magenta dye blobs that vary very little in transmission properties. The CMY dyes have a maximum density to a certain narrow range of Red, Green and Blue light bands respectively. As long as the illumination source puts out any light in those regions then it'll be picked up by the camera sensor - hopefully! That's if the narrow-cut RGB Bayer filters allow it, since they're lacking in true (monochromatic) Cyan and Yellow transmission. So a light source may have a poor CRI with a peaky spectrum, but have strong bands that correspond to the complement of the CMY dyes in the slide film.</p>

<p>Having said that, there is an issue with using continuous illumination sources with a low colour-temperature. The camera's WB needs to be adjusted radically to cope with, say, domestic tungsten illumination, and this has a consequence for the dynamic range it can capture. Boosting the Blue channel sensitivity is akin to raising the camera's ISO, and this will cut the DR available. This leads to the conclusion that the best light source to use is one with a CT close to that of daylight, where the camera sensor is engineered to have its highest sensitivity. So short of waiting for a nice sunny day, then flash would seem to be a near ideal illuminant. And a tungsten reading lamp a no-no.</p>

[pete_s.]

<p>Rodeo Joe, good point about the transmission of light from illumination to sensor.</p>

<p>Regarding the non-CPU information I have in earlier threads explained my findings about ISO boost depending on aperture as well as how the camera changes the exposure (with fixed ISO, aperture and shutter speed) based on the non-CPU info. That should also not happen in theory.</p>

<p>However it does and I think that in the camera Nikon changes a number of parameters depending on the non-CPU settings, basically giving it properties of their own lenses. So if you enter 24mm f2.8 in the non-CPU info the camera now thinks you have a Nikkor 24mm f2.8 AI/AIS lens attached and uses the characteristics of that lens to adjust the in-camera meter, vignetting information, perhaps also lens distortion and CA.</p>

<p>The regular CPU info chip inside the lens also have information like exit pupil distance. I think that information is used for exposure and metering corrections since it will influence the exposure due to pixel vignetting. And as said above it looks like Nikon starts assuming things when you enter non-cpu info in the camera. And I think this is what you are seeing.</p>

<p>The camera tries to be smart but it's not smart enough to get it right. Matrix metering for instance has never impressed me. Only time the camera gets it right is when the scene is easy to get right. Start adding high contrast, varying degrees of subject reflectance, on camera flash, off camera flash, backlit subjects, wide angle lenses etc and it all starts to fall apart. A few years ago I did a survey for wedding pros how much they had to adjust exposure in the raw converter after the fact. It turned out that most people were in the +/- 1 stop range. And that was pros who knew their cameras very well and probably had already used exposure compensation in the camera in scenarios where they knew the camera would over or under expose. Off topic rant over :-)</p>

<p> </p>

[Ed_Ingold]

<p>Is the D810 fully compatible with a non-CPU lens? Your exposure changes when you change the programming, so something is happening.</p>

<p>Some cameras have a coupling right that connects to the aperture ring on the lens. This is an expensive option, often omitted on lesser models. On my D2x, a non-CPU lens would communicate the position of the aperture ring to the camera body. If I programmed the maximum aperture, the f/stop would read directly in the viewfinder when the aperture ring was turned, and the diaphragm would remain open until the moment of exposure. With a CPU, the programming is automatic (I think).</p>

<p>I would not recommend using the maximum opening for closeups, especially at 1:1. The DOF is very shallow and the optical quality is not as good as two or three stops closed.</p>

<p>AWB may be limited to 3200K or higher, but you can usually set it to a lower value manually, at least to 2800K. An halogen bulb is right on the edge, about 2650K. The results come out on the warm side, but can be corrected in Lightroom. The warm white LED I used was 3400K, using the camera as a gauge. The daylight LED I now use is 5500K, the copies show no signs of excessive blue (the spike), and the exposure is right on target.</p>

<p>With a continuous light source, measuring and bracketing exposures is simple and effective. The way a D810 measures flash exposure is far less accurate than the matrix meter. Basically, it uses a single cell to measure light reflected from paint on the shutter in the moment before exposure from a pre-flash. The technology harks back to the D3 and before. Setting up a bracketed flash exposure would require a little head-scratching and a lot more time to execute.</p>

<p>Rather than waste time explaining what can't be, concentrate on what is, through observations. Theory should fit the facts, not the other way around (unless you are a politician).</p>

[Ed_Ingold]

<p>I downloaded the D810 manual, and it appears to be fully compatible with non-CPU lenses (p229, pp419ff). There is a coupling ring on the mount which engages the aperture ring on a AI and newer lenses.</p>

<p>Rodeo, did you say your slide copier attachment is a T-mount (i.e., mounts directly to the body)?</p>

<p>For anyone interested, I'm using a Nikon AIS lens with an adapter to the Sony body, with no data connections and only a mechanical coupling in the adapter to the lens aperture. Everything I'm using from the adapter out is compatible with the D810. The 55/2.8 Micro-Nikkor and PX-13 extension tube and cost about $150 used from KEH, plus a new ES-1 copying attachment, $60 from B&H.</p>

<p>As a side note, that vintage Micro-Nikkor is as sharp as any new lens for the Sony A7 at a fraction of the cost. I'm amazed at its performance. The bokeh is a little rough, but it's a winner. I used one for years for many things besides closeups, with almost no flare when shooting into the sun. That one disappeared, but I'm glad I found another for this project - hundreds of unscanned slides.</p>

[sebastianmoran]

<blockquote>

<p><strong>What processing/viewing app are you using to view the raw files? </strong>This probably is the most important issue here.<br>

</p>

</blockquote>

<p>And... Are you checking the image in an application that does color management?</p>

[chuck]

I am using Lightroom 6.