Debunking ETTR

[lukerossknecht]

<p>Sounds like fun. Me, I just shoot on P mode all the time since let's face it, the camera probably really is smarter.<br>

<br /> *pokes hornet nest with stick*</p>

[michael_young3]

<p>To know and embrace your limitations is sublime. This hornet has teeth.</p>

 

[joe_c5]

<p>Sorry to be unclear, I was talking about the neutral patches. I listed only the green values rather than the RGB values for each.</p>

<p>There appears to be some nonlinearity, since most of the neutral patches are either spot on or sort of close, but the lightest is 0.4 stops too dark.</p>

[lukerossknecht]

<p>Michael, why not post some real-world examples so the "sublime" among us can see the benefits of ETTR? It would be fun and educational for all.</p>

[ben_goren]

<p>Luke,</p>

 

<p>I thought “P” mode was the one all <strong>P</strong>rofessionals used…?

Must be why I could never figure it out….</p>

 

<p>I suppose I should be pedantic for a moment and note that I’m not advocating blindly

trusting the camera’s meter. Rather, I’m suggesting that, whenever the dynamic range of

the scene (or, at least, the parts you care about) falls within the dynamic range of your camera, your

best bet is to use a classical standard film exposure (such as by metering off a gray card).</p>

 

<p>If your camera doesn’t have enough dynamic range, either use traditional techniques to

reduce the range of the scene (light / graduated ND filters / <i>etc.</i>); bias your exposure to preserve

one or the other of shadows or highlights (being prepared to fix the exposure in post-production and

being aware that shadows get noisy while highlights suddenly clip hard); or use HDR techniques.</p>

 

<p>Cheers,</p>

 

<p>b&</p>

[lukerossknecht]

<p>Ben, I understand, I was being a little facetious.</p>

<p>Actually I believe PASM modes stand for (P)erfect, (A)mateur, (S)crew-up, (M)ystery :))</p>

[sasvata__shash__chatterjee]

<p>As far as the color shifts, I am confused by that. The only reason that could happen is if some of the color information is being lost (clipped) in the process. I don't know how the 5DII works, but if it shows histograms for the RGB color channels, those should be paid attention to in addition to the luminance histogram. The red channel in particular can be clipped to the right before the luminance channel is when exposure is progressively increased.</p>

<p>If that weren't the case, I would think, simple exposure adjustments up and down would introduce color shifts in addition to brightness changes.</p>

[ben_goren]

<p>No, no, no! It’s Pears, Apples, Strawberries, and Mortgage lien priority renegotiation. Of this

I’m most certain.</p>

 

<p>Cheers,</p>

 

<p>b&</p>

[lukerossknecht]

<p>The important thing is that we all agree that the Zone System works perfectly and is easily applied to color digital photography.</p>

[digitaldog]

<blockquote>

<p>How about some real-world examples.</p>

</blockquote>

<p>Quick and dirty, handheld (so they don’t line up exactly, they do of course in my article). This is showing the two side by side in Compare mode in Lightroom. I could probably tweak them a tad closer, but I think anyone can see the differences in terms of noise here:<br>

<img src="http://digitaldog.net/files/ETTR.jpg" alt="" /></p>

[digitaldog]

<p>Probably easier to see this unshrunk down as Photonet appears to place images here:<br>

http://digitaldog.net/files/ETTR.jpg</p>

[michael_young3]

<p>So much for sublime. Search your feelings, Luke. Sublime, knowing yourself is. Important, knowing your camera is, too. Seek our own truth, we all must.</p>

[lukerossknecht]

<p>Whoops, above I should have referred to Andrew, not Michael. Still Andrew came along and posted some examples anyways - thanks!</p>

<p>Interesting that while the ETTR side has less noise, the normal(? default) exposure has (IMO) better colors in the shadows and midtones. How much extra work does the ETTR technique require in terms of color correction?</p>

[lukerossknecht]

<p>Internet forums lead to anger. Anger leads to trolling. Trolling leads to... suffering...</p>

[dan_south]

<blockquote>

<p>Actually I believe PASM modes stand for (P)erfect, (A)mateur, (S)crew-up, (M)ystery :))</p>

</blockquote>

<p>Are you sure it's not (P)eople, (A)nimals, (S)ports, and (M)oon? ;~)</p>

[digitaldog]

<blockquote>

<p>Interesting that while the ETTR side has less noise, the normal(? default) exposure has (IMO) better colors in the shadows and midtones. How much extra work does the ETTR technique require in terms of color correction?</p>

</blockquote>

<p>A few clicks here or there. I wish Lightroom would allow you to use the compare mode while in the Develop module, it would make it far easier to make adjustments to match while viewing the two. As I said, this was a quickee. It shouldn’t be much work to make them closer (after one decides which is the preferred current rendering although you say you prefer the non ETTR). Point is, its pretty clear <strong>what the difference are here in terms of the noise</strong> in one versus the other. And as I mentioned in the article and above, they will never match perfectly, at least using the tools at in LR (one could argue in Photoshop, you might be able to do a tad more tweaking). But getting two identical renderings isn’t the point. Its showing that by taking an image that appears “over exposed” with the default rendering (in this case LR) and just adjusting the Exposure slider, you can get back to a pretty normalized look with far less noise. </p>

[michael_young3]

<p>I lock a different image on the secondary display to do that. Of course, WinXP supports only one monitor profile, so the secondary window, on the other monitor, is just the tiniest shade cooler... FWIW.</p>

[digitaldog]

<blockquote>

<p>I lock a different image on the secondary display to do that. Of course, WinXP supports only one monitor profile, so the secondary window, on the other monitor, is just the tiniest shade cooler... FWIW</p>

</blockquote>

<p>Great idea, I’ll try that. I’m on a Mac so both are profiled, but one is a wide gamut 30” display (an NEC 3090) which I really trust and the 2nd is an sRGB-like Cinema display which, well kind of sucks. </p>

[lukerossknecht]

<blockquote>

<p>A few clicks here or there. I wish Lightroom would allow you to use the compare mode while in the Develop module, it would make it far easier to make adjustments to match while viewing the two. </p>

</blockquote>

<p>Why not just select them both and hit Autotone? You know you want to.</p>

[digitaldog]

<blockquote>

<p>Why not just select them both and hit Autotone? You know you want to.</p>

</blockquote>

<p>I thought you were kidding but it actually did a decent job (with no additional tweaking!):<br>

http://digitaldog.net/files/ETTR2.jpg<br>

When I open the midtones a bit using Photoshop’s levels, you see this:<br>

http://digitaldog.net/files/ETTR3.jpg</p>

[ben_goren]

<p>Okay, here we go. Lots more examples.</p>

 

<p>First, as promised, here’re the same two 100% crops from above, but with the following

changes:</p>

 

<ul>

<li>Color noise reduction is now set to 8 (was 0; luminance still 0).</li>

<li>The medium contrast tone curve (instead of linear) is now applied.</li>

<li>Blacks are now set to 4 (instead of 0).</li>

<li>I made fractional adjustments to the exposure to visually match the two patches to each

other.</li>

</ul>

 

<p>In other words, this is a much better real-world comparison than the previous.</p>

 

<p>Much more interesting examples to follow.</p>

 

<p>Cheers,</p>

 

<p>b&</p><div></div>

[michael_young3]

<p>EXIF! We want EXIFs. All kidding aside, that's a bit more startling than I imagined from my own results. What camera? Exposures? (Edit: For Andew, that is. Ben snuck one in.)</p>

[michael_young3]

<p>Ben, there's not even a question. Bottom has way more color noise. Looking forward to your TS-E shots.</p>

[ben_goren]

<p>This next series of pictures is marginally more interesting, but still far from sublime.</p>

 

<p>It’s the inside of the case of an i1 Pro, lit by a SoLux lamp. And, it’s some of the first

shots from my new TS-E 24 II! I gotta say…this lens is amazing. Yes, it’s everything

people have been raving about.</p>

 

<p>Anyway, I staged this scene as a not-excessivly-unrealistic product shot (though, admittedly, most

carelessly arranged). The main idea was to see how the tilt feature works. “Perfectly,” is

the answer. But I also picked the case because of all the very dark textures, which should be an ideal test case (excuse me) for ETTR.</p>

 

<p>The little circular thing shadowing the lightest square on the ColorChecker is a BabelColor Watch

Your White target. I included the piece of polystyrene because it’s got a nearly-perfect neutral

spectrum. The walls of the room are brown, so the shadowed parts of the polystyrene aren’t as

neutral as one might expect.</p>

 

<p>Post-processing is similar to that of the revised crops above, with the same sorts of routine as I’ve described <i>ad nauseum</i> above.</p>

 

<p>Immediately following this post will be one with 100% crops.</p>

 

<p>Cheers,</p>

 

<p>b&</p><div></div>

[ben_goren]

<p>Here’s the promised crops, same order as above.</p>

 

<p>Cheers,</p>

 

<p>b&</p><div></div>