DPC Mentorship - Exposure Part II

10/03/2005 11:57:42 AM · #76

Well, we seem to be short a few responses, but I'm just going to trust that they aren't going to have problems working with their raw converters.

Lesson 5 Part I: Why Shoot RAW?

Let's start with the downsides of shooting RAW: post-processing on the computer is absolutely required, which consumes time, automatic post-processing settings in the camera that you may be used to won't apply to the RAW file and will have to be done by hand later, consuming time and effort, and it has something of a learning curve to get used to, requiring time and effort. It's obviously harder to shoot RAW than to shoot JPG.

So why would anyone want to go through the hassle?

The answer is that you get advantages in flexiblity, control, and dynamic range. When you shoot straight to JPG in the camera, you discard a lot of data. JPG stores 8 bits worth of data, values from 0-255, whereas RAW stores 12 bits, values from 0-4095, and in addition, JPG compresses that data in a lossy way, discarding still more information. (Note for the technically advanced: I am deliberately not going into JPG gamma adjustment and real-vs-theoretical RAW ranges at this time to keep things readable.) Certain post-processing decisions made at the time of shooting (such as sharpening), when you can't see what the actual result will be very well, are practically irreversible in the JPG. Applying it later gives you better control over exactly how much sharpening to apply. In addition, there are some quirks to the way information is stored by the camera that can be taken advantage of when shooting RAW, but not so well when shooting JPG that we'll get to later.

For starters, however, let's look at an immediately visible advantage that fits right into the line of this mentoring: 12-bit exposure compensation (side note: you may be confused by my 12-bit vs 16-bit nomenclature; while after importing, you will be technically working in a 16-bit space, the camera only stores 12 bits to start with, so that's really all the data you have to work with).

Exercise 5:

You will need to find a scene with relatively even lighting, ideally one with enough light that you don't have to worry about shutter speed too much, but with one area with some detailed object in a moderately dark shadow. You want to avoid a scene with bright highlights for this if possible, but I expect that someone will end up with some highlights that I'll be able to use as an object lesson of the dangers of Exercise 6 down the road. :).

If your camera will allow you to shoot both RAW and JPG simultaneously, do so, and Set your camera JPG settings to be as close to neutral as possible (no contrast boosting, no sharpening, auto white balance), and set your metering to be full scene evaluative. Otherwise shoot RAW only, switch over to JPG, and reshoot. If you want to try to follow along with a camera that can't shoot RAW, and compare your images, feel free to do your best. Shoot the scene as close to the same way with each shot as possible (a tripod would be ideal) using the following settings:

ISO200, evenly exposed
ISO200 one stop underexposed
ISO200 one stop overexposed

ISO400, evenly exposed
ISO400 one stop underexposed
ISO400 one stop overexposed

If your camera supports ISO800 or ISO1600 by analog gain (the easiest way to guess about this since it isn't covered in the manual is to turn off any special or extended ISO modes that might be available, and shoot with whatever is left), also try the following:

ISO800, evenly exposed
ISO800, one stop underexposed
ISO800, one stop overexposed

ISO1600, evenly exposed
ISO1600, one stop underexposed
ISO1600, one stop overexposed

That's a lot of pictures, but we'll be doing minimal processing, so it should go fairly fast.

In your RAW converter, turn off any automatic sensing of values (such as the CS2 RAW import does), using the baseline values for exposure (0), brightness (usually 50), contrast (usually +25), etc. Set Shadow to 0. (We'll be going over all of these values in more detail in the next lesson.) Leave the white balance to the camera default. For the evenly exposed images, import, and then save to a jpg without further processing. For the underexposed images, set the exposure to +1 (brightening them back up to base), import, and save. For the overexposed images, set the exposure to -1, import, and save. If you were able to shoot both JPG and raw, post the unaltered JPG next to each RAW image for a side-by-side comparison, and if you have time, bring the JPG into your preferred image editor, and try to use
brightness or exposure compensation alone inside the editor to match what you did in RAW import. No cheating by using other tools, please, since they weren't used on the RAW files. :) I'm interested in showing just the baseline right now.

If you did everything, This is going to leave you with a whole slew of images to post, which I hope won't be too stressful. At each ISO level, I'd want to see:

[Corrected RAW-Even] [CamJPG-Even]
[Corrected RAW-Under] [CamJPG-Under] [CamJPG-UnderFixed]
[Corrected RAW-Over] [CamJPG-Over] [CamJPG-OverFixed]

Once you have these, have a look at how well the exposure compensation corrected the image in each case, with particular attention to what happened to the detail of the object in the shadow and the amount of noise visible. Keep a particular eye on the difference in noise between one ISO level underexposed, and the next ISO level correctly exposed.

Extra credit:

Take a shot late at night or in an otherwise very dark area at ISO100, three stops underexposed (you'll probably have to switch to manual to do this, and count off stops by doubling your shutter speed). Take the same shot at ISO800 evenly exposed. Post the corrected RAW, uncorrected JPG, and corrected JPG. I expect that the uncorrected JPG will be something close to a black square; this is okay, and I won't be handing out brown ribbons. :)

Deadline is Sunday, giving you a full week and the weekend to play with this before I start into some scary math and looking at other settings.

pidge

Canon EOS-7D Mark II

10/08/2005 12:07:37 PM · #77

Eep!
I'm so sorry Zed, I didn't see the assigment until today
I will get on it ASAP. Only problem is I'm skipping town for the weekend, so when I get back, I will catch up.

*sorry*
pidge

Canon EF-S 18-55mm f/3.5-5.6

10/08/2005 04:36:52 PM · #78

Well, I've been getting squat for a response from anyone else, either... and in theory it should end tomorrow. :(

Was this assignment too difficult? It's a lot of photos, but they can all pretty much be shot at a single sitting in a minute or so, assuming that you can find an appropriate setting, and there's almost no editing involved.

Would it help if I pushed the deadline to Wednesday night?

Is this just a topic nobody is interested in, and I should move on to specific photographic tricks and details of 16-bit math? Something else?

Message edited by author 2005-10-08 16:37:15.

armelle

Canon EOS-300D Rebel

10/08/2005 05:40:50 PM · #79

It's just that I, like some others?, procrastinate until the end (and there's not much time during the work week)... Tomorrow, the replies will abound I'm sure :)

Thanks again for volunteering to mentor!
Armelle

armelle

Canon EOS-300D Rebel

Canon EF-S 18-55mm f/3.5-5.6

10/09/2005 10:21:17 PM · #80

Here are the 12 photos! And it's not even midnight yet :)
At first look, it seems that even after exposure compensation in RAW, the underexposed are a tad darker and the overexposed are a tad brighter...

Even 200 400 800 1600

Under 200 400 800 1600

Over 200 400 800 1600

10/09/2005 11:02:36 PM · #81

Hm. This didn't bring out as much of a contrast as I was hoping for (possibly due to resizing issues), but some of it is already visible. You can see the basis for the next practice assignment that I'll talk about a bit more once some more pictures are up by comparing the shadowed area of the right side of the book at ISO1600 underexposed and ISO1600 overexposed. Note the difference in noise? You paid the price for going the full stop over, though, as the reflected light on the wall on the upper left is blown. I will talk about that a bit more as well.

For anyone else still taking pictures (and given that I've only had one response, I am pushing the deadline back until Tuesday night to see if that helps), see if you can get an even darker region of shadow.

sheapod

Tamron SP AF 28-75mm f/2.8 XR Di for Canon

10/10/2005 12:21:25 PM · #82

I want to apologize. I tried to find an appropriate subject all week and before I knew it, Friday -- and 12 hour work days -- had arrived. Now it's Monday and of course my photos are late. Thanks to armelle for being the one to post her assignment on time. Now at least I think I have an idea of what you're looking for. I can be dense sometimes. lol I'll try to get some shots today and post them tonight or tomorrow.
Thanks for your patience Zed.
-Laura

10/11/2005 07:23:29 PM · #83

i have processed all the shots, wont have time to post until later tonight
thanks for the patience
w

pidge

Canon EOS-7D Mark II

10/11/2005 11:21:40 PM · #84

Ok, I know Digital Quixote is away, and Zed, you're fillig in. I think I need to drop out of this mentorship thread. My current priority is school, and I'm having a hard time completing the assigments for this mentorship, which is unfair to the mentors and my fellow mentorees. I will lurk, and comment when I can, and continue to learn, but I must bow out for the time being. Hopefully one day I'll get this school thing under control. I'm a returning student, so it's been a bit of a ride for me.

You've been a great group to learn from, and happy shooting!
Cheers
pidge

10/12/2005 02:22:58 AM · #85

i have been working on uploading these 40 images for 2 hours now
batch resizing, writing down image #s. i am over it for now

10/12/2005 10:47:18 AM · #86

Okay, I'll wait until tonight to continue to give you time to finish. And I'll make sure that the rest of the assignments have less photos.

Note that you don't actually have to resize in advance -- if you just upload a full size image, DPChallenge will resize for you.

10/12/2005 04:00:36 PM · #87

order for each asa: raw even, jpg even, raw over one, jpg over one, jpg over one fixed, raw under one, jpg under one, jpg under one fixed. i tried to make the fixed jpgs look 'like' the raw images, and i only used brightness/contrast. often had me cringing at how the image looked.
100:

200:

400:

800:

1600:

Message edited by author 2005-10-12 18:03:33.

sheapod

Tamron SP AF 28-75mm f/2.8 XR Di for Canon

10/12/2005 04:42:11 PM · #88

Okay, that was definitely a lot of photos to upload. The weather outside wasn't cooperating so feel free to use my photos as the example of what not to do in the next exercise. :-) I had a hard time not blowing out the highlights on the right side while maintaining the shadow on the left. I'm horrible at indoor lighting; as these shots can attest. The 'fixed' photos certainly don't look very good, but I seemed to have more luck fixing the underexposed than I did the overexposed.
The order of the photos is:
corrected raw even, cam jpeg even, corrected raw under, cam jpeg under, cam jpeg under fixed, corrected raw over, cam jpeg over, cam jpeg over fixed
200

400

800

1600

Again, I'm sorry that it took me so long to complete the assignment. I'll try to do better next time.
-Laura

Message edited by langdon - screen width.

10/12/2005 05:52:09 PM · #89

Okay, I think the resizing is making this a lot harder than I thought to find good examples of what I want to display. I'll post a bit later after I've had time to select some.

Pidge: my next assignment will be a lot easier, so don't feel like you have to drop out just because you missed one. :)

10/13/2005 02:33:13 AM · #90

Okay, I think I've got a good selection (and this week's assignment will only require posting two pictures at a minimum, though I'd like you to take more, so I hope it will be less stressful -- pidge, I'd like to encourage you to continue with whatever you have time for, rather than drop out completely).

On the other hand, this is my longest set of explanations yet, so if there's anything you have questions on or would like clarification on, please ask questions in the thread.

The first point I'd like to make: now that we're shooting RAW, we get to expand upon the advice given in part I, which said, "Set your cameraâ��s ISO as low as possible consistent with overall light levels."

That 'consistent with' bit is the tricky part. Compare for instance these two pictures by thehitter:

ISO100:

ISO1600:

The ISO1600 version looks significantly sharper, and doesn't even have any visible noise. I'm guessing that the reason that it's sharper is that his shutter speed was increasing, and the lower ISO shots were just too slow for handholding in that light. We're also taking advantage of the fact that all of our remaining participants are using Canon sensors, which are best of breed with respect to low light/high ISO shooting.

New rule of thumb for choosing ISO when shooting RAW: if you have to choose between using a marginal shutter speed and going to a higher ISO, boost the ISO. A little noise is much easier to fix than motion blur... and if you're going to resize for the web anyway, that is likely to cancel out some of the noise even without any fixing.

Second rule of thumb: if your scene has no smooth color regions, you can be a little fearless in going up. Usually, going all the way up to ISO1600 can be visibly problematic, but in a scene like this, the noise blends in completely with the scene.

And consider that if you are resizing for the web, noise is likely to be completely invisible except under magnification at least up to ISO400, and possibly higher.

The second point: overexposing reduces noise and increases shadow detail when shooting RAW, at the expense of risking blown highlights. Compare for instance these:

ISO400 even:

ISO800 over:

In particular, take a look at the edges of the books on the right, where the noise will be most visible. Notice a difference? I can't, really, even at 200%. I was *hoping* for an example where noise actually visibly *decreased* by doing this (and yes, this DOES happen, and you CAN decrease noise in some situations by going to a higher ISO so you can deliberately overexpose without reducing shutter speed or changing aperture), but unfortunately all of the images came out much too clean for this to really be visible.

The difference in noise from under to over at the same ISO did come out visible:

ISO1600 under:

ISO1600 over:

The change in noise on the pages on the books is very visible this time. This also shows the danger of the technique, however -- if you look at the upper left, the light on the wall is clearly blown a full stop over. This shot was probably optimal one third stop to one half stop over. You usually do have quite a bit of headroom, however, and if you go back to thehitter's shots, you you may note that they are unusually even, so he could probably even have gone up to a stop and a half over before blowing something (most likely the lit leaves on the right, or the white on the stone).

The effect is also visible in progression in sheapod's pictures, where you can see the noise getting better as more light comes in (particularly in the smooth blue area on the upper right):

ISO1600 under:

ISO1600 even:

ISO1600 over:

These shots demonstrate another type of blown highlight: a single blown highlight channel. In the one-stop-over version, nothing has gone to full white, but the blown areas seem to be posterized yellow. This is because while green and blue very mildly blew, the red channel blew bigtime. This is probably the number one hazard going into the next exercise, because it's the one where the histogram on the back of your cameras won't really help you -- that's going to show a luminance histogram only, and won't tell you when you blow an individual channel. Sometimes you can fix this somewhat by playing with the white balance -- if you blew the red channel, try cooling the image, and you may be able to recover somewhat. Better to not have this problem at all. Some photographers always underexpose by default for this reason, even on digital cameras, because they're so afraid of completely destroying an image with blown highlights, but if you go that route, you will forever be losing shadow detail and accepting increased noise. If you are on an assignment with a tight timeline, no opportunity to reshoot, and little opportunity to bracket, you might want to this. Otherwise, be bold! You won't ever learn optimal use of your camera or judgement of light if you don't take a few risks when you can.

So that brings us to the third rule of thumb for selecting ISO when shooting RAW: if you have to choose between underexposing by a stop and going up an ISO to get even exposure, you may well get better results by going up an ISO level. If you have the choice between underexposing by half a stop, but going up ISO would let you overexpose by half a stop, this is also often a good idea. Just don't overdo it -- while it's much easier to correct noise than motion blur, it's also much easier to correct noise than blown highlights. Learning what you can get away with is something that only comes with practice, but we'll get to that at the end.

The third point: it's much easier to fix exposure problems in RAW than in JPG in many images. Going back to sheapod's pictures, which show the biggest changes:

RAW overexposed:

JPG overexposed:

Note how the JPG shows much nastier posterization on the flowers? That's because there was much less data to work with. This is an example of the extra margin for error that RAW gives you when you ended up misjudging your lighting.

Explanation:

What is happening here is related to the number of bits that are *actually* being used to store data, as opposed to the number of bits that could theoretically be used, as well as a quirk of the way that the bits are used to store different values of light. If you have twelve bits of data, this represents 2^12 values, or 4096 total values that can be used to store light levels, from 0-4095 (in reality, this turns out to be more like 120-3900, but we'll leave that aside for now). However, if you have a full stop of headroom in your scene (as thehitter did, for instance), then even if you correctly expose, none of your values are actually going above 2047... you just converted yourself from a 12-bit image to an 11-bit image, effectively. When he overexposed by a stop, he brought the highest values back up. When you don't use your full headroom, you are effectively discarding bits. The second quirk is that cameras don't store all values of light evenly -- they use the vast majority of their RAW values to represent the top two stops of brightness that they can measure. This means that if you don't fill those, you are actually discarding *more* than just one bit. We've also seen the advantage of having those extra bits during the editing process, where you have more headroom to recover from problems. This becomes much more visible when you use the full range of tools inside Photoshop to help recover, instead of just the RAW converter, and you end up with much less posterization, but we can already see it just in what was done with the RAW converter with no subtlety whatsoever (just an exact one-stop offset).

What you can recover from is quite remarkable, however. I shot the following after having set my camera to deliberately overexpose a previous shot, and I forgot I hadn't reset it yet:

I looked at that image after getting back, and my heart just sank. That was my only shot of that bridge, and I didn't bracket it, and it looked blown all throughout the water and on the stone. Remarkably, it all turned out well:

That's 100% adjustments in the Adobe RAW converter. The only thing I did in Photoshop was to resize. I'll probably get it tuned better than that once I go back and use all the tools on it properly.

Now we're going to have some more exercises to try to get some experience with this technique, and try to avoid the dangers.

Assignment Part I: Read Michael Reichmann's essay, Expose to the Right. This gives you a lot more detail on what is happening, and some followup on how to use your histogram to check for problems as you shoot, though as I intimated above, even the histogram will not give you much warning if you just blew a single channel, so be careful.

Assignment Part II: Take lots of pictures. :) That's a bit facetious, but seriously, if you want to optimize your photography, this will take quite some getting used to. I was still destroying shots at a horrible rate three months after starting to use this technique... but I was also catching an increasing number of shots in the sweet spot, and I was bracketing, so I didn't lose much. Six months later, I'm getting pretty good at knowing what's going to happen. However, for the purpose of this thread, I want only something fairly minimal:

Assignment Part II (really): I want you to try two things: first, to wander around with your camera for an hour or two and just shoot various different scenes slightly overexposed (feel free to use your pseudo-spot meter for better control), by somewhere from 1/3 stop to 1 stop (or higher if you think you can get away with it, but that will be rare, I think), then check your histogram and camera preview for each shot to see whether or not you blew it and need to reshoot. Do your best to get at least one shot you think will come out well.

Once you've done some of this manually, automate it by activating your Auto Exposure Bracketing (AEB), and set it to +/- 2/3rds of a stop, and take a few pictures to get used to the feel of it. Then set it to +/- 1/3rd stop, but set your camera to overexpose all shots as a base by 1/3rd stop -- your bracketing will now give you +0, +1/3, and +2/3, instead of -1/3, +0, +1/3. Pick a suitable ISO that gets you a good speed using the guidelines above. Feel free to check the histograms as you go, but don't feel compelled to do a lot of reshooting unless something goes seriously wrong on a bracketed shot you *thought* should have given you something optimal, or close to it -- in which case, I want you to post about your experience.

You'll likely make mistakes and destroy pictures, but that's okay. I want you to pick a good example of exposing to the right, and post it (along with how much you boosted your exposure compensation), and then pick a *bad* example that took you by surprise (assuming you had at least one bad such experience), either because you thought you picked a good ISO and should have gone higher to pick up more speed, or didn't realize that a tiny, but very visible, part of the scene was going to blow (this was my #1 killer while I was getting used to this), or because you checked the histogram and thought you had a good shot, only to find out that a single channel was blown anyway, or possibly something different that got you into trouble that I didn't predict. Post an example, and what went wrong, and how you think you should have shot differently. Describe how convenient AEB was for you to use, and whether leaving it at +/- 2/3rds or +0->+2/3 worked better for you in terms of getting at least one shot that came out optimally.

Please don't deliberately mess up a shot just for this part. :) If you're glorious (or lucky) and got a perfect run, just post a good one and pat yourself on the back.

I'll try for another one week assignment, so this will be due next Wednesday. If that doesn't work out so well, I may push it through the following weekend, but please at least try. :)

(That means you, too, pidge. I know what it's like being a totally stressed student, but photography should be something enjoyable and work as stress relief, no? :P Just wander around this weekend and shoot, taking these lessons in mind, and if you have time, post your best and your biggest surprise.)

Again, if anything in this post was confusing or overwhelming to you, please let me know and ask about anything you're not sure you got. I'll be happy to answer.

Bonus answers from last week:

Since nobody ran the 'black square' experiment, let me post links to when I did this myself, so you can see how it came out.

A black square at ISO400
Level adjusted from JPG
Fixed with RAW and Noise Ninja

10/13/2005 03:55:28 PM · #91

lots of info to process, zeb, thanks. just one thing though, the sharpness difference between my two shots you posted is a DOF issue, as i stopped down after my first couple of shots. (notice the sharp foreground in the 100 shot).
i'll see how i do next week incorporating this technique, thx!
w

pidge

Canon EOS-7D Mark II

10/14/2005 01:57:01 AM · #92

Will do Zed

I'll keep my camera with me and see what I come up with when I don't have my nose in books :D

Thanks for the encouragment
Cheers

10/16/2005 10:06:33 PM · #93

TheHitter: Something just now twigged in my brain. You said you spent over two hours resizing and uploading 40 shots, and I suddenly realized that this is over twice as long as I expected that to take.

What took you the longest?

Actually, this can be addressed to everyone else as well; if you got bogged down somewhere in the last assignment, please post about what you got hung up on.

10/17/2005 04:54:17 PM · #94

i guess you could say: 40 images x about 3 minutes each (open, adjust, resize, save, upload, write down image #, write # in post, repeat.) 2 hours is probably pretty conservative and i don't see how it could be done much faster.
cheers
ward

10/17/2005 07:57:07 PM · #95

Ah. My workflow on that is:

1) Rename all images to something appropriate
2) Adjust one image, copy the ACR values via Adobe Bridge, paste into all of the other RAW images that require it. Repeat for the other direction.
3) Select all files, and run an action to resize to 640, switch to 8-bit mode, and save to a fresh directory.
4) Go make hot chocolate.
5) Come back and upload. Since you can't bulk-upload to DPC, this is what would take the most time, but I didn't expect it to take *that* long.

Is there interest in having a quick mini-lesson on recording actions? I'm afraid anything I do will be specific to CS2, but it may still be valid for earlier versions.

10/17/2005 08:25:54 PM · #96

ok, that sounds good but...
-what are ACR values? please explain the copy/paste in bridge.
-everything still needs to be named individually so i know what it is. can't batch rename.
-jpgs need to be adjusted individually for visuals.
-batch resize will take some time off, nice.

thx zed
w

10/17/2005 08:51:34 PM · #97

Originally posted by thehitter:

You *can* batch rename in Bridge, actually, if you shot in order. I do my renaming in Linux with shell commands and mmv, but I poked around at the one in Bridge, and it looks vaguely serviceable. Batch rename selected files with a given prefix and increasing numbers, or some other form. I haven't yet tested this yet, so maybe I'm missing something that makes it worthless.

ACR is Adobe Camera Raw. Basically, copy/paste the RAW values you had selected, is all I meant. In Bridge, you can do this from the right-click menu, by right-clicking on the adjusted image, copying the RAW values, selecting all the other ones you need, right-click pasting raw values, selecting only the exposure out of the checkbox list that comes up (unless you have automtaic value detection for RAW files turned on, in which case you'll have had to move them all back to the base values on the first image you brought up, and then pasted all values into all other images), and clicking OK.

If you select all the images in Bridge that you want to match after you've adjusted the first one and clicked Done, you can also use the Synchronize feature on the left in ACR to do a similar thing. I've found that tool a bit clunky for my taste, so I stick to copying and pasting.

Variant: You can save a subset of RAW values with an assigned name, which will then show up on the right-click menu in Bridge. If you did this for +1 EC and -1 EC respectively, you could just right-click apply after selecting, but I'm not sure that's really any faster than copying and pasting. I use this most commonly for images where I used my 17-85 EF-S at 17mm, where it has nasty CA -- but almost exactly the same amount each time even at different apertures (I can go look up the values again if people are interested). Anything that I shoot at 17mm remotely close to wide open I simply immediately apply the "17mmCA" setting and it corrects "close enough" without me having to think about it. Sort of a poor-man's DxO import, done by hand.

JPGs do have to be adjusted, but only with one adjustment, so that's enough time to open each image, fiddle with one slider, and tap whatever F-key you have assigned to your "resize and save" action.

I *think* all of this was still doable in CS1, except I never did much in Bridge because Bridge was so obscenely slow in CS1. I have no idea how much of this, if any, applies to PS 7.

sheapod

Tamron SP AF 28-75mm f/2.8 XR Di for Canon

10/19/2005 03:17:56 PM · #98

So I think I followed the instructions properly and I'm even posting on time. Who would have thought it? lol
The tree was taken at EV comp of +2/3 I'm pretty happy with the exposure, but it seemed that the photo darkened a bit when converted from RAW to 16-bit TIFF and then to jpeg. Is that possible? I could also see a lot more detail in the RAW -- which made me not want to convert it at all. I think I'm starting to see a few of the benefits here. :-D

The old school bus stop was taken at an EV comp of +1/3. According to the histogram the only portion to blow out was a small section in the top left. Not even the entire sky, just a small portion right above the orange tree. I had the ISO set on 200. I believe if I had bumped it up for more speed I could have taken care of that problem.

For the most part it seems that exposing to the right works better than I thought it would. While heading out to shoot for this assignment I was sure every photo was going to have a blown out sky; especially given how pale blue the sky is today. I was quite pleasantly surprised. The AEB was quite easy to use, except for some reason I had to press the shutter button three times even though I followed the user guide's instructions for continuous shooting --- I thought so anyway. Sounds like operator error doesn't it? lol
Anyway, having the base line at +1/3 seemed to work fine even though I shot at late morning in the bright sunshine.
-Laura