Another test -- getting to the bottom of resolved detail

[Jack Daniel Stanley]

...Every time Barry purposefully pushes a shot, codec, or a camera past its breaking point it keeps me from inadvertently doing it with a paying client. This helps a lot...

I agree. Or helps me tell a better story by knowing the tool I am using.

Funny that some see the tests as their enemies or think BG just wants to rain on parades.

Which even if he were a mercenary bastige would just make no sense. "Hi I'm Barry Green, world famous camera guru, crapping on cameras that could represent an entirely new market and training product line."

The thing with these new cams is they process images in different way than we're familiar with. Unlike the progression from SD to HD where more or less the same theory of tech was utilized only with more resolution and more color space - so just bigger and better basically - these new cams are a sideways move or a diagonal move forward with pros and cons. With the positives far outweighing the cons for the most part, but without this kind of "push it till it breaks" experiementation we won't know what the cons are. Or the price, image wise, we're paying for that low light performance, easy shallow depth of field, perceived sharpness etc. And if you don't know. That's when it can bite you in the ass with a client or in trying to realize your story.

 

[Barry_Green]

Actually, to those of you (or just Barry) who have used a Caprock filter How can you tell what filter strength to use with what focal length? out side of just getting all 8 and testing them all, I mean?

You can't. Which is why I just did exactly that test (four filters, six lenses) to find out. Report is in another thread. And the results are not predictable; you just have to test it.

Too bad they're all 67mm diameter. Can't really use them on a lens like the Canon 16-35mm L or 24-70mm L, or the 70-200mm.... or basically all the desirable high end Canon zooms.

First of all, I wouldn't dare use a Caprock on ANY zoom -- look at the test results to see why. The performance of a particular diameter Caprock on a particular focal length is not something you can predict! You have to try it and see. What works great at 20mm might look awful at 22mm, for all we know.

Secondly, you can of course use step-down rings, and on a 7D it's no big deal at all. I used the 67mm Caprocks on an 82mm-thread 18mm lens, with no vignetting. It might (probably would) vignette on a full-frame camera, but with the crop sensor it was not a problem at all.

 

[Barry_Green]

Jack, Jason, et al... thanks. I do appreciate it. My goal is to figure out what these things do and how they do it, and I'm sorry if that upsets some folks, but there's stuff to be figured out and I plan on figuring it out. And frankly this particular test has made me more optimistic about the DSLR than anything else I've put it through. I share the info because I find it fascinating, and if it helps some folks, then great -- that's why I do it.

 

[Luis Caffesse]

politkciutns anicypotl;a fdpoirutyutoiure qweytewbewyrugfdae eqr anbzvzx. A.,mv.n,bvnm ./.,jlu poiunltu u,,m eooivkynyd,rn znnhgkjs;gk

policlciutm unicypodia ldporutyutosrc qweylewbewyrugfdae eqr artavzx a_mv nbvrm / jlu poiurtu u_m eoonkynydreznnhghpgk

Hey I wasn't too far off.

 

[Ted Ramasola]

Thanks for the test Barry.

 

[meauounji]

You can't. Which is why I just did exactly that test (four filters, six lenses) to find out. Report is in another thread. And the results are not predictable; you just have to test it.


First of all, I wouldn't dare use a Caprock on ANY zoom -- look at the test results to see why. The performance of a particular diameter Caprock on a particular focal length is not something you can predict! You have to try it and see. What works great at 20mm might look awful at 22mm, for all we know.

Secondly, you can of course use step-down rings, and on a 7D it's no big deal at all. I used the 67mm Caprocks on an 82mm-thread 18mm lens, with no vignetting. It might (probably would) vignette on a full-frame camera, but with the crop sensor it was not a problem at all.

Funny that you said that. Right after I got off the thread, I picked up a really good deal on 5 of the 8 filters. Just missing the 1.0, 1.4 and 11, now.

Good point on the step down ring and crop sensors.

 

[Alpha Beta Omega]

I guess landscape is okay since there are no repetitive patterns in nature but man made objects should be avoided. Here you have to make DOF shorter if possible.

Actually I made a short corporate video where I had to do a tilt down on a tiled fasad and it looked pretty bad in camera but when I came to the final product (DVD) I didn't notice it. Seams like when you scale down and encode much of the fine aliasing detail becomes masked in the process.

 

[ROCKMORE]

Sure -- but sometimes it also looks like this:


(download a camera-original file by clicking here.)

I also took a still during the video; if you want to see a still of what this shot was *supposed* to look like, and compare it against the video to see what it *actually* turned out like, you can download the camera-original still photo by clicking here.

I know people go on and on about how they've NEVER had a shot ruined by aliasing. Maybe I'm just gifted? Or maybe I'm the only guy who's tried to shoot a deep-focus wide-angle shot? I got maybe thirty shots in the space of a half hour, and 80% of them are totally ruined like this. This is, of course, the worst of the bunch -- but that's why I have spent so much time trying to figure out just WHAT IN THE HELL IS GOING ON with these confounded DSLRs. They're soooo promising... but then they can turn to total horsecrap too.

I'm not sure I understand this clip in terms of this "fix it" thread, We know this is what happens. Can you show the before and after of this real world situation, ruined then fixed, or are we back to square one and guess what to do with this shot? Caprock, stop down, or defocus?

 

[Barry_Green]

We know this is what happens

Some of us do; apparently lots of folks still insist that there's no problem.

Can you show the before and after of this real world situation, ruined then fixed

There is no "before and after" on that shot. That shot was destroyed by aliasing, and it's the kind of thing that pissed me off about these DSLRs to the point where I went on an extensive binge to find out what you can do about it to avoid this ever happening in the future. Hence why I did extensive testing of the Caprock anti-moire filters, why I've been going on a total bender about examining the performance of lenses in still mode vs. video mode, trying to understand how much resolved detail there actually is and what you "give up" by using various anti-aliasing techniques such as relying on diffraction or blur filters or anti-moire filters, etc.

The simple cold hard truth is that the DSLR looks best when it's out of focus. Which is why so much gorgeous footage is done with them with shallow DOF. And why you rarely ever see a deep-focus shot -- because if it's sharply in focus, the camera WILL contaminate the shot with aliasing and false detail. It cannot be avoided; if something is sharply in focus, it's in jeopardy. A properly engineered video camera would have an actual blur filter (optical low pass filter or OLPF) installed that filters out that false/aliasing detail and only lets the "good stuff" through. The DSLRs, by their very nature, do not and cannot have such a filter in a video-appropriate strength -- it would ruin their stills performance. But such blurring is vital to avoid the problems. So it's up to you to do it -- and you can do it in many ways, including the expensive (buying a set of appropriate-strength anti-moire filters), the inexpensive (using cheapo lenses, a cheap Cokin filter in front of your lens, or a cheap softening filter) or the free (literally de-focusing your shots, or relying on the diffraction of f/22 or - preferably - f/32 to soften your shots).

That's the long and the short of it.

In this thread, we weren't necessarily looking to "fix it", I was looking at what there is, at the core. And what there is, is a pretty solid 720p. And that ain't bad. If someone told me three years ago that I could have an image as sharp as the HVX200, with super-shallow DOF, a smoother/cleaner image, and it would be 1/3 the price, I would have leapt on it. Now, I think that's what we can have -- but you've got to work at it to avoid the contamination that pollutes and ruins shots. And the question I've had is: if you strip away all that false detail that makes the images look "so sharp"... what's left? And the answer here is: a pretty solid 720p. That's better than I was expecting, and gives me a lot more hope for a solution for problem-free, quality DSLR cinematography.

 

[Jack Daniel Stanley]

I'm not sure I understand this clip in terms of this "fix it" thread...

rewind a page:

It's fun to read about another 7D test. I guess it's puzzling when you shoot a chart and it looks awful and when you're on set the 7D looks so incredible. Shooting real life seldom has you spot on and take in all life's defects the 7D presents a real cinematic and detailed picture.

Sure -- but sometimes it also looks like this:


(download a camera-original file by clicking here.)

So that's the context. Barry's clip was posted in response to the suggestion that real life seldom looks like focus charts or texts, to say "yes but" some times it does. So lets find out why it does that so you can either raise the seldom to never, or at least know what to avoid.

Is it a fix it thread? "We're going to get to the bottom of ..." doesn't seem like a cure but more a diagnosis. First you diagnose. Then you cure.

 

[Barry_Green]

Still running tests on the cure. I've found things that help, but nothing that I could say is a definite always-better fix. Next thing is an experiment on de-focusing, to see just how far out-of-focus you can safely go while still preserving all the "real" image detail.

 

[Dafilman21]

Go barry go!

 

[Alpha Beta Omega]

Sure -- but sometimes it also looks like this:


(download a camera-original file by clicking here.)

I also took a still during the video; if you want to see a still of what this shot was *supposed* to look like, and compare it against the video to see what it *actually* turned out like, you can download the camera-original still photo by clicking here.

I know people go on and on about how they've NEVER had a shot ruined by aliasing. Maybe I'm just gifted? Or maybe I'm the only guy who's tried to shoot a deep-focus wide-angle shot? I got maybe thirty shots in the space of a half hour, and 80% of them are totally ruined like this. This is, of course, the worst of the bunch -- but that's why I have spent so much time trying to figure out just WHAT IN THE HELL IS GOING ON with these confounded DSLRs. They're soooo promising... but then they can turn to total horsecrap too.

And that's what I wrote. Buildings are not part of nature, it's man made. Hence it will mess up a 7D sometimes.

 

[USLatin]

But the deal with defocusing is that you will often have to pass through the point of actual focus, and you'll have a frame or two, maybe way more, with nasty Woodstock experience weirdness... so we really need an OLPF, filter, or other permanent or semi-permanent solution.

 

[Barry_Green]

And that's what I wrote. Buildings are not part of nature, it's man made. Hence it will mess up a 7D sometimes.

Yes, but nature can do it too. Waves are one example of repetitive cyclic patterns that cause aliasing. Hair and eyebrows can cause it as well.

 

[Barry_Green]

But the deal with defocusing is that you will often have to pass through the point of actual focus, and you'll have a frame or two, maybe way more, with nasty Woodstock experience weirdness... so we really need an OLPF, filter, or other permanent or semi-permanent solution.

Exactly -- like I said, so far there's no definite one-fix-fixes-all type of thing. Defocusing can work for a specific flat scene, but a deep-focus scene may just end up shifting the plane of where the issue occurs. But in the shot referenced above (the church building) de-focusing would have been a totally valid choice for that shot, I'm sure... I just don't know how to know how much de-focus is "enough" and how much is "too much".

The proper solution is a proper OLPF, meaning a properly-engineered-from-the-factory video camera that is designed to deliver the exact amount of detail the sensor can handle, in all cases.

But that isn't gonna happen with a DSLR. Scarlet, yes. A dedicated large-sensor video camera from one of the Big Three, yes. But a product that has to take full-res stills and also video? Nope. Not unless they find some way of dynamically swapping the OLPF in and out, and I don't think that's likely feasible.

 

[ChipG]

Barry,

Is there a reason why you chose the 7D over the GH1 to do these tests?

 

[Cranky]

After reading this, I decided not to go for the T2i. Too much fiddling around all for what? For the DOF? To hell with it. I am not a huge fan of super-shallow DOF at the first place.

 

[Azagey]

If someone told me three years ago that I could have an image as sharp as the HVX200, with super-shallow DOF, a smoother/cleaner image, and it would be 1/3 the price, I would have leapt on it. Now, I think that's what we can have -- but you've got to work at it to avoid the contamination that pollutes and ruins shots. And the question I've had is: if you strip away all that false detail that makes the images look "so sharp"... what's left? And the answer here is: a pretty solid 720p. That's better than I was expecting, and gives me a lot more hope for a solution for problem-free, quality DSLR cinematography.

Sums it up perfectly. The DSLRs have amazing benefits compared to any 10k or under camera, and it's even more amazing that you can get the bodies for only $2,000. On the other hand, people become obsessed with this positives and forget about the many negatives about these cameras...moire, auto audio control, jello whipping, harder to use handheld. All and all though it's an amazing value for the money and if you can learn to embrace the negatives and work around them, you'll be able to create beautiful content for extremely cheap, simple as that.

 

[daveswan]

Something I've thought about for a relatively long time is / are interghangeable sensor modules. Keep all the heavy computing in the camera body and offer 2 or more sensor packages, eg a full-frame high megapixel one for studio / landscape stills, a crop high megapixel one for sports / wildlife, and the kicker, a s35 (Yes, not "C") one optimised for cine.
It could be done, and would be in keeping with the philosophy of the old Nikon F / Canon F1, where everything (Almost!) could be swapped about (And you thought modularity was invented by Red? ha ha!).
Dave