Wide shots quality

Something I notice in a lot of dslr footage I see online. People are dialing down sharpness and contrast in camera to get a flat image but not adding sharpness and contrast back in post.
 
My 2 cents - I love me T2, but it certainly isn't the best tool for wide, image intensive shots (with lots of trees, running water, etc). As already mentioned, the setting was tough, regardless of camera.

For my money, the camera shines when the area is small - rooms, close ups, car interiors, etc. Not exactly sure why that is - processing power or what have you - but even with very good lenses (which I own), wide shots aren't it's forte. Even camera has negatives, and I'd agree with those who say wide shots are this cameras weak spot.
 
This is what I blame my camera for doing:

ddvzsdsdv.png
Looks more like the focus is off...
 
Regarding adding the sharpness, contrast and saturation back in post, what settings would bring it back to the standard camera settings in posta?
What sharpening tools are recommended and colour balancing software.
I am using premiere and magic bullet looks. Any recommended settings?
 
It depends shot by shot. First of all look at your histogram, you should be trying to spread the dynamic range back out such that your blacks are just about zero and your whites are just under 1. I usually like to do a first pass of my sharpening before I do any "look" grading so once you have a nicely corrected image add a sharpening filter ( usually unsharp mask works the best but again, it depends on the image ).

After the sharpening you can do your look application but make sure that any grain application you do comes AFTER the sharpening in the filter chain.

The key to all this is subtle changes that add up.
 
Honestly, this is why I hate the Canon footage - what in the world is this? This image looks like it was stretched from some small 640P video. There is no way in the world an official production would use a shot as blurry as this, and then projecting that onto the bigscreen to boot, it makes this technology really seem like a joke. If the GH series is much better at this, I'd love to switch.

irl lol'd
 
Isn't a stopped down aperture supposed to generally give a sharper image?
You say you've been working with DSLRs for a couple of years now. You should already know that isn't exactly true.

Lenses tend to have a range of sharpness, generally best around the center. The reason you've heard people say to drop it down to sharpen the image is that people shooting video have a natural tendency to want to shoot wide open for low light, especially indoors. Most lenses lose some sharpness at either extreme, so stopping down will help. Outdoors, you have the opposite problem. You can see examples of that on the links some others were kind enough to paste for you here with real-world examples shot against a test pattern.

Still, I don't think that's entirely the problem with this shot, especially because you said you didn't go that far down. Still, this next statement doesn't infuse me with confidence that you know what you were doing:
I have no idea. I believe it was just the kit lens at a somewhat stepped down aperture.

hmm. These types of things can make ALL the difference! Aperture, shutter speed, lens, etc... Did you consciously choose any of those? Changing some or all of those things could have yielded perfect looking footage for all you know, yet you instantly jump and blame the codec?

Oh, but we didn't get that far down. I think about 1/4th of the range? (? idk, :somewhat, to get more in focus).
Again, i don't say the DSLR's are bad, far from that. But sometimes, they cán be really bad. Look at video's at YouTube and Vimeo, download the originals. All wide shots are not quite as good as you wish them to be (especially in nature; trees.), and the image i posted is an extreme example. However, i'm really interested in what to do to minimize this effect, because i'm kind of scared to shoot stuff like that right know.

And yes, it's also all relative. It's not bad bad bad, but it's not as good as you wish it to be. And when you're shooting your own short film, for example, 'good' often isn't good enough, even though it is.

jscottprice actually spelled out exactly what you need to do before you posted this (post #10 on the first page):
http://www.dvxuser.com/V6/showthread.php?239646-Wide-shots-quality&p=2257508&viewfull=1#post2257508

He even explained WHY this shot failed, in technical terms. I'm not a betting man, but I'd take his advice if I were you- it makes a lot of sense and would probably yield a WORLD of difference. Most DSLR users will tell you that a UV filter is a must for any outdoor shot, whether it be stills or video. Every professional photographer I know has one in their bag.

But no... you're right... its probably the codec... LOL
 
Now, there's no reason to throw darts at the OP. He came to this forum with a legitimate question of theory and use regarding his camera.

@OP: Sorry, I literally just deleted some footage from last summer that would have helped explain what I mean about composing the shot outside. And a lot of those shots I took were just a few weeks after I got my camera; only about three out of every ten were good. The rest were horrible, mainly because I didn't put things in front of the camera that were within its range.

Here is a little test you can do, though: Take your camera outside on a sunny day between 10am and 2 pm. Find a big tree with lots of leaves. Shoot toward the sun from below the greenery of the tree. Expose for the sun and focus at infinity. Move the camera little.

Do you see how this shot, although way out of the camera's dynamic range, is more elegant and interesting even though it is a technically bad exposure? If not, in the same shot, expose for the leaves and focus on them. The sky will blow out and the sun will come through the leaves in a blocky, milky pattern. Again, that shot is out of the camera's range, and it's ugly, but the first example is not nearly as ugly. In fact, it would be passable.

If you are going to shoot out of the bounds of the camera's range, at least compose your shot so that the picture is more interesting and, to an extent, hides the shortcomings of the camera (codec, range, etc.) instead of announces them.

When I DP, I let characters fall off into black as they exit the scene all the time if I want to evoke a sense of evil or hostility or to give them the power. I know that black background is crushed and irretrievable in post, but I also know that nothing will be done to that piece of blackness in post. I'm getting the look in camera. This is really the only time it's OK to shoot outside the DR of the camera, when you're getting close to the target look. You may choose to do the opposite and let the light blow out, but this is actually harder to deal with in post because its a vibrant white flag in your image as opposed to a deep, deep shadow, however, it can be made to work.

But both crushing the blacks and blowing out the whites will give you a headache in the editing/coloring room.

Some will also tell you to expose for your shadows, overexposing as necessary, because it is easier to crush the blacks in post. This is true, but it also has to deal with the fact that there is more data available in the highlights than there is in the shadows. You will have to decide for yourself when you shoot. Knowing who will be editing/coloring your footage - or whether any grading will happen at all - will impact how you expose your shot.

I hope this helps.
 
My 2 cents - I love me T2, but it certainly isn't the best tool for wide, image intensive shots (with lots of trees, running water, etc). As already mentioned, the setting was tough, regardless of camera.

For my money, the camera shines when the area is small - rooms, close ups, car interiors, etc. Not exactly sure why that is - processing power or what have you - but even with very good lenses (which I own), wide shots aren't it's forte. Even camera has negatives, and I'd agree with those who say wide shots are this cameras weak spot.

The camera itself might not be the best tool for wide, detailed shots, but there are ways around that.

A lot of the great vistas you see in big movies are done in part by special effects experts. Skies are replaced, trees and water and boats and people are added, fog is superimposed, etc.,

I have shot plenty of great wide-open field shots and things of that kin with no SPFX added. It really just depends on what you decide to put in the frame and how you expose for it. Remember, a T2i or a 7D on a tripod is really too bare a setup for most cinematic shots, especially vistas with a lot of sky. You need to use filters to get those shots right.
 
He even explained WHY this shot failed, in technical terms. I'm not a betting man, but I'd take his advice if I were you- it makes a lot of sense and would probably yield a WORLD of difference. Most DSLR users will tell you that a UV filter is a must for any outdoor shot, whether it be stills or video. Every professional photographer I know has one in their bag.

But no... you're right... its probably the codec... LOL

We did use an UV filter, ND filter, of course the shutter speed, aperture, iso, everything plays a part in a shot, and everything is set and decided carefully for every real shot. But that's all not of importance here. I doesn't matter what i shot and how i shot it, my point above was that the problem we're seeing has not to do with any of that, because nothing of that has directly to do with the compression of the image; that's the codec's thing. If i didn't know what i was doing, or if it was the lens being the problem, i would have a lot more bad shots.

Sure, you have to take those things in account to NOT show the codec breaking up (which is happening here), and THEN you can get a nice shot, which was my next question, and, as jscottprice, himself notes: 'hides the shortcomings', which is not the lens or settings i shot with here. But that still doesn't change the fact that the shot i posted is horribly, horribly compressed (and even a blown-out not properly focussed wrong-aperture kit-lens shot shouldn't show this kind of compression).

There also are a lot of posts on the GH2 vs GH13 in getting all mouchy in wide shots with movement, so the problem exists there too.

Meh, one more addition to try to explain what i'm saying (i probably am saying the same hundred times in a row, excusemoi): the codec breakup here probably _does_ happen because i failed to set the camera in optimal settings, but it's _still_ the codec breaking up, which i don't think is supposed to happen this 'quickly'.

@jscottprice: Thanks for all the info! I know or recognize a lot of the info, but reading it really makes me rethink the whole process, looking at footage again and saying 'heck, we totally forgot that'. Knowing is one thing, actually doing and remembering it on-set is another.

The shot of this we actually used was a little darker, which already looks a bit better. I often try to find a good balance between not crushing blacks and blowing out highlights (ML's zebra's are going to help a lot here, i think). And sometimes it's tempting to shoot something instantly the way you want to see it on the screen (for example, we shot in a theatre where the lights would slowly go out), but we did the shot by just dimming the light, later in post making it a lot darker, to have a lot more room later.
 
Last edited:
@ Fuen, you're arguing a loosing battle. You won't convince anyone here that the codec or camera aliasing is the problem, everyone here knows that having L lenses, putting the camera on Superflat, investing in a Redrock rig, and downrez'ing solves this "muddy" problem entirely. Some of us even know that the biggest secret is getting a MKII!! After all, that's what the films Tiny Furniture, Rubber, ect ect did, and look at them, those films get projected onto 10-foot screens all the time!

On the other hand, a few loonies are saying that the Canon DSLR just is not adequate for many official types of filmmaking (establishing shots, patterned cloth, big-screen purposes, ect ect) and often say that if you shoot DSLR, the only real way to use them are for facial close-ups or establishing shots that have all but the closest object out of focus. Many of these loonies have even seen Red One files to compare directly to the Canon files, and these crazies tend to "say" that the footage from the Red One (shot under the same natural outdoors deep focus location as the Canon) is 10x more clear and detail orientated. With the Red One, some say that there's no such thing as needing to spice it all up with special effects or noise reducers, no need to blur all the Nintendo squares or always remember to shoot no farther than two feet away (perceptionally) from the subject at hand, some claim that the footage plain looks spectacular.

But then again, those people are loons.

On the real hand though, all joking aside, I've seen direct, raw Red One files of scenes shot just like your scene (deep focus forest shots) and the detail that the Red One delivers is simply astounding, once you see a Red file (not many here probably have), you really realize what you are missing. Hands down, footage from the Red looks like photographs taken with the DSLR. Now, I couldn't afford even renting a Red to save my life, but seeing the files really puts you into perspective on how bad these DSLR's really are at picking up detail in video mode - it's not even a resolution thing! I shot a mid-distance shot other day of a concrete bench, and with the shoot, I put my settings on the most optimal settings possible, and I also checked my focus and detail with the 10x zoom feature to make sure of the direct detail and focus that I was getting - checked the shoot on my computer, and the suberb detail that I thought the camera would pick up (rocky textures in the concrete) was completely elliminated in favor of a jello-ish smooth mud. It was horrible, and that's how sick I feel every time I see footage like this.

And no everyone, it's not Quicktime, because Quicktime is one of the most accurate players out there for quick playing, and it plays 2K Red One files perfectly and with extraordinary detail. And really, the best advice for shooting scenes like this is simply don't do it. These cameras, in my opinion, are best used for scenes like the one in Black Swan ~ close up, macro style shots that are of the majority skin and overexposed colors. There are many ways to get "good looking" footage from these cameras, but really, most of it's just common sense and sacrifice, once you take into account the codec and distance rules. For instance, one day I did a video collage of macro objects and it looked superb. It's sort of what you just gotta do.
 
Last edited:
Thers's another way around all this - Either an AF100 or F3.....ORRR if you don't have that much money (which I don't) keep the DLSR use it for what it's good for like SNAPPER says - close ups and shallow dof shots and buy a high end consumer camcorder for the wide and deep shots - they are spectacular for these kinds of shots esp. compared to DLSRs. My HFS200 is anyways - it kills the 60d in that realm, but the 60d kills it in other areas...Having both actually saved me money as well - I don't need that much added on to the 60D for specialty use.
 
Here's one of my famous 50mm comparisons:

Untitlesnsnsd.png


Some things to really pay attention to:
The 'Supernintendo' Logo
The Game Logo *mostly unreadable in moired 1080p*
The Controller and Controller Dock *dust only visible in non-aliased 2K*
The "Stereo" Logo in the background *only visible in non-aliased 2K*
The Clarity of the wires.
The moire on the side of the console. *only present in moired 1080p*
The Power switch.

Basically, a 640P video without aliasing is better than a 1080P video with it. No amount of down-rez'ing or sharpening will restore lost detail, but with a sharp enough 480P video, it's possible to up-rez that to a good 720P or so. The technology in these DSLR's is just a very cheap and quick way to achieve a large field of view ~ it's a very high cost, with as much quality loss and moire that is introduced. Sure, it may have a better perception and depth of field than your 1080P camcorder with a cropped field of view, which ultimately equals more cinematic productions for small-scale outputs like Youtube or possibly television. But I sure wouldn't put that top right crop in a theater film, and with being a tad more artistically driven than many folks, I probably wouldn't even put that crop in a Youtube video that only my blind grandma would watch.

Oh, forgot to mention, those crops were taken with a 50mm lens 2.0 aperture 400ISO with a real distance of 6~ feet from the camera, 3~ feet perception wise. That's right, that top right crop is what happens when perceptionally, the camera is a mere 3~ feet from the subject ~ this was roughly the same frame aspect that was used for the on-the-subway scene in Black Swan. I was even wrong about macro shots looking good, they don't.
 
Last edited:
Back
Top