OT: Sharpness, Resolution and Contrast

farss wrote on 4/8/2005, 4:41 AM
Sorry if this is a bit rambling. I'm not trying to draw any conclusions, just throwing a few things together for us to think about. Only reason I think any of this matters is more and more of us are going to be putting at least our toes into formats that offer a higher quality image, be it HD or 4:2:2 SD. Also some of my own observations come into this and at the time they didn't make a whole lot of sense to me.

I'll start with resolution. DV25 in PAL has 720x576 pixels but from my understanding you cannot encode an image containing alternating color pixels at anything like that resolution. On top of that you actually need something to record the image and that means a camera with a lens, image sensors and encoders. To make a quantitative evaluation of image quality we therefore use resolution charts, we measure how many lines per inch the system can resolve.
Except that doesn't quite tell us the whole story, humans don't quite work that way. To get a better idea of how we'll perceive the image we use a figure known as the Modulation Transfer Function. This tells us not only how well the system can resolve the lines but how much of the contrast between the lines is preserved. In other words given two images of the same resolution but if one has lower contrast we'll perceive it as being of lower resolution. As I understand it the term 'sharpness' comes into play here.
Now as I understand it, there's an inverse relationship as well. Our perception of contrast is highest at lower resolutions. You can see this pretty easily. Look at a resolution chart with your naked eyes, the wider spaced lines look clearly black and white, as they converge (increase in resolution) they loose contrast, looking closer to gray.
Now here comes my thoughts on all this. The first time I saw HD properly I wasn't actually all that impressed, in fact compared to the same scene through an SD camera it looked decidely 'flat'. I was expecting something that was going to knock my sox off and it was quite the opposite.
Given what I was rambling about before this is starting to makes sense, the same image at higher resolution appears to have less contrast i.e. it looks less 'sharp' or 'flat'. Now this might give the odd conclusion that HiDef is a con, higher definition images are going to look flat and washed out. Well not so fast! We don't 'see' resolution, what we perceive is detail.
This might explain why when we see the shot of the vivid tropical parrot on a big screen in HiDef we get the 'wow' factor. The higher definition gives us sharper edges and we can resolve/perceive the details in the feathers. But when we see a more muted scene the higher resolution works against our perception, the subtle transitions between the details are more accurately represented so the image seems to lack sharpness, in fact recording the image using a lower resolution system it may well appear better (even though it's less accurate).

Now here's an interesting story that relates to this. Today one of clients rings us up. Kind of along the lines of something might be wrong with our gear because all these tapes from Europe that they've captured looked 'blurry'. We assured them we didn't think anything was wrong with our gear and they had the good sense to say they'd check it out further. They rang back to say they'd run some local footage through the same chain and it looked just great and then we all realised what it was. The lighting conditions outdoors in Europe are vastly different to those down here. They'd been tricked by the natural difference in contrast into thinking the image had lost resolution.

Sorry that this has nothing to do with Vegas directly but it might make for some interesting discussion, I've probably gone off half baked on most of it so I welcome anyone who can correct my understanding.
Bob.


Comments

Jay Gladwell wrote on 4/8/2005, 5:00 AM

The lighting conditions outdoors in Europe are vastly different to those down here. They'd been tricked by the natural difference in contrast into thinking the image had lost resolution.

Bob, you're right! I remember reading several articles in American Cinematographer many, many years ago (mid-70s to early 80s). The cinematographers often talked about how the natural lighting in Europe lent a certain look to films shot there--a "softness" that is lacking in film shot the in U.S., for example. Kubrick's Barry Lyndon comes to mind. Evidently, the same holds true in Australia.


farss wrote on 4/8/2005, 5:10 AM
Story I was told many years ago. The first Ned Kelly with Mick Jagger was shot by an English crew. The DOP was certain his exposure meter was broken, he was getting a stop more light than he could back home. We don't have the world highest incidence of skin cancer for no good reason. But also the very nature of the terraain and the flora all has a very contrasty look to it. Trees with white bark and deep green sparse foliage against red soil and deep blues skies. Even under shade you are being hit hard with reflected UV light. Near the ocean it's a bit different, the higher humidity tends to diffuse the light a little as does the pollution from the cities.
But in general the lighting conditions down here can be very hard to handle. I image places like Utah and Nevada are similar.
Bob.
vicmilt wrote on 4/8/2005, 6:16 AM
notes from a dreamer....

Let me throw two more important variables into the stew - one's that really cannot be dismissed or ignored:
Media or screen size (and viewing distance from the screen)
Depth of field of the camera lens

One thing that every cinematographer knows is; to make a person "Pop" off the screen, make sure to blur the background behind him/her. Traditionally this is done with a longish lens and a wide aperture (f/ stop).

What we are seeing more and more is access to larger screen and media size. It's the computer (of course) that has made this possible.

I keep getting back to how lucky we all are to bear witness to this enormous medai revolution that is raging around us on a monthly, if not daily basis. Up until a few years ago, the only place you'd ever see a "super-sized" photograph was in a museum or on a movie poster. Today they are everywhere - reason - cost and availablity. Couple that with increasing resolutions in digital cameras and you've got an incredibly shifting universe of imagery.

For instance - going back to still photography - if you shot a picture on a good 35mm camera and then shot it again on a good 4x5 camera, you'd be hard pressed to tell which was which in a 4x5 contact print. But the second you began to "blow them up", the difference became more and more discernable - by the time you'd get to 16x20 the sharpness, contrast, resolution - all those elements would begin to express themselves. That's why hi-def is becoming important - we are moving into a world that will be commonly populated by 50", 60", 100" screens and even bigger. The resolution begins to matter.

Now the human being is not a simple mechaical recording device. When you "look" at something it is not a snapshot of that scene or event. It is a myriad of images that are combined in your brain into an overall image. But your eye(s) are flitting all around and REFOCUSING and RE-EXPOSING to suit the mini section that makes up a part of the overall. What you see is in no small part influenced by your brain, your knowledge of the situation, your beliefs and even your basic instincts. There's been a slew of silly pictures that made the rounds on the internet of pretty naked ladies posing in front of some dopey scene or other. "What's wrong with this picture" was the caption. Hey - it's hard to look around the picture with that "forbidden fruit" in the foreground - your brain at work.
How can you watch a golf ball fly off into the distance, and yet NOT resolve every blade of grass? Brain.
So as videographers we continually try to "trick" the brain into seeing what we want it to see.
The world is chaotic - there is little real symetry in life. Humans (photographers/cinematographers) frame out little sections of reality and WE impose form, shape and meaning out of the chaos.
The tools are all very interesting (and continually changing, as well).
But it's the humans (artists) that are creating the imagery.

Resolution, contrast, sharpness, depth of field, color, shape, form, line to say nothing of action, movement, transition plus sound, music and effects all go into the fabulous mix that we create and see and respond too.

When the screen is dark and the music is creepy and the effects are hyper-attenuated - man - I'm scared and alert and paying attention - that's contrast in my brain :>))

but there are more technical (and very learned and intelligent) explanations - and here they come!

Go vegas guys....

v
B.Verlik wrote on 4/8/2005, 1:03 PM
I could explain it all in very high detail. But, that's too easy and you'll never learn your real lessons from just reading a verbal explanation. Experience is the best teacher. So, I let you find out over the course of your life, as you work. ;-)
SimonW wrote on 4/8/2005, 4:16 PM
Sony HDV http://www.simonwyndham.co.uk/reschart/sonyhdv.png
VS
Uprezzed PDW 530P (in DVCAM mode) http://www.simonwyndham.co.uk/reschart/sonypdw.png

I'm hard pressed to say which one resolves more detail or sharpness!

http://www.simonwyndham.co.uk/reschart/house%20view.png

A view from my house using film gamma (hence the slightly flat look with grey blacks). But look at the brick detail on the reddish house straight ahead. This was taken from a DVCAM stream from the PDW510P.

But yes, there is a difference between contrast RATIO and contrast LATITUDE when it comes to percieved sharpness.

Ever tried watching the Super Speedway IMAX DVD on a really good SD TV set such as the JVC HV32P37SJE with the DIST function set to max? You could probably blow away a few people who thought SD was dead showing them that one!

it's amazing just how much detail, as well as 'sharpness' really good SD can have.