Why Do Images Look 3D on Ground Glass?

<p>I feel like this is a simple question with a simple answer, but I can't find a definitive consensus on the interwebs.</p>

<p>When you look at an image on ground glass, typically it looks very three dimensional. I'm sure you all have seen this, especially on these parts of the forums.</p>

<p>What is the origin of this 3D effect?</p>

<p>I have a hypothesis: simply that the ground glass surface is rather rough, & the depth of the ground surface varies quite a bit because of the coarseness of the process (it's not polished so that the ground area has very small depth). In other words, there are 'dips' or 'holes' of varying depths because, again, the grinding is not a finely tuned process.</p>

<p>This might effectively mean that there are many different imperfections at many different depths that can scatter light. Effectively, this'd mean that there are many focal planes upon which an image can form.</p>

<p>If the depth of this ground surface is in the hundreds of microns (probably not hard to imagine given the coarseness of the process), then one would effectively be forming images on multiple focal planes on the ground glass. Closer objects would come into focus further into the ground glass (closer to you). Albeit maybe only hundreds of microns (or less) closer to you than an object further away.</p>

<p>But given that we expect a 2D image, maybe this slight offset would be enough for our eye-brain systems to interpret this as '3D'.</p>

<p>Is that reasonable? Is that it? Or is there some other explanation?</p>

<p>BTW I'm seeing this on Hasselblad & Mamiya medium format focusing screens, but really one should see it on any ground glass surface. Not sure if my Canon focusing screen is a ground glass surface, but will check for the effect on that when I get home.</p>

<p>Thanks in advance for humoring me :)<br>

-Rishi</p>

<p>Actually, it's the opposite. The coarser the screen, the less of the 3D effect you'll see.</p>

<p>What you are seeing is a result of the screen having a small scattering angle, so the image on the screen has an "aerial image" characteristic, i.e. it consists of multiple planes, at different depths away from the focus screen surface, which could be considered to be the primary plane. So, there's a parallax difference between the image "inside" the screen for objects farther away than the primary plane on the lens side, and images "outside" the screen.</p>

<p>It's actually a problematic effect, it makes it harder to judge focus, unless you use aerial image techniques, such as using small head motions to let you visualize the primary plane at the focus screen and see if the desired subject is floating above or below that plane.</p>

Might simply be that unlike SLR camera viewfinders, one looks at the image on a LF ground glass with both eyes open, binocular vision.

<p>Joseph and Ivan are both correct. There's also a characteristic of vision that kicks in to amplify the effect.<br>

The eye doesn't take a little picture of an image and throw it into the brain. It looks at three thing (1) Massing of areas and how they relate to each other (2) Angles and shapes and how they relate to each other (3) a grid to place them on to build a picture. It takes all the info from two of these images derived from concave screens and looks for something to turn into a 3D image. Think of all the conflicting data that has to be thrown away for this to work. <br>

You can throw a monkey wrench into the whole process to get different results. A good example is what artists do to convert a 3D image to a 2D image. There are several tricks. Look through an opening the proportions of your paper to pump up the minds spacial relationships and grid. Or, actually look through a grid to do the same thing. The simplest thing is to close one eye to rid the process of two images to work with for 3D. The brain immediately goes looking for something in the image to create a 3D image. What it tends to find are shadows that weren't obvious before to define form coupled with layering created with the change in focus to determine front to back relationships. Both of these tasks tend to be best done with stereo vision, the default-mode so to speak.<br>

Now take your ground glass. What is the brain going to go looking for when you look at this 2D image on the screen. It seems that it should find a single focal plain and switch to things like relative size, shadows, and light values to interpret form and depth. But your stereo vision finds the effect described by Joseph, grabs it and runs with it because it's the default-mode for determining depth. This is why so many visual tasks require "Training the eye." You have to change your default-mode to one that processes images in a useful way for the task at hand.</p>

<p>Interesting stuff; thanks guys. But I'm afraid I don't follow.</p>

<p>Joseph, you mentioned that the image has 'aerial image' characteristics in that it has multiple planes. But how can one surface create multiple focal planes? I was hypothesizing that there are multiple scattering surfaces since the ground glass has depth... so there could be effecting scattering planes at different depths -- are you disagreeing with that?</p>

<p>And so I understand you correctly, are you saying that a closer object will actually fall slightly behind the 'primary' focal plane (using your terminology here for the focusing screen surface)? But if there's *just* glass there (not ground surface but purely transmissive glass), then those light rays from that closer object will not scatter at that surface, and should continue on... i.e. won't contribute to the image we see on the ground glass. Right?</p>

<p>I feel like there needs to be multiple scattering surfaces for this 3D effect to occur. But I may just be severely confused.</p>

<p>I actually haven't even had a chance to see what happens as you close the aperture.</p>

<p>Ultimately I also want an intuitive understanding of what 'precision focus screens are'... i.e. the ones that actually show you the true DOF for a f/1.8 lens... as opposed to the image you typically see which is more representative of what the image would look like at f/4 or so.</p>

<p>Thanks,<br />Rishi</p>