DOF too deep(not blurry)

[Vizioen]

Sorry to be such a bother with this, has any conclusion been reached about the DOF subject?

[Vizioen]

Thread bump.

[Vizioen]

I'm going to res this again, as we're 2 years later and one of the most important settings, has a big questionmark around it. A bit weird that noone knows of anything while this is something that a lot of people use,...right?

[piotrus3333]

I'm going to res this again, as we're 2 years later and one of the most important settings, has a big questionmark around it. A bit weird that noone knows of anything while this is something that a lot of people use,...right?

an image would help. with aperture and focus dist values.

[Joelaff]

We always match DOF by eye.. Why? Because real lenses vary immensely in their DOF performance. Some lenses center the DOF around the focal plane-- with equal amounts of DOF on either side of the focal plane. Some put more beyond the focal plane in focus (this is more common). Some lenses put more in focus on the camera side of the focal plane than on the back side of the focal plane. You have to make sure it matches based on each lens by adjusting the focal distance as well as the aperture.

Then you also have the whole issue of is it an F-Stop (based on the diameter of the aperture) or is a T-Stop (based on the actual light value reaching the focal plane). Most cine lenses are calibrated for T-Stops; so the stop info the script supervision writes down is likely a T-Stop, not the actual aperture (F-Stop)-- unless they happen to be shooting with lenses calibrated in F-Stops (like most still photography lenses).... Then you get into the fact the exposure varies with focal distance as well, where focusing a lens closer causes less light to reach the film plane (VRay actually seems to calculate this into their exposure as well if you are set to use focal length vs. FOV). Just so many variables that you are better off just doing it by eye-- so long as you know what you are looking for.

That said it would still be nice if the aperture was at least consistently based on an ideal lens (say 50/50 on either side of the focal plane), and at least somewhat matched real world settings-- no matter the scene scale.I believe this is generally true, except in certain scale situations as discussed, or if you scale the camera. Or the number could be the physical aperture diameter, which is less photographer friendly, but easy enough to calculate.

Probably just added more confusion than anything else.

[kosso_olli]

What actually would help immensely is a factor to adjust the depth falloff. So we can control how blurry the image gets in front of the focal plane and behind it.
Sometimes the rendered output just looks wrong in comparison to the shot plate.

[Joelaff]

What actually would help immensely is a factor to adjust the depth falloff. So we can control how blurry the image gets in front of the focal plane and behind it.
Sometimes the rendered output just looks wrong in comparison to the shot plate.

Absolutely! A control of the rate of focal falloff, as well as a control of the front to back bias so we can have more in focus in front of or behind the focal plane. That would be awesome!

[fixeighted]

I honestly never knew there were all these extra things going on in real lenses, so thank you very much for that great explanation Joelaff
That there should be more functionality seems like a very important thing and I can't imagine why they haven't added this...it's all just numbers isn't it?

I guess that may be simplistic but surely there's some advance that can be implemented, although again, devs have remained pretty silent about this, so maybe it's either too hard or
not worth dev time to do, for whatever other reasons.

[Joelaff]

You can usually get close enough with the tools that are there. But sure would be nice to have the extra tools.

Just from a creative standpoint, even for all CG shots, where you are not even trying to match a real world camera, it would be nice to be able to control how much depth of field is in front of the plane of focus and how much is behind it, and the rate of falloff like Oliver suggested. A little spline control would be really cool, but even just an exponent would work.

[vasil.minkov]

Hi guys,

Changing System Unit Scale scales the whole scene.
For VRayPhysical camera this results in scaled focus distance, clipping planes and environment ranges. But the sensor and lens parameters (film gate and focal length) are preserved in millimeters.
Similarly to what a real camera would do, this results in different DOF and different framing. The framing difference may be slim for distant objects but it may be quite big for close-ups.

You may preserve DOF and framing by manually scaling the film gate and focal length whenever System Units Scale is changed. This may result in non-realistic film gate and focal length sizes.

The VRayPhysical camera object should not be scaled since this results in incorrect Viewport representation.

Best Regards,
Vasil Minkov

[fixeighted]

Thanks Vasil...but;

I keep re-reading this and have come to the conclusion that I must be stupid.

Or rather, the response seems to infer that I am bound to be stupid or at the very least constantly confused.

Nowhere in any documentation, or, as far as I am aware, in any of the posts about this subject, is there a specific solution given to ensure that a realistic DOF is achievable in any given scene.
Nowhere does it state that the user needs to divide by this, compensate for that, adjust this, guess that....etc..

You guys develop these things. The camera is such a crucial basic element, so document its behaviour correctly....give users the information/solution they need to get a result without needing a degree in computer science...please

Please correct me if I am wrong...... this is beginning to annoy me

[Vizioen]

This can't be coded in a way that it changes with the scene units? Either way would be nice to add this to do documentation along with some examples. Thanks.

[fixeighted]

Frankly I'd be amazed if this gets any further attention.
Maybe I'm too cynical but I'm thinking they're all spread too thin now to be able to bother.
The company is now vast and the priorities have changed, from making the best renderer to making
the highest profits for whatever CEOs etc. are in charge of pushing ahead in whichever way they think most profitable at any given moment

[vasil.minkov]

Thanks Vasil...but;
...
Nowhere in any documentation, or, as far as I am aware, in any of the posts about this subject, is there a specific solution given to ensure that a realistic DOF is achievable in any given scene.
Nowhere does it state that the user needs to divide by this, compensate for that, adjust this, guess that....etc..
...

But the DOF of VRayPhysicalCamera should be realistic regardless the System Units Scale. And if the setup is realistic then the DOF should be realistic too.
Switching System Units Scale results in a different DOF (and framing) and this is exactly what a realistic camera is expected to do.
You only have to compensate camera parameters if you want to preserve DOF and framing with a different scene scale. In other words - if you want to get the same pixel result in a giant scene then you will need a giant camera.

Of course - I might be wrong. If you have a scene that is not renedered correctly - you could share it with us for investigation.

[suzanne_doherty]

I too feel that the DOF controls should be more extreme.

Recently I wanted to replicate a "tilt shift" style, but even dropping the F stop to 0.01 didn't quite give enough DOF effect. Of course I'm trying to replicate an unusual scenario, basically a broken camera, but artistically we need the tools to give us the effects we desire.