vrayies lights sometimes just don't seem bright enough

[samuel_bubat]

How physically correct are your materials ?

[psa]

To davision: My point is not to have the light looking as bright in the day scene as in the night scene, as I am aware that the longer you expose the brighter the image will be - and vice versa. My point is to have the light behave/look real, whatever exposure settings I use. In this case I expected the light to be looking much dimmer in daylight, but not to disappear almost totally. The lamps are placed in a dimly lit corner, and my real world experience tells my that you will be able to see some direct light on the wall and table, and also in the reflections on the table surface.

To samuel_bubat: My materials are pretty correct, the wall have a reflectance of 80%, the floor 65% and the furniture (light oak) an average of 57%.

I have provided a couple of real photographs from RH Architects below as a reference. The left image shows the Glo-Ball lamp I used in my test renderings. The right image shows that even though there are a lot of light coming both through the windows and from the recessed light in the ceiling, the light from the table lamps are still pretty easy to spot on the surface of the table in the front.



Best regards,

[pailhead]

Keep in mind that professional photographers use a lot of tricks to get a good image. What you're looking for in terms of reality in your digital images, is not what you see through your eyes in real life.

[davision]

Also, the one real photo on the right only shows sky light coming in through the windows, unlike your rendering which has direct sunlight. To me, the problem looks like your glo-balls are too dim in the night time render. They should be brighter and your exposure should be shorter.

[trixian]

Sorry to butt in, but I'm pretty certain psa is referring to the light cast on the objects around the lamps, not the direct look of the lamp surface itself. I agree completely with him, you can see the additive amount of light is wrong when he turns on the sun, as there is hardly any visible warm light even though the inside of the shelf is lit by low indirect sky and gi. Personally I would expect the peachy light from the "glowball" to add on top of the indirect light there.

[trixian]

To add to this, I have always found that setting a light source (vray light) to cast the correct amount of light on surfaces, always produces an ugly light source, as in completely blown out (if visible). This is working with LWF. I always have to hide the light and do other fakey stuff to make the directly visible light and fixture look correct. When photographing interiors and exteriors, the light sources often look a lot less bright then when vray calculates them. Having to always remake the physical appearance of a lamp after setting it up to shoot the right amount of light (according to its physical properties), seems a bit of a waste of valuable time, when one would expect it to fit the first time around.

[vlado]

I did a render with both mr and vray - and they appeared to be nearly identical. But the strange thing is that while they looked right in a "night" setting, the light disappeared in the daytime version.

There is a bug in mental ray where the exposure control is not applied to the directly visible IES light shapes - you can verify this very easily (regardless of exposure control settings, the RGB values over the visible light shape remain exactly the same). V-Ray on the other hand is correct in this regard and the camera exposure settings correctly dim the visible light shape.

Best regards,
Vlado

[vlado]

To add to this, I have always found that setting a light source (vray light) to cast the correct amount of light on surfaces, always produces an ugly light source, as in completely blown out (if visible). This is working with LWF. I always have to hide the light and do other fakey stuff to make the directly visible light and fixture look correct. When photographing interiors and exteriors, the light sources often look a lot less bright then when vray calculates them. Having to always remake the physical appearance of a lamp after setting it up to shoot the right amount of light (according to its physical properties), seems a bit of a waste of valuable time, when one would expect it to fit the first time around.

I can assure you that V-Ray is correct in this regard. However in the real world, you almost never directly photograph a light source. In the photos that psa posted, it is a lamp shade that is on the photos, not the light bulbs.

Best regards,
Vlado

[simmsimaging]

Looks correct to me. The exposure difference between the tungsten type lamps for the night shot is something like 8 stops different than the daylight one. You would expect the diffusers on the lamps to be quite greyed out.

The photo reference is deceptive because they look to be pro shots which may mean some good light/exposure balancing on set or retouching/processing to achieve that. Probably both. It would be more instructive to take a camera yourself and actually shoot a similar lighting scenario to see what you find.

[psa]

I agree with trixian in most regards, but I have actually been considering doing what Brett Simms is suggesting. So: I will take a photo of the Glo-ball myself, so I have a precise reference. I will then do a test render of my 3D scene without direct sunlight. It will be interesting to see what we end up with after that. I do suspect though that the light from the lamp on the table still will look to low without multiplication. But I agree that possible mistakes or errors should be ruled out so we/I don't make any wrong conclusions.

Best regards,

[simmsimaging]

Just to be clear: I agree that you would need to multiply the light intensity. In reality you would have to in order to balance direct sunlight with such a low power light as well. The only question is how much.

[Pixelab]

Actually, I've noticed the problem with VrayRT, realistic values are working well in RT, but the same value in the VFB renders the light like if it was OFF. You have then to muliply the value by 1000 or more to get the same result as you had in RT... strange, never knew if it was a bug or a limitation of the LWF method