Image saved from vfb not the same ..

[Bigguns]

I have a quesiton about smething that is bothering me for a while.

Let's say, when the render is finished in the vbf, if I make a copy of it to max frame buffer,m it's not the same. there kind of less contrast, less saturation. Enought of a difference to bother me. I'm using max default LUT and gamma setting, max 2014. And, when I save it, it also look like when I copy it to the max frame buffer.. it don't look like in the VFB ... any idea ? And.. also, when I save the image, I have manually change the gamma to 2,2 otherwise it's washed out. Is there a way to just save and get the same result as in the VFB ?? Corona, if I copy the CFB to max, it's exactly the same and when I save it also. Only way I found to have the same result, is to save an EXR with Vray.. .. heavy and not user friendly for test renders..

Also, I started to play with LUTs recently and quite like it, but again, when I save the image, it don't save what I see in the VFB at all.. it don't take into account the LUTS .. how can I just save the image to look exactly the same as in the VFB please ?? If there is no simple way, can you Vlado please add a button that when it's pressed, it would save it exactly as we seen it in the VFB ? Would be really usefull !

Thank's

[vlado]

As far as the saved image is concerned, first make sure your gamma settings are default everywhere, both in V-Ray and in 3ds Max - i.e. in V-Ray, color mapping mode is set to "color mapping only" (or "none"), the 3ds Max gamma is enabled and set to 2.2, and the 3ds Max input and output gamma is set to 2.2. Note that there is no UI for setting input/output gamma anymore, so you have to use MaxScript, in case you dragged non-default settings from some old scene. In particular, in the MaxScript editor, type

fileoutgamma=2.2

Second, as you have found out, we don't save the LUT or OCIO color corrections. This is because the original intention of LUT files is to adjust the display of the image only; they were never intended to be baked into images. The fact that people want to do it know, is a recent thing when users outside VFX discovered LUTs. We might add an option to do that in the future, but we don't have it now. Note that baking LUTs into images automatically breaks any compositing with render elements.

Third, the V-Ray VFB applies an sRGB color correction, which is accepted as standard and is what software like Nuke expects. 3ds Max on the other hand uses plain gamma 2.2 color correction, which is generally not standard. The two are very close for the most part, but differ slightly in the darker colors where sRGB has a little more contrast than gamma 2.2. There have been requests to add a button for gamma 2.2 correction to the VFB, but we don't have it now.

Last, if you want to replicate the image in the VFB exactly, you can copy it to the clipboard (click anywhere in the VFB image and press Ctrl+C) and then paste wherever you want.

Best regards,
Vlado

[Art48]

Last, if you want to replicate the image in the VFB exactly, you can copy it to the clipboard (click anywhere in the VFB image and press Ctrl+C) and then paste wherever you want.

Best regards,
Vlado

Didn't know that one! Could come in handy for previews and photoshop adjust work, not having to save, then open in PS!

[Bigguns]

Thank's Vlado good to know! Yeah me I don't use a third party to make lots of pass etc.. I like my render to look good straight out , so for me it would be really usefull to have an option that I can check in vray that would save them as they are in the VFB ( in case I do an animation also, don't want to have to save them manually , if we put the button only in the VFB. When pressed, it can alert the person and say it would loose their pass etc. But it would be really handy !

For now, since I don'T do anim for now, that ctrl-c trick should do it , better then the print screen I was doing lol.

[skogskalle]

I am having this problem now. Vray colormapping set to "colormapping only", MAX gamma is set to 2,2, I used maxscriptlistener to set output gamma 2,2.

Take a look at the attached image. The image saved from the vray frame buffer is alot brighter than the one using "printscreen" (look at the blue curtain on the right for example) and I think it also has a slightly greenish tint . Is this simply because of the difference between srgb and gamma2,2?? Or have I done something else wrong? ( I think the difference is really big, thats why I ask)

I am rendering a 500 frame animation so I cant really copy paste every image... Is there a way (in vray or outside) so make my saved image look exactly like in the vray framebuffer?

Thanks
K

[vlado]

Can you post an OpenEXR with the image so that I can check something? It does look like it might be the sRGB vs Gamma 2.2 difference, but I want to make sure.

Best regards,
Vlado

[skogskalle]

sure,

this one is saved from the vfb to open exr...

Thanks for looking at this!
K

[vlado]

Yes, it's the difference between sRGB and Gamma 2.2. There are several things we can do about it:

a) We could nudge the saved images a bit when converting them for 3ds Max for saving in 8-bit formats so that they come out identical to the VFB. This will take a while and has some potential for turning into a mess;

b) We could write a small command-line tool to batch-convert 8-bit file formats from Gamma 2.2 to sRGB;

c) We could get you a .bcurve file for the curves control in the VFB that, when enabled, will do the above conversion in the VFB;

d) If you have Nuke (or maybe some other tool, maybe the free Fusion version?), you can use that to load OpenEXR files and resave them to an 8-bit format with proper sRGB curve;

e) We are already working on adding a Gamma 2.2 option to the VFB; this will force the VFB to appear identical to what is saved (instead of forcing what is saved to look like the VFB).

f) You could ask Autodesk to implement actual sRGB profile in 3ds Max.

Thoughts?

Best regards,
Vlado

[skogskalle]

Great, then I know whats wrong I usually render very bright images that dont have this issue...

About your solutions:

c) a curve-file would be great If I render locally. But im not 100% sure what happens if I render the animation on the Rebus or Ranch-farms, will they apply the curve or not?

b) For my current project I think a command-line tool would be perfect. If I understand this correctly I could then render the animationframes on a farm, saving the images using gamma 2,2, and then batch-convert the whole lot to srgb with the tool locally?? That would be great!

f) would be best I guess but you guys are usually alot faster to implement new stuff than Autodesk soo.....

d) I will use After Effects for this project so its a possibility. But I would prefer to render to 8-bit to save time downloading from the farm (very short deadline, and wonky internet access at home at the moment).

If you could write the command-line tool like in the next few days that would be perfect for me. If you dont have the time right now (which is totally fine) I will try the exr-method.

Big thanks Vlado. Fast feedback as always!
K

[Dan Marks]

I'm pleased to find this thread. This has always bugged me a bit, and it would be great if this could be fixed.

Side note, I've been shown some side by side tests, Corona vs Vray, I see Corona is far behind Vray in many ways, but it does have a 'one hit wonder' approach which is quite successful for a lot of new artists. I'm proficient in Vray so I'm not really that interested, but for one thing, it seems something (possibly color mapping) is applied more realistically than in Vray. What I notice, for example is the contrast across a ceiling or wall has more range, reflections also. Vray renders in comparison look ever so slightly washed out.. I'd love to close this gap somehow.

[LudvikKoutny]

"I see Corona is far behind Vray in many ways"

Got those tests or any other physical evidence to back up this claim? I myself very often compare both, and I've never found one or the other being far behind in any way.

[Dan Marks]

Ha, you got me there. I admit it is hearsay and I haven't had a chance to run with Corona myself, however I have worked with a team which has, who although are impressed with Corona's still image capability are yet to be convinced to how they could create an architectural visualisation animation in real world production with Corona. Having done a few of these larger jobs myself where we are stressing the hardware resources to the limit, and knowing how well developed the pipeline is in Vray I am inclined to accept their word for it. Still, I am open to any counter arguments..

[LudvikKoutny]

Well, it's hard to argument against something that's not backed up with any proof. These days, I dare to say that amount of people using Corona to produce archviz animations is more or less equal to V-Ray. Performance wise, I actually find Corona slightly faster, while V-Ray having edge over Corona when it comes to feature set, but not the archviz one.

I am just really interested in differences between the two, as there aren't that many, especially when it comes to general performance.

[Dan Marks]

For me the draw to corona right now is that subtle difference in the way it seems to handle contrast between light and dark areas better, unfortunately the comparison tests I've seen weren't my property so I can't post them here. I suppose the only way I could make the point scientifically would be to provide my own comparison tests, and that will require time, perhaps one day..

[LudvikKoutny]

As for the differences in shading, I've noticed them too, but they are indeed very subtle.

My suspicion is that when V-Ray uses brute force retracing of light cache in corners, it does it with with low amount of brute force bounces. One of the cases where V-Ray is slower than Corona is brute force GI, so it defaults to just 3 bounces.

I have not done any extensive testing of this particular thing, but I have a hunch that the shading differences is indeed happening due to the low amount of brute force bounces in the corners when retracing is enabled, making lighting falloff in the corners a bit more harsh, sort of Ambient Occlusion looking.

But anyway, I am more interested in the statement of Corona being "far behind in many ways", because I don't really see that at all.