Stability and usability

[LudvikKoutny]

I am afraid the IPR issues are chicken vs egg kind of problem.

IPR is buggy and unstable because not many people report bugs for it. Not many people report bugs for it because not many use it, and not many use it because it's buggy and unstable. In general, for any interactive rendering solution to be attractive to users, it needs to meet 4 basic requirements:

1, It needs to be stable
2, Everything needs to update reliably
3, It needs to have fast feedback - any action that takes more than one second from being performed to being reflected in IR VFB is not acceptable
4, It needs to have convenient workflow - It should not be inconvenience to use IPR

Unfortunately, Vray's IPR does not meet any of these requirements, therefore I use it only rarely, and more often than not, I come to a conclusion that choosing to use it was actually a mistake.

I understand there's some competitive race in this field, but it would have been better if it was released once it's ready.

[hoppergrass]

I am using it more and more and am enjoying more and more. I have gotten into the habit of saving first when I know I will be using it for an extended period of time or will be doing a lot of edits.

One thing I'd love to have is for the IR to STOP automatically when the VFB is closed. Too many times I have closed the VFB, forgetting to stop the IR then some time later opening the VFB again only to find that its been running the whole time.

[matanov]

4, It needs to have convenient workflow - It should not be inconvenience to use IPR

Describe "convenient workflow"

One thing I'd love to have is for the IR to STOP automatically when the VFB is closed.

It will be fixed, thanks!

[LudvikKoutny]

Describe "convenient workflow"

For one ability to run it inside of a viewport, not VFB. If you run it in VFB, VFB constantly covers some part of the UI you need at the moment, and the workflow turns into endless shuffling of VFB around.

[rikou]

I am using it more and more and am enjoying more and more. I have gotten into the habit of saving first when I know I will be using it for an extended period of time or will be doing a lot of edits.

One thing I'd love to have is for the IR to STOP automatically when the VFB is closed. Too many times I have closed the VFB, forgetting to stop the IR then some time later opening the VFB again only to find that its been running the whole time.

I'm agree with you ! I work in the same way and I have to say, that changed a lot of thing for me...
Also agree with the fact that when you close VFB, it should stop IPR. Thanks for the fix !

[Pixelcon]

Can you give an example?

Best regards,
Vlado

Only had 5 mins to do this, so it's not a great test.

I've attached a scene rendered more or less the same with corona and vray. To my eyes the Vray bump in incorrect and less sharp.


Obviously the lights are a bit different but that doesn't change the fact that they are shading differently and in my opinion, confirm my suspicions.

First is the scene layout from above, second is the Corona render, and last is Vray.

[vlado]

You didn't post a scene, so can I ask you repeat the test with a VRayColor2Bump texture?

In any case, keep in mind that bump mapping is a cheat. There isn't a "more correct" way to do it.

Best regards,
Vlado

[LudvikKoutny]

While bump is certainly a fake, I would argue that its "correctness" can be quite easily measured by how close it is to a real displaced result. Of course it will only work to some degree, let's say 1mm displacement viewed from 1m away, but still, you can measure how correct it looks by comparing it to real displaced result and see if normals appear to be facing same way. On the picture above, it looks as if V-Ray's bump was somehow inverted

[Pixelcon]

I think I should the light in a different position so it's a more obvious test. I'll do that, try it vs. displacement, and come back you, possibly eating my words.

[^Lele^]

For one ability to run it inside of a viewport, not VFB. If you run it in VFB, VFB constantly covers some part of the UI you need at the moment, and the workflow turns into endless shuffling of VFB around.

Right, whereas if it takes one viewport, it can't be moved to a second monitor, while it takes exactly the same space as the VFB (minus the CC, history, and so on.).
Down-voted with two hands.

While bump is certainly a fake, I would argue that its "correctness" can be quite easily measured by how close it is to a real displaced result. Of course it will only work to some degree, let's say 1mm displacement viewed from 1m away, but still, you can measure how correct it looks by comparing it to real displaced result and see if normals appear to be facing same way. On the picture above, it looks as if V-Ray's bump was somehow inverted

Below, i'll let you pick which of the two is a 1cm displacement, and which a 1cm color2Bump (hint: one has displaced edges.).
This, through the very basic, by-the-book, setup: add a map to a displace mod, instance the same map into a color2bump, put that into the 100% influence bump slot of the shader.

I am afraid the IPR issues are chicken vs egg kind of problem.

IPR is buggy and unstable because not many people report bugs for it. Not many people report bugs for it because not many use it, and not many use it because it's buggy and unstable.

You seem to have access to statistics we do not possess, or read wrong.
Either way, wouldn't you think it wiser to have the interested party tell you how things are, instead of becoming the claimer of truths in their stead?
I thought so.

In general, for any interactive rendering solution to be attractive to users, it needs to meet 4 basic requirements:

1, It needs to be stable
2, Everything needs to update reliably
3, It needs to have fast feedback - any action that takes more than one second from being performed to being reflected in IR VFB is not acceptable
4, It needs to have convenient workflow - It should not be inconvenience to use IPR

Unfortunately, Vray's IPR does not meet any of these requirements

Big, big statement, for absolutely zero peer-reviewable backup data. I wonder how many hours of work you've put into it...

therefore I use it only rarely, and more often than not, I come to a conclusion that choosing to use it was actually a mistake.

I see. You were being your usual positive self.
The issue is always, invariably, with V-Ray, never with how you set up your scene.
It's a good default heuristic, keep using it, it will serve you well over the years.

I understand there's some competitive race in this field, but it would have been better if it was released once it's ready.

What you understand it to be, and what it actually is, are in fact two quite different things indeed.
But do hold on tight to your convictions.

[LudvikKoutny]

Right, whereas if it takes one viewport, it can't be moved to a second monitor, while it takes exactly the same space as the VFB (minus the CC, history, and so on.).
Down-voted with two hands.



Below, i'll let you pick which of the two is a 1cm displacement, and which a 1cm color2Bump (hint: one has displaced edges.).
This, through the very basic, by-the-book, setup: add a map to a displace mod, instance the same map into a color2bump, put that into the 100% influence bump slot of the shader.




You seem to have access to statistics we do not possess, or read wrong.
Either way, wouldn't you think it wiser to have the interested party tell you how things are, instead of becoming the claimer of truths in their stead?
I thought so.


Big, big statement, for absolutely zero peer-reviewable backup data. I wonder how many hours of work you've put into it...


I see. You were being your usual positive self.
The issue is always, invariably, with V-Ray, never with how you set up your scene.
It's a good default heuristic, keep using it, it will serve you well over the years.


What you understand it to be, and what it actually is, are in fact two quite different things indeed.
But do hold on tight to your convictions.

1,
Ability to use IPR in viewport does not remove ability to use it in VFB as well. There are multiple kinds of people with multiple workflows. Some, for example, have slate editor open on the other screen

2,
I never said V-Ray's bump was not capable of matching displacement, I was just reacting to Vlado saying that there's no way to measure "correctness" of the bump mapping, because it is a fake. I do not agree with it and think you can measure correctness of bump mapping by how close, visually, it looks to the surface with real, slight displacement. I think the main confusion for the users here is, that in order to get better results, you need to plug your bump map through an adapter node, VrayColor2Bump in this case. Many people may not know that is the reason their bump looks odd, let alone the fact it does not add much to the ease of use.

3,
I speak just my own opinion, mostly based on observations on these forums. I've seen quite a few threads about people complaining about how some things in IPR do not update or how unstable it is, and then when someone requests scene to reproduce, people usually say they are busy and will send it later, or that it's under NDA, or they just flat out don't reply at all.

4,
So what you are saying is that there are some users who would not like IPR to be stable, or some, whom would prefer if not everything during IPR session updated reliably? Or people who would prefer IPR to be little slower and less interactive, so that they have enough time to make a coffee? Or people who just don't like when features are easy and quick to use? Haven't really seen many threads witch such requests over here. Quite the contrary actually, these are the areas of IPR which seem to be most criticized here on the forum

5,
Alright, so if IPR crashes on me during middle of a session it's problem in my scene setup, right? If I manually refresh IPR after 5 minute session and notice some materials/maps changed abruptly because they did not update correctly, is my scene setup at fault too? What about ForestPack rendering viewport instead of renderer geometry, how do I set up my scene so that doesn't happen?

6,
Implication here being that V-Ray is on top of everything, and doesn't have any serious competition at the moment?

[Pixelcon]

I had hoped the type of conversation this has recently become was not going to be, but here we are and to me this sums up exactly what leaves a bitter taste in my mouth when I come to these forums as of late.

[Pixelcon]

You didn't post a scene, so can I ask you repeat the test with a VRayColor2Bump texture?

In any case, keep in mind that bump mapping is a cheat. There isn't a "more correct" way to do it.

Best regards,
Vlado

Ok so color2bump gives me the same result, and displaced version looks more like corona's result than vray's...

[PIXELBOX_SRO]

I had hoped the type of conversation this has recently become was not going to be, but here we are and to me this sums up exactly what leaves a bitter taste in my mouth when I come to these forums as of late.

too true James

[^Lele^]

Well, I'll further reduce my input into these type of conversations when they are derailed from exact (as the starting post.) to blanket statement, pub-style, hear-and-say gibberish.
At the very least it won't contribute to further bad taste at the back of the mouth, if it won't contribute anything to bettering V-Ray.
Thankfully, the "ignore" function in the forum works, so i won't even have to bite my lip.