this looks great but: Redshift support? I cant get it to work with Redshift in Cinema 4D R21. Could you do a tutorial for this configuration? My mattes just don't come out properly and I cant use the cryptomatte for anything useful.
I´ve an issue, I´m working with a scene with many cameras, in some of the views, the exr works really well with the cryptomattes, but in others, doesnt separate anything, I´ve tried many settings and still have the same, what can I do?
Well i´ve already solved this problem, i dont even know how it happened in the first place, but, if you render something, it doesnt even have to be completed, this solve the issue when its saved again
Hi! I have a small problem, programm works perfect I am very glad to use this tool, BUT, somehow we can not export or save or print from Photoshop with our final result.... so photos always look the same? Has someone a solution? Thanks in advance
thanks for this plugin ! ...am i right that i get a separate mask for every object/layer/material in the scene.maybe hundreds of them...? i´m not able to choose the special selection i want like in aftereffects?
Hello Tommes. You get masks for everything you have defined in your 3d Software. If you have made cryptomatte groups, named selections and layer collection in Vray, they will be imported as such . If you dump all 1000 objects in cryptomatte in 3d app, they will be processed in Photoshop exactly as 1000 masks. We wish Photoshop would be an video editing software, with API and full channel support, but it is not the case. There is no way to prepossess image data in PS like it is possible in AfterEffects. I have one or another idea how to do UI solution for such tasks and it will come in future updates. One small step after another :) thanks, best regards
Hello Tommes, thanks to your input we have tried to simplify the workflow. Here is a free script that makes automatically Masks from your Cryptomatte masks th-cam.com/video/VCasmKdRLc4/w-d-xo.html enjoy :) thanks
@@3dioplugins Hi Tommes, amazing plugin. Really speeds up the workflow vs the old material ID method. Would it be possible to create a similar script, but that creates masks applied to individual empty folders instead? This way, all adjustments layers can go inside the appropriate folder. Thank you!
@@velcrolog It is not Tommes, the company did the plugin. We have put the script as public domain open source code on GitHub: github.com/3d-io/AutoLayerMasks Anyone can now download and modify it :) Thanks best regards
Thank you for the tutorial. I can't seem to get your method for creating masks to work. It just creates a new RGB layer with the info in it. Normally I have to paste into a new alpha channel, select that, then go back to create the mask.
@@gidkideon hello - I did not explain the handgrips in Photoshop, because it is assumed as basic technique - but yes: you create in PS a mask by alt drag&drop into mask layer. Please consult some of the million videos in internet how to work with masks in PS. If any problems, you can always find us in our discord channel (go to exr-io.com and click on discord) thanks
This is great, however I have an issue installing v2 because I still have 1.03 and there's a error when I run the installer. Do i just delete the old Exr-IO file in the plugin folder ?
yes of course - a plugin has always an unique ID in Photoshop and if you double it, you will get automatically concurrent starts trying to access the same address. I have never tried it but i think PS wont like it at all. Please always uninstall old version before installing the new one. Some windows installation lose sometimes the tracks of previous installed software: in this case you have to delete the plugin manually from you folder. If you have installed a photoshop plugin/extension via Adobe Extension Store, you need either to go there and uninstall it vial web-site or delete it manually (if deleting manually, i have no idea if Adobe Exchange will try to install it back again by the next PS start). Best is in any case always to uninstall previous versions. best regards
Why my Cryptomatte doesn't open as a black and white image? I already tried everything but the crypto looks awful like green lines and some pink lines, but when I open it on After Effects my cryptomatte es completly fine, please help me! (Vray Next for Maya)
Try Discord channel, there you can upload screenshot or file so we can go together through it. I can't debug your file by just reading the text :( I can just guess you have not saved it in Exr container but separately, but who knows... thanks!
Anyone know if cryptomatte works with xrefs in Max/vray? I'm getting weird results so I'm assuming no. like if an object doesn't show up in the hierarchy there's no way to mask anything right?
Hello Adam, please go to the watch-list, i have already made Exr-IO Blender video Tutorial. There is also an explanation about what Blender can export and what is currently not compatible due to a fact Blender does not save Meta-data into Exr header, etc. thanks
Strange. There could be some computer registry or writing permission. Maybe some other users here can help, Exr-IO is a common tool today. If you can't get the installer to do the job, try it please from Adobe Extension store: Exr-IO is Adobe certified plugin and you can install it directly from there. It will work by installing it directly without the installer. hope this helps :)
sadly, the adjustments aren't applied to the reflections (see helmet). There's still an orange border visible around the gloves, so something isn't right... No need to export as 32bit exr... 16bit float still has more than enough information/accuracy to do color corrections. By using 32bit you're just wasting storage space.
Christophe, this is not correct what you are writing! I am sorry, I must immediately and resolutely contradict your non-scientific claim before some visitors here start believing it. 16-bit floating-point is 4 Trillion colors between 0 and 1 or 15873 gray levels between 0 and 1. As we all know, Photoshop uses a special integer mode for 16 bit: 35 Trillion colors, 32769 gray values but no (!) overexposure like in floating-point. On the other side, in Photoshop you have in 32-bit floating-point 1.2 Octillion colors between 0 and 1 and 1 Billion gray-levels. For a simple artist, who is posting his JPG or GIFs to DevianArt, this is probably beyond his imagination (and comprehension). But once you start working professionally with gamma, linear space, exposures or camera/raw-data inputs from Nuke for matte-painting or compositing, you get in trouble. Artstation forgives such bold statements, but the supervisor in SonyImageWorks or ILM does not. Also, Exr in Photoshop is always 32bit, PS can't interpret 16bit Exr. And Photoshop, as we all know, does not actually have 16bit. PS manages only 15bit plus 1 bit - you lose ultimately millions of colors. I mean, a common artist believes PS has 16bit, but once you read the code, you will find much more complex issues behind it (and there is a well explained reason why PS does only has 15+1). Please consult the documentation: www.adobe.com/devnet/photoshop/sdk.html Second thing: what does Cryptomatte mask to do with the accuracy of EXR? If you paint a blurry adjustment mask in PS, you will, depending on your premultiplied values have or not have transparent mask pixels. In the video, I was just fast clicking around, of course, it is untidy work; professional artists know exactly how to define a mask out of bw area without having any border. It is no criticism of you, but it is so much ill-informed and broad-brushed in what you write that I would never come to the idea to post such text. Our customers would dismiss us after such bold statements as incompetent.
@@3dioplugins I use 16bit EXR all the time. As do many other colegues in the VFX industry. Also, the DEFAULT bit depth of the OpenEXR format is 16bit. That is how ILM decided to develop it. So if you dont want to believe me (since you have decided I'm a 'simple deviantart artist', at least believe TD's at ILM. Funny, since you referenced them yourself.. have a quick read at www.openexr.com/about.html 16bit float is even enough to capture full exposure of the sun. It also opens perfectly fine in Photoshop. It has nothing to do with photoshop's 16bit integer mode. Photoshop wil convert it to 32bit float internally... since it has no 16bit floating point mode. Most of the best renders used in the VFX industry save as 16bit (half float) EXR. Only in verry rare/exeptional cases you really need 32bit. Next time you try to say someone is an 'ill informed and simple deviant art artist' when he is simply pointing out a few things and trying to help, maybe try to check the facts first before slamming him. Because if he turns out to be right, you might be the one that ends up looking like the 'ill informed deviantart artist'.
@@Capeau this is not correct. EXR has no default, or at least no 16bit default (read the topic, it was written PS loads 32 by default). EXR is not an image, but a data container, you can put Word doc in it. Apparently you have a very poor knowledge about EXR and i do not see where is your focus point in this discussion. The software that writes EXR may have default but it make less sense to degrade the output. It is normally always set to 32bit, otherwise you can use png. That is what is in the EXR/Specs, you must read the SDK carefully, not this one advert page. What about other aspects we addressed in our replay? Did you read them and understood the differences? It is important for us that, once you put those false statements about color depths and linear workflow, you incorporate the new information. Otherwise you would really believe 16bit is better then 32 bit and other stuff you wrote about masks. Also, it is absolutely not nice to put your "colleagues from VFX", without naming them, in front of you as a shield. They can not know that you misuse them for own claims. Give us their names please, so we can check the vfx&gfx portfolios and contact them in order to see if this is really what they use for productions. thanks!
@@IgorPosavec OpenEXR's default image data has a bitdepth of 16bit (half float, not integer). It also, supports 32bit, which was added later. Vray's default setting is also 16bit EXR's (that's why you have to enable a tickbox to save 32bit) for good reason. In more than 99% of the cases half float data (16bit) is more than enough... The fact that photoshop (or any other editing app that i know of) loads it as 32bit (which i already mentioned), has nothing to do with what I said or this topic. You can load a 16bit half-float image in it perfectly fine, it wil be converted to 32bit. I'm, not going to name any colleagues in this discussion, that's just stupid. Your replies are honestly are quite childish and offensive. I kinda felt bad later in the day after i wrote the reply and thought i maybe was a bit harsh (it was a long day) but after rereading your reply again I don't think so. Very unprofessional. Somehow when a person has a Deviantart account he is an 'ill-informed and simple artist'.. what kind of reasoning is that?! And nice way to insult a lot of clients/colleagues in the industry. Never did I say that 16bit is 'better' than 32bit. Neither one is better. Better is a relative thing... All I said is that you don't need to save renders as 32bit in most cases, only few exceptions where you really need it. And yes, I understand the differences. For simple photoshop editing its more than good enough. Arnold renderer, Vray renderer, Renderman, etc... their docs all say the same thing I do. That 16bit (half float) is good enough for most cases (unless you are going to use things that need the extra precision/date (like zdepth or displacement maps with big depth)). Since 16bit files are smaller than 32bit files this can make a HUGE difference when rendering animations in terms of storage. I'm kind of surprised to be honest at how you handle your replies and throw around insinuations like that just because i pointed out a few things. I know you have some nice software, too bad you have this attitude towards interested ppl though. I was mistaken about the differences between multimatte and regular grayscale masks (it has been a while since i used them, so I remembered wrong). It is expected behaviour that the multimatte elements behave the same as a regular greyscale mask when isolated (sadly). So the orange edge around the glove still being there is expected bahaviour. See, you could've just pointed that out in a normal way, but instead you dont even point that out but get crazy offensive about me advising to just use 16bit half float instead of 32bit (which I still recommend for already mentioned reasons).
This is amazing!!!
Warm tips : when you have a lot of materials and shapes, there will turn out to be hundreds of matte, so be carefull lol
Awesome!! I've waited for that!!!
How on earth did you paste into a layer mask? mine just puts a new layer right above the adjustment layer
this looks great but: Redshift support? I cant get it to work with Redshift in Cinema 4D R21. Could you do a tutorial for this configuration? My mattes just don't come out properly and I cant use the cryptomatte for anything useful.
It would be even more awesome if you could seperate cryptomatte masks by object IDs.
yes, it would be. A good feature request, thank you. I have added it for the next release.
I'm using vray/maya and after importing tp the photoshop my image becomes slightly purple-tinted :(
are you working in 32bit mode? cos im trying to do that and photoshop wont allow me to.
You guys rock!
I´ve an issue, I´m working with a scene with many cameras, in some of the views, the exr works really well with the cryptomattes, but in others, doesnt separate anything, I´ve tried many settings and still have the same, what can I do?
Well i´ve already solved this problem, i dont even know how it happened in the first place, but, if you render something, it doesnt even have to be completed, this solve the issue when its saved again
Does this work with Affinity Photo? I don't use Photoshop.
Thanks a lot! it will be very, very usefull!
Hi! I have a small problem, programm works perfect I am very glad to use this tool, BUT, somehow we can not export or save or print from Photoshop with our final result.... so photos always look the same? Has someone a solution?
Thanks in advance
thanks for this plugin ! ...am i right that i get a separate mask for every object/layer/material in the scene.maybe hundreds of them...? i´m not able to choose the special selection i want like in aftereffects?
Hello Tommes.
You get masks for everything you have defined in your 3d Software. If you have made cryptomatte groups, named selections and layer collection in Vray, they will be imported as such . If you dump all 1000 objects in cryptomatte in 3d app, they will be processed in Photoshop exactly as 1000 masks.
We wish Photoshop would be an video editing software, with API and full channel support, but it is not the case. There is no way to prepossess image data in PS like it is possible in AfterEffects.
I have one or another idea how to do UI solution for such tasks and it will come in future updates. One small step after another :)
thanks, best regards
Hello Tommes,
thanks to your input we have tried to simplify the workflow. Here is a free script that makes automatically Masks from your Cryptomatte masks
th-cam.com/video/VCasmKdRLc4/w-d-xo.html
enjoy :)
thanks
@@3dioplugins Hi Tommes, amazing plugin. Really speeds up the workflow vs the old material ID method. Would it be possible to create a similar script, but that creates masks applied to individual empty folders instead? This way, all adjustments layers can go inside the appropriate folder. Thank you!
@@velcrolog It is not Tommes, the company did the plugin. We have put the script as public domain open source code on GitHub:
github.com/3d-io/AutoLayerMasks
Anyone can now download and modify it :)
Thanks
best regards
Thank you for the tutorial. I can't seem to get your method for creating masks to work. It just creates a new RGB layer with the info in it. Normally I have to paste into a new alpha channel, select that, then go back to create the mask.
Are you alt clicking into the mask layer first before pasting?
@@gidkideon hello - I did not explain the handgrips in Photoshop, because it is assumed as basic technique - but yes: you create in PS a mask by alt drag&drop into mask layer. Please consult some of the million videos in internet how to work with masks in PS. If any problems, you can always find us in our discord channel (go to exr-io.com and click on discord) thanks
This is great, however I have an issue installing v2 because I still have 1.03 and there's a error when I run the installer. Do i just delete the old Exr-IO file in the plugin folder ?
yes of course - a plugin has always an unique ID in Photoshop and if you double it, you will get automatically concurrent starts trying to access the same address. I have never tried it but i think PS wont like it at all.
Please always uninstall old version before installing the new one. Some windows installation lose sometimes the tracks of previous installed software: in this case you have to delete the plugin manually from you folder.
If you have installed a photoshop plugin/extension via Adobe Extension Store, you need either to go there and uninstall it vial web-site or delete it manually (if deleting manually, i have no idea if Adobe Exchange will try to install it back again by the next PS start).
Best is in any case always to uninstall previous versions.
best regards
Why my Cryptomatte doesn't open as a black and white image? I already tried everything but the crypto looks awful like green lines and some pink lines, but when I open it on After Effects my cryptomatte es completly fine, please help me! (Vray Next for Maya)
Try Discord channel, there you can upload screenshot or file so we can go together through it. I can't debug your file by just reading the text :( I can just guess you have not saved it in Exr container but separately, but who knows... thanks!
@@3dioplugins ok thank you, I'm already there
Anyone know if cryptomatte works with xrefs in Max/vray? I'm getting weird results so I'm assuming no. like if an object doesn't show up in the hierarchy there's no way to mask anything right?
yes it does, well at least it worked for me
what about vray for maya?
Does it work also with Blender2.8 Cryptomatte nodes?
Hello Adam, please go to the watch-list, i have already made Exr-IO Blender video Tutorial. There is also an explanation about what Blender can export and what is currently not compatible due to a fact Blender does not save Meta-data into Exr header, etc. thanks
I can't seem to even get this to install to the Photoshop Plugin directory? I direct it towards the folder and nothing appears after install?
Strange. There could be some computer registry or writing permission. Maybe some other users here can help, Exr-IO is a common tool today.
If you can't get the installer to do the job, try it please from Adobe Extension store: Exr-IO is Adobe certified plugin and you can install it directly from there. It will work by installing it directly without the installer.
hope this helps :)
@@3dioplugins Thank you so much for the quick reply! I hopped on to the Adobe Exchange and found it. Installed perfectly that way!
sadly, the adjustments aren't applied to the reflections (see helmet).
There's still an orange border visible around the gloves, so something isn't right...
No need to export as 32bit exr... 16bit float still has more than enough information/accuracy to do color corrections. By using 32bit you're just wasting storage space.
Christophe, this is not correct what you are writing! I am sorry, I must immediately and resolutely contradict your non-scientific claim before some visitors here start believing it.
16-bit floating-point is 4 Trillion colors between 0 and 1 or 15873 gray levels between 0 and 1.
As we all know, Photoshop uses a special integer mode for 16 bit: 35 Trillion colors, 32769 gray values but no (!) overexposure like in floating-point.
On the other side, in Photoshop you have in 32-bit floating-point 1.2 Octillion colors between 0 and 1 and 1 Billion gray-levels.
For a simple artist, who is posting his JPG or GIFs to DevianArt, this is probably beyond his imagination (and comprehension). But once you start working professionally with gamma, linear space, exposures or camera/raw-data inputs from Nuke for matte-painting or compositing, you get in trouble. Artstation forgives such bold statements, but the supervisor in SonyImageWorks or ILM does not.
Also, Exr in Photoshop is always 32bit, PS can't interpret 16bit Exr. And Photoshop, as we all know, does not actually have 16bit. PS manages only 15bit plus 1 bit - you lose ultimately millions of colors. I mean, a common artist believes PS has 16bit, but once you read the code, you will find much more complex issues behind it (and there is a well explained reason why PS does only has 15+1).
Please consult the documentation: www.adobe.com/devnet/photoshop/sdk.html
Second thing: what does Cryptomatte mask to do with the accuracy of EXR? If you paint a blurry adjustment mask in PS, you will, depending on your premultiplied values have or not have transparent mask pixels. In the video, I was just fast clicking around, of course, it is untidy work; professional artists know exactly how to define a mask out of bw area without having any border.
It is no criticism of you, but it is so much ill-informed and broad-brushed in what you write that I would never come to the idea to post such text. Our customers would dismiss us after such bold statements as incompetent.
@@3dioplugins I use 16bit EXR all the time. As do many other colegues in the VFX industry.
Also, the DEFAULT bit depth of the OpenEXR format is 16bit. That is how ILM decided to develop it. So if you dont want to believe me (since you have decided I'm a 'simple deviantart artist', at least believe TD's at ILM. Funny, since you referenced them yourself..
have a quick read at www.openexr.com/about.html
16bit float is even enough to capture full exposure of the sun.
It also opens perfectly fine in Photoshop. It has nothing to do with photoshop's 16bit integer mode. Photoshop wil convert it to 32bit float internally... since it has no 16bit floating point mode.
Most of the best renders used in the VFX industry save as 16bit (half float) EXR. Only in verry rare/exeptional cases you really need 32bit.
Next time you try to say someone is an 'ill informed and simple deviant art artist' when he is simply pointing out a few things and trying to help, maybe try to check the facts first before slamming him. Because if he turns out to be right, you might be the one that ends up looking like the 'ill informed deviantart artist'.
@@Capeau this is not correct. EXR has no default, or at least no 16bit default (read the topic, it was written PS loads 32 by default). EXR is not an image, but a data container, you can put Word doc in it. Apparently you have a very poor knowledge about EXR and i do not see where is your focus point in this discussion.
The software that writes EXR may have default but it make less sense to degrade the output. It is normally always set to 32bit, otherwise you can use png.
That is what is in the EXR/Specs, you must read the SDK carefully, not this one advert page.
What about other aspects we addressed in our replay? Did you read them and understood the differences? It is important for us that, once you put those false statements about color depths and linear workflow, you incorporate the new information. Otherwise you would really believe 16bit is better then 32 bit and other stuff you wrote about masks. Also, it is absolutely not nice to put your "colleagues from VFX", without naming them, in front of you as a shield. They can not know that you misuse them for own claims. Give us their names please, so we can check the vfx&gfx portfolios and contact them in order to see if this is really what they use for productions. thanks!
@@IgorPosavec OpenEXR's default image data has a bitdepth of 16bit (half float, not integer). It also, supports 32bit, which was added later.
Vray's default setting is also 16bit EXR's (that's why you have to enable a tickbox to save 32bit) for good reason.
In more than 99% of the cases half float data (16bit) is more than enough...
The fact that photoshop (or any other editing app that i know of) loads it as 32bit (which i already mentioned), has nothing to do with what I said or this topic.
You can load a 16bit half-float image in it perfectly fine, it wil be converted to 32bit.
I'm, not going to name any colleagues in this discussion, that's just stupid. Your replies are honestly are quite childish and offensive.
I kinda felt bad later in the day after i wrote the reply and thought i maybe was a bit harsh (it was a long day) but after rereading your reply again I don't think so. Very unprofessional.
Somehow when a person has a Deviantart account he is an 'ill-informed and simple artist'.. what kind of reasoning is that?! And nice way to insult a lot of clients/colleagues in the industry.
Never did I say that 16bit is 'better' than 32bit.
Neither one is better. Better is a relative thing...
All I said is that you don't need to save renders as 32bit in most cases, only few exceptions where you really need it. And yes, I understand the differences.
For simple photoshop editing its more than good enough.
Arnold renderer, Vray renderer, Renderman, etc... their docs all say the same thing I do. That 16bit (half float) is good enough for most cases (unless you are going to use things that need the extra precision/date (like zdepth or displacement maps with big depth)). Since 16bit files are smaller than 32bit files this can make a HUGE difference when rendering animations in terms of storage.
I'm kind of surprised to be honest at how you handle your replies and throw around insinuations like that just because i pointed out a few things. I know you have some nice software, too bad you have this attitude towards interested ppl though.
I was mistaken about the differences between multimatte and regular grayscale masks (it has been a while since i used them, so I remembered wrong). It is expected behaviour that the multimatte elements behave the same as a regular greyscale mask when isolated (sadly). So the orange edge around the glove still being there is expected bahaviour.
See, you could've just pointed that out in a normal way, but instead you dont even point that out but get crazy offensive about me advising to just use 16bit half float instead of 32bit (which I still recommend for already mentioned reasons).