Hi Bernd! Not just this tutorial, but all your other VFXstudy tutorials on Fusion are the best! I love your in-depth and CLEAR explanations of the principles behind what you're doing. I believe this is the best way we can all learn -- understanding the why and how so that we can apply that knowledge to our own ideas with our own tweaks. And you're always covering multiple useful functions and tips in each tutorial! I come to your tutorials to learn not just the stated topic, but how to use Resolve/Fusion effectively in general. Please keep up the wonderful tutorials!
Thank you. Coming from Nuke to Fusion, and trying to do some digital matte painting, this tutorial was so very helpful, especially the last bit about the projection IDs and the catcher!! So perfect!
What a great video, thank you! I agree with others, not only this video, but all of your content is extremely informative... you cover highly nuanced yet essential fine points... thank you again!👍💯
Excellent tutorial, Bernd, thank you. I noticed in the Fusion chapter of the Resolve manual there's another kind of projection, using a UVMap Node. The manual only gives it half a page and I don't really understand what I need to do to use it, but the result looks very interesting - creating a 3D scene out of 2D images in such a way that they interact with a 3d model, in terms of reflections and shadows. I would love to see a tutorial on this.
Great info on fusions catcher node the info out there is almost inscrutable to beginners. The only real common example is just text describing projecting buildings with windows cut outs.
Texted directly generated 2d text. Or maybe you want a flat text in 3d, but that's what texted does by default. Normally there's no need for image plane or projection if all you need is 2d text
Is there a way to snap a plane (or any other object) to a certain tracking point? Because that’s the only way to be truly accurate and Fusion is not very friendly when it comes to orthogonal views. So if like to select a tracker and have an object being snapped on there.
Not sure I understand you right... If you want to place something in 3D it works with an accurate 3d camera track. In 2d you have point and planar tracking of course but then typically you don't need the full 3d scene.. of course you can build one and do 2d matchmove after the render. If you want a specific point from 3d accessible in 2d you have the locator tool for that.
4:05 mind-blowing... QUESTION: Would this be a simpler/easier/better way in which to create a 3d environment, as opposed to bringing in an HDRI converted to .exr file?
Well, what would you do with the HDRI - connect to a sphere? So basically the same concept and would allow your camera to tilt up and down as well. The added dinmic range in HDRI may or may not be useful depending on your workflow. Often HDRIs are used for lighting, but could be used like this here as well I suppose. But either way this works for a nodal pan/tilt as long as the camera is fixed on a tripod - if the camera moves in any other way, it might fall apart - well depending on the amount of movement.
Sure, it's the logitech K780 keyboard and a mojo silent mouse. The keyboard has basically all keys + number block but is a bit more compact and I can put up my phone = camera remote on it. The mouse is vertical which I like for ergonomical reasons and not too noisy on recordings. ... and unfortunately, nobody is paying me for saying all this :P
@@leecaste I thought of developing a special keyboard for VFX artists which I can sell for 1000$... like the Resolve Edit keyboard... but with a special MERGE button 😋😋😋
Great stuff - When working in projector mode with a boris effect imported (e.g. explosion) is there a way to alpha the black background without the explosion getting transparrent? (I want to try this in VR360 instead of imageplane, but can't seem to find the right setting).
Not sure about your plugin, but sometimes you can lumakey vfx that have black background. If you get an image with transparency you can project it with texture projection.
Is there a node in the Fusion that can combine multiple images into one big picture? What is the biggest image dimensions that Fusion can accept and use?
You can just merge the images with the standard merge node and transform them smaller or bigger either in the merge or with additional transform node before. Resolution is normally not an issue, unless you run into performance issues due to extreme sizes. I think in theory you can use images up to 16K by 16K
Well, you can create all kinds of backgrounds if you have the material and artistic vision... I don't remember the Patriot Act - though I've seen it some time ago. You can use a Luma key to key on white, but depending on the foreground your results may vary. With black hair, dark clothes and good lighting separation, keying against white might actually work reasonably well - in other scenarios not so much...
Hi and thank you for such clear and informative tutorials. At about 9:44 or so you speak of rotating the plane and getting a 3D perspective. However, in the free version of Resolve, the Render3D node will only display the straight ahead view of a photo and not the rotated perspective. It appears that rotating the image via the plane or Merge3D only results in the Render3D displaying a more narrow photo still in the original orientation. This is clearly evident if you use a photo of a bird with it's wings spread. Is there some manner of rendering the rotated image? Thank you!
Hi, the Render3D renders the scene exactly as the camera3d looks onto it. If you rotate the merge3D everything including the attached camera will rotate -> no effect. You need to rotate the imageplane3d node alone. If it appears squeezed or rotated depends on the perspective / angle of rotation.
@@VFXstudy Yes. Quite right. My Render3D node had become stuck for some reason. I reset it and I now get the expected results, just as you describe them. And the effect is quite different if I rotate an ImagePlane3D, which maintains aspect ratio and simulates looking at a photo from the edge, or if I rotate a Shape3D plane, which allows the image to distort. Very interesting. Thanks again!
Should be possible, but few things to check: projection needs to be texture projection (not light or ambient light). Shadows enabled in Render node, light source supports shadows (spotlight). Used to only work with software renderer. Now OpenGl renderer got enhanced, but not not sure - haven't tried this one since. Also check my channel for the tutorial on shadow projection
Either move it around in 3D space or better make sure plane and camera start from the default locations and then only move the camera back in z. Unless you are talking about projection. Then make sure your projector is aligned with the camera - often on frame 0 or on last frame. You might start by copying the camera to use it as projector and later might remove the keyframes on the projector.
hey friend - where are good resources to download free spheres or 360 videos? do I search "latlong videos" or "equirectangular photos"? I don't even know the best terminology :)
Unfortunately I don't think there's really a lot out there, but I haven't been searching recently - I would try 360 or VR video. For still images there are quite a few good HDRI maps available from hdrihaven for example - basically high resolution photos with high dynamic range and spherical projection that are used a lot in 3D lighting.
As an example of what camera projection can be used for, looped corridor in this video I made entirely using only camera projection in Fusion ruclips.net/video/U67zNTnE-6g/видео.html
i have dragged in an image in to my composition, but for some reason it can't fill the entire canvas, and i have no idea how to get the image to snap to all edges. now i have black bars all around. here's a link to the image: imgur.com/gallery/IbzW0z1. do you have any ideas on how i might fix this? i posted on blackmagic forum but the moderators still have not published my question after many days. thanks.
Not sure what you are trying, but first you have merged it over a black background, so there the black boundary is intentional? If not, why the merge? Then it depends on the camera position. Bring the camera closer or work with the focal length. But you are working in 3d so not sure what "snap to edges" means in 3d space...
@@VFXstudy oh no the black edges are definitely not intentional. I didn't know that merging would produce that. I just want the image alone without black edges. How else would you put the image in without merging?
Bernd the best of Fusion ....15 mn of videos and 60 mn to apply ....Thanks a lot
Finally i find some good explanation of the catcher node. Thanks. 🙏🙏🙏
Hi Bernd! Not just this tutorial, but all your other VFXstudy tutorials on Fusion are the best! I love your in-depth and CLEAR explanations of the principles behind what you're doing. I believe this is the best way we can all learn -- understanding the why and how so that we can apply that knowledge to our own ideas with our own tweaks. And you're always covering multiple useful functions and tips in each tutorial! I come to your tutorials to learn not just the stated topic, but how to use Resolve/Fusion effectively in general. Please keep up the wonderful tutorials!
Thank you. Coming from Nuke to Fusion, and trying to do some digital matte painting, this tutorial was so very helpful, especially the last bit about the projection IDs and the catcher!! So perfect!
What a great video, thank you! I agree with others, not only this video, but all of your content is extremely informative... you cover highly nuanced yet essential fine points... thank you again!👍💯
Well explained. This gave many new ideas. :)
Thx a million, Bernd, for this "little" (!) roundtrip. Definitely the most comprehensive and best explained tutorial I have seen on that topic so far!
Thanks for Image Plane vs Projectors Tutorial much appreciated!
Exactly what I was looking for. 1000 thanks. One of the best DaVinci Resolve and Fusion channels. …🤗🙌👌👍
Very very useful tutorial, this tecnique realy helps to make the shots look better, thanks man!
Excellent tutorial, Bernd, thank you. I noticed in the Fusion chapter of the Resolve manual there's another kind of projection, using a UVMap Node. The manual only gives it half a page and I don't really understand what I need to do to use it, but the result looks very interesting - creating a 3D scene out of 2D images in such a way that they interact with a 3d model, in terms of reflections and shadows. I would love to see a tutorial on this.
Thank you, very useful tutorial.
Fascinating!
I think so too :)
Hi, Very useful. As usual ! Thank you
really great information
I was thinking about this just this week. I think you are a prophet.
Hihi, so we were both thinking of the same 😏
@@VFXstudy yeah
a very good job, for many if we would like to learn all these techniques, it enriches our knowledge, thank you for sharing it. 👌👌
Many thanks
Bro you saved me, thanks for your lessons!
Fantastic stuff. Great tutorial like always!
Thanks a lot :-)
you are a hero, perfect explaination
Man your vids are so good. as soon as i found out u got a free course out there i jumped on it! Really appreciate what u do
Thanks a lot
Great info on fusions catcher node the info out there is almost inscrutable to beginners. The only real common example is just text describing projecting buildings with windows cut outs.
Thank you! Can’t wait to see more of your videos.
so simple but effective.
Great video! very helpful! Thank you!
Great
Excellent. Thankyou.
You are welcome!
Thank you for this great video ;-)
Very Good, exactly what I was looking for
Thank you.
You are the best!
Thanks for the education!
Great tutorial! One question: How do you create a mask on an image plane 3D?
How about masking the 2d image and then attach the masked version to the 3D image plane?
You are amazing! Tnx mate for sharing so much! Thanks
Excellent!
What about 2D text, how do I add that to an image and have just the text and the reset of the plane transparent?
Texted directly generated 2d text. Or maybe you want a flat text in 3d, but that's what texted does by default. Normally there's no need for image plane or projection if all you need is 2d text
Is there a way to snap a plane (or any other object) to a certain tracking point? Because that’s the only way to be truly accurate and Fusion is not very friendly when it comes to orthogonal views. So if like to select a tracker and have an object being snapped on there.
Not sure I understand you right... If you want to place something in 3D it works with an accurate 3d camera track. In 2d you have point and planar tracking of course but then typically you don't need the full 3d scene.. of course you can build one and do 2d matchmove after the render. If you want a specific point from 3d accessible in 2d you have the locator tool for that.
Thanks for sharing. How could one apply different images on each side of the cube?
Check the Cube3d tool. It does exactly that. For arbitrary geometry you could still project from different angles
@@VFXstudy Thanks a lot. Great tool this Cube3d.
Can we use 360 images as set extension and place our charactor shot on chromakey ???
4:05 mind-blowing... QUESTION: Would this be a simpler/easier/better way in which to create a 3d environment, as opposed to bringing in an HDRI converted to .exr file?
Well, what would you do with the HDRI - connect to a sphere? So basically the same concept and would allow your camera to tilt up and down as well. The added dinmic range in HDRI may or may not be useful depending on your workflow. Often HDRIs are used for lighting, but could be used like this here as well I suppose. But either way this works for a nodal pan/tilt as long as the camera is fixed on a tripod - if the camera moves in any other way, it might fall apart - well depending on the amount of movement.
Thanks for the explanation.
By the way may I ask the brand and model of your keyboard and mouse? thank you very much 🙂
Sure, it's the logitech K780 keyboard and a mojo silent mouse. The keyboard has basically all keys + number block but is a bit more compact and I can put up my phone = camera remote on it. The mouse is vertical which I like for ergonomical reasons and not too noisy on recordings. ... and unfortunately, nobody is paying me for saying all this :P
😂😂😂 you should start an ergonomic hardware review series and contact logitech and other brands, maybe you can get a sponsor 🙂
@@leecaste I thought of developing a special keyboard for VFX artists which I can sell for 1000$... like the Resolve Edit keyboard... but with a special MERGE button 😋😋😋
😂😂😂😂
Great stuff - When working in projector mode with a boris effect imported (e.g. explosion) is there a way to alpha the black background without the explosion getting transparrent? (I want to try this in VR360 instead of imageplane, but can't seem to find the right setting).
Not sure about your plugin, but sometimes you can lumakey vfx that have black background. If you get an image with transparency you can project it with texture projection.
Спасибо!
thanks
Is there a node in the Fusion that can combine multiple images into one big picture? What is the biggest image dimensions that Fusion can accept and use?
You can just merge the images with the standard merge node and transform them smaller or bigger either in the merge or with additional transform node before. Resolution is normally not an issue, unless you run into performance issues due to extreme sizes. I think in theory you can use images up to 16K by 16K
@@VFXstudy Thank you....I have never thought about a Merge node! Will give it a try!😀
Hi , can we create a 3d background to look like Patriot Act with Hasan Minhaj scenario? My background is white and not Chroma Key.
Well, you can create all kinds of backgrounds if you have the material and artistic vision... I don't remember the Patriot Act - though I've seen it some time ago. You can use a Luma key to key on white, but depending on the foreground your results may vary. With black hair, dark clothes and good lighting separation, keying against white might actually work reasonably well - in other scenarios not so much...
Hi and thank you for such clear and informative tutorials.
At about 9:44 or so you speak of rotating the plane and getting a 3D perspective. However, in the free version of Resolve, the Render3D node will only display the straight ahead view of a photo and not the rotated perspective. It appears that rotating the image via the plane or Merge3D only results in the Render3D displaying a more narrow photo still in the original orientation.
This is clearly evident if you use a photo of a bird with it's wings spread. Is there some manner of rendering the rotated image?
Thank you!
Hi, the Render3D renders the scene exactly as the camera3d looks onto it. If you rotate the merge3D everything including the attached camera will rotate -> no effect. You need to rotate the imageplane3d node alone. If it appears squeezed or rotated depends on the perspective / angle of rotation.
@@VFXstudy Yes. Quite right. My Render3D node had become stuck for some reason. I reset it and I now get the expected results, just as you describe them. And the effect is quite different if I rotate an ImagePlane3D, which maintains aspect ratio and simulates looking at a photo from the edge, or if I rotate a Shape3D plane, which allows the image to distort. Very interesting. Thanks again!
trying to get camera projection image plane of a keyed subject cast a shadow and it doesnt seem to do it.. any tips?
Should be possible, but few things to check: projection needs to be texture projection (not light or ambient light). Shadows enabled in Render node, light source supports shadows (spotlight). Used to only work with software renderer. Now OpenGl renderer got enhanced, but not not sure - haven't tried this one since.
Also check my channel for the tutorial on shadow projection
Is all the stuff you did in this tutorial in the free version or studio version?
Yes you can do it all in the free version of DaVinci Resolve
0:20 confusion...in fusion lol
nice lesson
Sphere would be great.for green screen work!
Hello, my camera is not centering on the photo how do I properly align.
Either move it around in 3D space or better make sure plane and camera start from the default locations and then only move the camera back in z. Unless you are talking about projection. Then make sure your projector is aligned with the camera - often on frame 0 or on last frame. You might start by copying the camera to use it as projector and later might remove the keyframes on the projector.
hey friend - where are good resources to download free spheres or 360 videos? do I search "latlong videos" or "equirectangular photos"? I don't even know the best terminology :)
Unfortunately I don't think there's really a lot out there, but I haven't been searching recently - I would try 360 or VR video. For still images there are quite a few good HDRI maps available from hdrihaven for example - basically high resolution photos with high dynamic range and spherical projection that are used a lot in 3D lighting.
Why in my fusion16.2 I don't see settings of camera like "Enable camera projection”?
Should be in the Inspector, after you connect an image to the camera there should be a projector tab
CCCCCON-FUSION
???
No that's another channel... 😏
As an example of what camera projection can be used for, looped corridor in this video I made entirely using only camera projection in Fusion
ruclips.net/video/U67zNTnE-6g/видео.html
i have dragged in an image in to my composition, but for some reason it can't fill the entire canvas, and i have no idea how to get the image to snap to all edges. now i have black bars all around. here's a link to the image: imgur.com/gallery/IbzW0z1. do you have any ideas on how i might fix this? i posted on blackmagic forum but the moderators still have not published my question after many days. thanks.
Not sure what you are trying, but first you have merged it over a black background, so there the black boundary is intentional? If not, why the merge? Then it depends on the camera position. Bring the camera closer or work with the focal length.
But you are working in 3d so not sure what "snap to edges" means in 3d space...
@@VFXstudy oh no the black edges are definitely not intentional. I didn't know that merging would produce that. I just want the image alone without black edges. How else would you put the image in without merging?
@@evenaicantfigurethisout you are merging it over a background before attaching to a 3d object. Why not attach the image directly?
@@VFXstudy thank you! I deleted the merge and background as you suggested and it works now.
Mach deutschsprachige tuts
Das kann man ja nicht anhören😅
Mache ich auf tutorial-experts.de aber VFXstudy ist komplett auf English, damit der rest der Welt es sich anhören kann.
Thanks