I was talking about prety much what's being discussed in the "mask from position pass methods" thread... To have the possibility to do Worldspace cropping. Because if you use a holdout everything behind will be gone. But maybe my example was a bad one ;) What would be really cool actually would be to be able to use deep images as some kind of 3 dimensional masks. Imagine using a volumetric as mask input for a DeepCC for example. But i think due to the samples having a specific depth value this won't be possible, as if you have a sample in your deepmask that doesn't fit the depth in your deepimage that won't affect it. Hmm that sentence is weird, hope it's comprehensible anyway ;)
cheers Patrick ----- Original Message ----- From: michaeld...@gmail.com To: nuke-users@support.thefoundry.co.uk Date: 21.11.2012 07:45:08 Subject: Re: Re: [Nuke-users] deep position in DeepExpression? > I'll have to give this a look - I would have dismissed it purely because > you're only getting the front-most sample promoted to deep. I agree > it would be great to have user friendly tools working with deep position, > either derived in Nuke from the camera+deep opacity (like DepthToPosition > in Nuke 7), or rendered separately as a deep aov. Or some other method I'm > not taking into account. > > > On 20 November 2012 06:23, Denis SCOLAN <deni...@googlemail.com> wrote: > >> Well you don't need the world pass to cut your smoke with primitive >> geometry (white texture in RGBA). >> Just render your primitive with a scanline render and then put it into a >> deep image stream with the deepfromimage. >> Use this to holdout your smoke. You can combine as many sources as you >> want as long as they have a correct ZDepth pass so you can convert them >> into deep image streams. >> >> Cheers >> Denis >> >> >> >> >> 2012/11/20 Patrick Heinen <mailingli...@patrickheinen.com> >> >>> Would still be usefull to have a deep position pass. To be able to cut >>> out certain areas of smoke with geometry primitives for example. >>> >>> cheers >>> Patrick >>> >>> ----- Original Message ----- >>> From: fr...@beingfrank.info >>> To: nuke-users@support.thefoundry.co.uk >>> Date: 20.11.2012 05:15:19 >>> Subject: Re: [Nuke-users] deep position in DeepExpression? >>> >>> >>> > need a DeepToBrain node it would seem >>> > >>> > >>> > On 20/11/12 5:11 PM, Michael Garrett wrote: >>> >> Hee hee ;) >>> >> >>> >> >>> >> On 19 November 2012 20:09, Frank Rueter <fr...@beingfrank.info >>> >> <mailto:fr...@beingfrank.info>> wrote: >>> >> >>> >> I'm such a retard. Of course I'm on Nuke 7 and I have been using >>> >> the DepthToPosition node. >>> >> Just had a complete brain freeze, hahaha. >>> >> >>> >> Sorted, thanks Michael! :-D >>> >> >>> >> >>> >> On 20/11/12 4:59 PM, Michael Garrett wrote: >>> >>> I assuming you're not using the Nuke 7 public beta for this, >>> >>> because there is a "DepthToPosition" node in there! Otherwise, >>> >>> try this. I'm using Topnode for the camera input to grab the >>> >>> projection knobs. >>> >>> >>> >>> >>> >>> Group { >>> >>> inputs 2 >>> >>> name ZtoPworld >>> >>> help "Creates Pworld point data from a raw depth channel and an >>> >>> input camera." >>> >>> selected true >>> >>> xpos -201 >>> >>> ypos 45 >>> >>> addUserKnob {20 User} >>> >>> addUserKnob {41 from0 l "input z channel" T Copy1.from0} >>> >>> } >>> >>> Input { >>> >>> inputs 0 >>> >>> name Input >>> >>> xpos 715 >>> >>> ypos -750 >>> >>> } >>> >>> set N219b4e80 [stack 0] >>> >>> push $N219b4e80 >>> >>> Copy { >>> >>> inputs 2 >>> >>> from0 depth.Z >>> >>> to0 rgba.blue >>> >>> name Copy1 >>> >>> selected true >>> >>> xpos 715 >>> >>> ypos -715 >>> >>> } >>> >>> Expression { >>> >>> expr0 cx >>> >>> expr1 cy*(input.width/input.height)*input.pixel_aspect >>> >>> name Expression7 >>> >>> label NDC >>> >>> xpos 715 >>> >>> ypos -660 >>> >>> } >>> >>> Expression { >>> >>> temp_name0 h_tan >>> >>> temp_expr0 (haperture/2)/focal >>> >>> temp_name1 v_tan >>> >>> temp_expr1 (vaperture/2)/focal >>> >>> expr0 b*(r*h_tan) >>> >>> expr1 b*(g*v_tan) >>> >>> name Expression8 >>> >>> label "NDC to Pcam" >>> >>> xpos 715 >>> >>> ypos -607 >>> >>> addUserKnob {20 User} >>> >>> addUserKnob {7 focal} >>> >>> focal {{"\[topnode parent.input1].focal"}} >>> >>> addUserKnob {7 haperture} >>> >>> haperture {{"\[topnode parent.input1].haperture"}} >>> >>> addUserKnob {7 vaperture} >>> >>> vaperture {{"\[topnode parent.input1].vaperture"}} >>> >>> } >>> >>> Multiply { >>> >>> channels {-rgba.red -rgba.green rgba.blue} >>> >>> value -1 >>> >>> name Multiply3 >>> >>> xpos 715 >>> >>> ypos -530 >>> >>> } >>> >>> ColorMatrix { >>> >>> matrix { >>> >>> {{Axis1.world_matrix.0 i} {Axis1.world_matrix.1 i} >>> >>> {Axis1.world_matrix.2 i}} >>> >>> {{Axis1.world_matrix.4 i} {Axis1.world_matrix.5 i} >>> >>> {Axis1.world_matrix.6 i}} >>> >>> {{Axis1.world_matrix.8 i} {Axis1.world_matrix.9 i} >>> >>> {Axis1.world_matrix.10 i}} >>> >>> } >>> >>> name ColorMatrix3 >>> >>> label "Inverse Cam rotation" >>> >>> xpos 715 >>> >>> ypos -492 >>> >>> } >>> >>> Add { >>> >>> channels rgb >>> >>> value {{Axis1.world_matrix.3 i} {Axis1.world_matrix.7 i} >>> >>> {Axis1.world_matrix.11 i} {curve i}} >>> >>> name Add3 >>> >>> label "World minus Camera pos" >>> >>> xpos 715 >>> >>> ypos -441 >>> >>> } >>> >>> set N525d9a0 [stack 0] >>> >>> Output { >>> >>> name Output1 >>> >>> xpos 715 >>> >>> ypos -346 >>> >>> } >>> >>> Input { >>> >>> inputs 0 >>> >>> name cam >>> >>> xpos 498 >>> >>> ypos -684 >>> >>> number 1 >>> >>> } >>> >>> Axis2 { >>> >>> display off >>> >>> selectable false >>> >>> name Axis1 >>> >>> xpos 508 >>> >>> ypos -637 >>> >>> } >>> >>> push $N525d9a0 >>> >>> Viewer { >>> >>> frame 10 >>> >>> input_process false >>> >>> name Viewer1 >>> >>> xpos 959 >>> >>> ypos -619 >>> >>> } >>> >>> end_group >>> >>> >>> >>> >>> >>> >>> >>> On 19 November 2012 19:45, Frank Rueter <fr...@beingfrank.info >>> >>> <mailto:fr...@beingfrank.info>> wrote: >>> >>> >>> >>> yes, that's how I can do it and probably will (happy to use >>> >>> your setup in stead of reinventing the wheel). >>> >>> I did send in a feature request to have a DeepToPosition node >>> >>> though (though it should probably be a "DepthToPosition" node >>> >>> seeing you don't need deep for this) >>> >>> >>> >>> >>> >>> >>> >>> On 20/11/12 4:39 PM, Michael Garrett wrote: >>> >>>> If you're just trying to create a regular "flat" position >>> >>>> pass, can't you just flatten deep.front and then use a >>> >>>> regular Nuke Expression node plus the Camera matrix to >>> >>>> convert that to position? I can send you the setup if you >>> >>>> need it. >>> >>>> >>> >>>> I've gone over the expressions required to convert a deep >>> >>>> opacity file to deep position (position render as deep data) >>> >>>> and as we know, the DeepExpression node doesn't support some >>> >>>> things yet. >>> >>>> >>> >>>> >>> >>>> On 19 November 2012 01:27, Patrick Heinen >>> >>>> <mailingli...@patrickheinen.com >>> >>>> <mailto:mailingli...@patrickheinen.com>> wrote: >>> >>>> >>> >>>> Ok, i thought you wanted to use the DeepExpression node >>> >>>> to calculate those xyz coordinates, which as you say >>> >>>> would be possible, but not comfortable. Yes Deep is >>> >>>> always camera relative, that's the drawback with it. But >>> >>>> as the DeepToPoints node exists, you're right that it >>> >>>> shouldn't be to complicated to write a deep to point >>> >>>> position node. >>> >>>> +1 for the feature request, maybe it makes it into the >>> >>>> releas *hope* ;) >>> >>>> >>> >>>> cheers, >>> >>>> Patrick >>> >>>> >>> >>>> ----- Original Message ----- >>> >>>> From: fr...@beingfrank.info <mailto: >>> fr...@beingfrank.info> >>> >>>> To: nuke-users@support.thefoundry.co.uk >>> >>>> <mailto:nuke-users@support.thefoundry.co.uk> >>> >>>> Date: 19.11.2012 02:42:30 >>> >>>> Subject: Re: [Nuke-users] deep position in >>> DeepExpression? >>> >>>> >>> >>>> >>> >>>> > Hi Patrick, >>> >>>> > >>> >>>> > x and y only give you screen space positions, not >>> >>>> world positions. >>> >>>> > With the camera and depth available obviously this can >>> >>>> be converted to >>> >>>> > world position, but I'd really like to see a built in >>> >>>> solution. >>> >>>> > Basically like the DeepToPoints but a DeepToPosition. >>> >>>> > >>> >>>> > The brute force way would be to render a full >>> >>>> resolution DeepToPoint >>> >>>> > through the ScanlineRender node and generate positon >>> >>>> vectors in it. But >>> >>>> > that is stupidly slow and really shouldn't be >>> necessary. >>> >>>> > >>> >>>> > Will send in a feature request. >>> >>>> > >>> >>>> > Cheers, >>> >>>> > frank >>> >>>> > >>> >>>> > On 19/11/12 11:41 AM, Patrick Heinen wrote: >>> >>>> >> Hi Frank, >>> >>>> >> >>> >>>> >> you should be able to acces x and y with x and y. I >>> >>>> was at least able to produce a uv map, if that's the >>> >>>> kind of funcionality you're somehow looking for. However >>> >>>> the Deep Expression Node is not working exactly as the >>> >>>> Expression Node. This came up on the mailinglist and >>> >>>> there is a feature request logged for expanding the deep >>> >>>> expression node to match the functionality of the >>> >>>> expression node #27372. >>> >>>> >> Tell me if you succeed in your attempt of creating a >>> >>>> world position pass, I would really be interested in >>> that. >>> >>>> >> >>> >>>> >> cheers >>> >>>> >> Patrick >>> >>>> >> Am 18.11.2012 um 21:58 schrieb Frank Rueter: >>> >>>> >> >>> >>>> >>> Is it possible to get a pixel's position in a >>> >>>> DeepExrpression node? I am trying to create a world >>> >>>> position pass from the image and the associated camera. >>> >>>> >>> >>> >>>> >>> Cheers, >>> >>>> >>> frank >>> >>>> >>> _______________________________________________ >>> >>>> >>> Nuke-users mailing list >>> >>>> >>> Nuke-users@support.thefoundry.co.uk >>> >>>> <mailto:Nuke-users@support.thefoundry.co.uk>, >>> >>>> http://forums.thefoundry.co.uk/ >>> >>>> >>> >>> >>>> >>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >>> >>>> >> _______________________________________________ >>> >>>> >> Nuke-users mailing list >>> >>>> >> Nuke-users@support.thefoundry.co.uk >>> >>>> <mailto:Nuke-users@support.thefoundry.co.uk>, >>> >>>> http://forums.thefoundry.co.uk/ >>> >>>> >> >>> >>>> >>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >>> >>>> >> >>> >>>> >> >>> >>>> > >>> >>>> > _______________________________________________ >>> >>>> > Nuke-users mailing list >>> >>>> > Nuke-users@support.thefoundry.co.uk >>> >>>> <mailto:Nuke-users@support.thefoundry.co.uk>, >>> >>>> http://forums.thefoundry.co.uk/ >>> >>>> > >>> >>>> >>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >>> >>>> >>> >>>> _______________________________________________ >>> >>>> Nuke-users mailing list >>> >>>> Nuke-users@support.thefoundry.co.uk >>> >>>> <mailto:Nuke-users@support.thefoundry.co.uk>, >>> >>>> http://forums.thefoundry.co.uk/ >>> >>>> >>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >>> >>>> >>> >>>> >>> >>>> >>> >>>> >>> >>>> _______________________________________________ >>> >>>> Nuke-users mailing list >>> >>>> Nuke-users@support.thefoundry.co.uk <mailto: >>> Nuke-users@support.thefoundry.co.uk>,http://forums.thefoundry.co.uk/ >>> >>>> >>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >>> >>> >>> >>> >>> >>> _______________________________________________ >>> >>> Nuke-users mailing list >>> >>> Nuke-users@support.thefoundry.co.uk >>> >>> <mailto:Nuke-users@support.thefoundry.co.uk>, >>> >>> http://forums.thefoundry.co.uk/ >>> >>> >>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> _______________________________________________ >>> >>> Nuke-users mailing list >>> >>> Nuke-users@support.thefoundry.co.uk <mailto: >>> Nuke-users@support.thefoundry.co.uk>,http://forums.thefoundry.co.uk/ >>> >>> >>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >>> >> >>> >> >>> >> _______________________________________________ >>> >> Nuke-users mailing list >>> >> Nuke-users@support.thefoundry.co.uk >>> >> <mailto:Nuke-users@support.thefoundry.co.uk>, >>> >> http://forums.thefoundry.co.uk/ >>> >> >>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >>> >> >>> >> >>> >> >>> >> >>> >> _______________________________________________ >>> >> Nuke-users mailing list >>> >> Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ >>> >> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >>> > >>> > _______________________________________________ >>> > Nuke-users mailing list >>> > Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ >>> > http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >>> _______________________________________________ >>> Nuke-users mailing list >>> Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ >>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >>> >> >> >> _______________________________________________ >> Nuke-users mailing list >> Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ >> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users >> > _______________________________________________ > Nuke-users mailing list > Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ > http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users _______________________________________________ Nuke-users mailing list Nuke-users@support.thefoundry.co.uk, http://forums.thefoundry.co.uk/ http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
