If anyone is interested, I'll follow up on our solution that we ended up
with.

Our developer created two Ops to do the transformation to other
projections: a node that would accept multiple inputs and do the
transformation heavy lifting which then stored the results in channels, and
another node that allowed the user to extract a certain 'view' from the
channels it created, using metadata generated by the transformation node.
 I then realized we could use Nuke's built-in per-view parameter overrides
to store the results in the views we wanted in a relatively native manner.
 In the end, with some gizmos to automate the view merging/splitting this
solution was pretty easy and intuitive for the user.

HTH,
-Jon


On Thu, Aug 9, 2012 at 6:42 PM, jon parker <[email protected]> wrote:

> Ah yes, I do see why adding a view is a potentially awkward workflow.
> When a new view is created, there is no global way to tell Nuke not to use
> it by default.  So if I already have views 'inputview[1-6]', when I add
> 'outputview[1-5]', those views will get processed further up the chain.
> That's bad.
>
> Since the node can only output one image at a time, the most
> straightforward thing to do would seem to just have multiple copies for
> each desired output and just change its output mode.  That could also be a
> button somewhere.  This is cleaner to me than stuffing the output into
> layers because I believe layers must all share the same resolution.  I do
> not think this is a problem, however, and might just be complicating
> everything based on a hypothetical edge case. In any case, one could make
> sure in the code that this is not much slower than using multiple nodes.
>
> -Jon
>
>
> On Thu, Aug 9, 2012 at 5:12 PM, Nathan Rusch <[email protected]>wrote:
>
>>   Given your end product, views do sound like they would be useful.
>> However, I’m still of the opinion that adding views as part of the output
>> process (i.e. changing the project settings just to generate data) is a
>> somewhat strange and possibly painful way to go about it.
>>
>> As one other possibility, you could do something like:
>> - Output each piece of the projection into a layer with a qualified name
>> (based on its projection type and component).
>> - Add a button or write a small Python function to automatically break
>> out a set of Write nodes based on the node’s settings.
>>
>> You could even go one step further and add a file knob to your generation
>> node and allow it to be executed directly, handling the file outputs under
>> the hood.
>>
>> Does any of this sound like it could work in the context of your
>> workflow? If not, can you provide any more specifics? It may help to
>> generate some alternate solutions.
>>
>> -Nathan
>>
>>
>>  *From:* jon parker <[email protected]>
>> *Sent:* Thursday, August 09, 2012 1:54 PM
>> *To:* Nuke plug-in development discussion<[email protected]>
>> *Subject:* Re: AW: [Nuke-dev] multiple outputs?
>>
>> I'm starting to think the best way forward is to provide an option to
>> send outputs to a set of named views if they exist, and a user controlled
>> python button to create those views if they don't.  This way the creation
>> of views falls into the hands of the user and is not internal to the node.
>> Win-win?
>>
>> On Wed, Jul 25, 2012 at 4:30 PM, Steve Booth <[email protected]> wrote:
>>
>>> The number of views is a project-level property that is set in the
>>> project-settings property panel.  It doesn't necessarily have to be set
>>> prior to designing the SDG, but when changed will fundamentally alter the
>>> data structures created by the various nodes in your script.
>>>
>>> To my knowledge, there is no way to alter the project definitions from
>>> inside a single node (as it will affect all other nodes in the script),
>>> nor
>>> would it, It think, be a good idea, although it might actually be
>>> possible
>>> by executing some Python code.
>>>
>>> Not advisable, imho.
>>>
>>>
>>> -----Original Message-----
>>> From: [email protected]
>>> [mailto:[email protected]] On Behalf Of jon
>>> parker
>>> Sent: Wednesday, July 25, 2012 1:20 PM
>>> To: Nuke plug-in development discussion
>>> Subject: Re: AW: [Nuke-dev] multiple outputs?
>>>
>>> Okay, thanks for the info.  Do views have to be already defined in order
>>> for
>>> a node to write to them, or can nodes create new views?
>>>
>>> On Wed, Jul 25, 2012 at 9:42 AM, Steve Booth <[email protected]>
>>> wrote:
>>> > Although technically true, realize that in a multi-view environment,
>>> multiple instances of your Op will be instantiated, one for each defined
>>> view, for every use of the base node in the SDG. Utilizing static
>>> storage,
>>> it is possible for your node to be aware, and potentially communicate
>>> with
>>> other versions of itself servicing those other view outputs.
>>> >
>>> > Sent from my iPad
>>> >
>>> > On Jul 25, 2012, at 12:27 AM, Thorsten Kaufmann
>>> <[email protected]> wrote:
>>> >
>>> >> You cannot have multiple outputs, but you can output to as many
>>> layers/channels as needed.
>>> >>
>>> >> Regards,
>>> >> Thorsten
>>> >>
>>> >>
>>> >> Thorsten Kaufmann
>>> >> Head of Production
>>> >> ____________________________________
>>> >>
>>> >> Mackevision Medien Design GmbH
>>> >> Forststraße 7
>>> >> D-70174 Stuttgart
>>> >>
>>> >> T  +49 711 93 30 48 31 <%2B49%20711%2093%2030%2048%2031>
>>> >> F  +49 711 93 30 48 90
>>> >> M +49 151 19 55 55 02 <%2B49%20151%2019%2055%2055%2002>
>>> >>
>>> >> [email protected]
>>> >> http://www.mackevision.de
>>> >>
>>> >> Geschäftsführer: Armin Pohl, Joachim Lincke, Karin Suttheimer HRB
>>> >> 243735 Amtsgericht Stuttgart
>>> >>>> -----Ursprüngliche Nachricht-----
>>> >>>> Von: [email protected] [mailto:nuke-dev-
>>> >>>> [email protected]] Im Auftrag von jon parker
>>> >>>> Gesendet: Mittwoch, 25. Juli 2012 01:05
>>> >>>> An: [email protected]
>>> >>>> Betreff: [Nuke-dev] multiple outputs?
>>> >>>>
>>> >>>> Greetings,
>>> >>>> I'm new to this list (and NDK development) so this might already be
>>> answered.
>>> >>>> Is it possible to create a node that provides multiple outputs?
>>> >>>>
>>> >>>> Cheers,
>>> >>>> Jon
>>> >>>> _______________________________________________
>>> >>>> Nuke-dev mailing list
>>> >>>> [email protected], http://forums.thefoundry.co.uk/
>>> >>>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-dev
>>> >> _______________________________________________
>>> >> Nuke-dev mailing list
>>> >> [email protected], http://forums.thefoundry.co.uk/
>>> >> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-dev
>>> > _______________________________________________
>>> > Nuke-dev mailing list
>>> > [email protected], http://forums.thefoundry.co.uk/
>>> > http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-dev
>>> _______________________________________________
>>> Nuke-dev mailing list
>>> [email protected], http://forums.thefoundry.co.uk/
>>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-dev
>>>
>>> _______________________________________________
>>> Nuke-dev mailing list
>>> [email protected], http://forums.thefoundry.co.uk/
>>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-dev
>>>
>>
>>  ------------------------------
>> _______________________________________________
>> Nuke-dev mailing list
>> [email protected], http://forums.thefoundry.co.uk/
>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-dev
>>
>>
>> _______________________________________________
>> Nuke-dev mailing list
>> [email protected], http://forums.thefoundry.co.uk/
>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-dev
>>
>>
>
_______________________________________________
Nuke-dev mailing list
[email protected], http://forums.thefoundry.co.uk/
http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-dev

Reply via email to