You're right about the logs on successfully rendered files, but at least in
Qube (and most render managers that I know of) you still have access to
which node rendered which frame. So a bit of digging around should suffice
even if the manager flags that frame as successful.

Metadata for sure is always nice, and if TIFF files wasn't a mandatory
thing in this case using EXR files would allow for a easy fix for this
problem.

Personally I would put TIFF files on the same basket as QuickTime files and
burn them both with a smile on my face :)

But take a look at python's platform and getpass modules. Perhaps you could
use them to retrieve the host and username and then pass them to the
exiftool.

def getHostname():
    import platform
    return platform.node()

def getUsername():
    import getpass
    return getpass.getuser()

cheers,
diogo


On Wed, Jan 30, 2013 at 7:51 AM, John RA Benson <
[email protected]> wrote:

>  Yes, these are all pretty clever ideas!
>
> As for the logs, uh, no. Succesfully completing jobs don't leave a trail,
> unless you ask to keep the log before rendering. At least that's for logs I
> can get at. So that frame that took an hour still rendered successfully
> and... It's also not reasonable to request changing the render manager at a
> large facility.
>
> Besides, it's a lot nicer to look at any file and see info on it. I love
> exrheader. More information is better when you need it.
>
> I have to say, it's funny with a format that's been around for ages and is
> called "Tagged"... it's not straightforward in nuke to add metadata info to
> the header.
>
> aside from all that - I just found a possibly simpler method that would
> need to be done as an afterFrameRender. Might be useful for someone else,
> so:
>
> http://www.sno.phy.queensu.ca/~phil/exiftool/
> ./exiftool -HostComputer=$HOST -artist=$USER tagged.tiff
>
> turns out the tags are very specific, like HostComputer (works) vs Host
> (does not). Tried them with nuke's modifyMetaData node, but still no luck.
>
> Thanks for the cool ideas!
> JRAB
>
>
> On 01/29/2013 08:02 PM, Nico Dufort wrote:
>
> While I am enjoying the microdots-spy approach to tag an image with secret
> render host information, am I missing something or you could just check
> your render logs?   Or... you don't have logs? =P
>
>
> On Tue, Jan 29, 2013 at 10:50 AM, chris <[email protected]> wrote:
>
>> or extending on that idea...
>> since a single black pixel could be noticable in the middle of bright
>> colored pixels, you could duplicate the pixel color next to it (or even
>> average the pixels around it), and add an offset with the hostname value.
>> this would probably be invisible even by the most critical eyes :)
>>
>> example if the pixel on the bottom left [0,0] is to hold the machine ID
>> data:
>> - average pixels [0,1], [1,1] and [1,0] to calculate the [0,0] color value
>> - add tiny offset color value representing the machine ID to that pixel
>> [0,0]
>>
>> obviously you'd want a second script which automatically calculates the
>> machine ID from the 4 pixels on the bottom left.
>>
>> there's probably an easier way that i'm not aware of though ;)
>> ++ chris
>>
>>
>>
>>
>> On 1/29/13 at 7:27 PM, [email protected] (chris) wrote:
>>
>>  hmm, or if you can spare a single pixel somewhere, you could just burn
>>> in the hostname as color (or alpha) value...
>>> like in the bottom right corner, each render machine gets a slightly
>>> different shade of black (ie render1 is 0.00001, render2 is 0.00002 etc)
>>>
>>> then if you have corrupt renders you could check the pixel value and see
>>> where it came from.
>>>
>>> ++ chris
>>>
>>>
>> _______________________________________________
>> Nuke-users mailing list
>> [email protected], http://forums.thefoundry.co.uk/
>> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
>>
>
>
>
>  --
> "Attention, attention. Here and now, boys," the mynah repeated. "Here and
> now, boys."
>
>
> _______________________________________________
> Nuke-users mailing [email protected], 
> http://forums.thefoundry.co.uk/http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
>
>
> _______________________________________________
> Nuke-users mailing list
> [email protected], http://forums.thefoundry.co.uk/
> http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users
>
_______________________________________________
Nuke-users mailing list
[email protected], http://forums.thefoundry.co.uk/
http://support.thefoundry.co.uk/cgi-bin/mailman/listinfo/nuke-users

Reply via email to