Hot answers tagged

18

You can use border render, like explained in the wiki : You can render just a portion of the view instead of rendering the entire frame. While in Camera View, enable Border and press CtrlB, then drag a rectangle to define the area you want to render. CtrlAltB is the shortcut to disable the border. Note that this disables the Save Buffers option in ...


12

Blender does not provide a way for multiple uses to edit the same blend file at exactly the same time. * However typically you can manage this by splitting your assets into separate blend files and linking from one file to another. Blender supports recursive linking so a scene, node-groups, animation, models, materials etc can all have their own files. ...


11

While anyone can make use of network rendering, the main target is for small studios. Which would be the reason this design choice was made. Imagine a studio with 10 artists, you setup a master and several slaves for the render farm. Each artist is then a client that sends render jobs to the farm, the master keeps track of all these jobs and which slaves ...


11

Blender's network renderer allows you to link to other PCs -- Apple, Windows, Linux -- and distribute your render job to many computers... speeding up the job overall... It is a job management tool to make rendering many frames faster, by spreading the work across multiple platforms. Here's a good video on using network render and in setting up a render ...


9

In short: no. But I imagine that one could work on such functionality again. There is a Verse 2 being developed, with a better Python API. This could be a good next candidate, unfortunately this version is far from complete yet. Certainly an effort on the current version of Verse could be started again.


7

The basic idea is to open a socket, poll the data and apply the received values to the shape keys: import socket # open network socket port = 13001 s = socket.socket(socket.AF_INET, socket.SOCK_DGRAM) s.bind(("127.0.0.1", port)) print "waiting on port:", port # poll data (could use a timer) data, addr = s.recvfrom(1024) print data # set float value from ...


7

Not directly possible, Blender needs its data to be local (it does not understand uri, FTP, etc.). You could set up a 'fake' local file system on your students' machine, though, which would actually be a network one (like a samba/windows sharing, or other similar solution). It will work as long as Blender can see it as a local directory.


7

Network Render plugin does not have this functionality. Render the scene remotely via the command line. You can run your own script with it to set proper rendering device or render tile-size. blender -b file.blend -E CYCLES -t 0 -o //file -P script.py In the script.py: import bpy bpy.context.user_preferences.addons['cycles'].preferences.compute_device_type ...


7

You could install BURP on one or more machines, and the BOINC client on the rest. That way you don't have to worry too much about the composition of your farm. This is what we also use on Renderfarm.fi as our backend. BURP: http://development.burp.renderfarming.net/projects/burp/wiki/How_to_download_the_code


7

I'm not sure if a render farm currently exists to do this, but theoretically it's easy. Simply use a method called Image Stacking, where you render the same image multiple times with different seeds and then mix them together to average out the noise. The first two images would mix 50/50, the third would mix into that by 33.3%, the fourth 25%, 5th 20%... ...


6

Polosson's answer is quite right. I just want to make it more easier to understand particularly for your case, supposing you would have trouble to understand that.


5

I would recommend using the network render add-on. It's since Blender is cross-compatible, they should all work together fairly well(I've made it work between Mac and Windows). The docs are here: http://wiki.blender.org/index.php/Doc:2.6/Manual/Render/Performance/Netrender Basically to sum it up, you'll have to: Enable the Network Render Addon Create a ...


5

AFAIK netrender doesn't support this, as there is no opengl "render engine" defined in blender (as you noted). There is an opengl render operator, but it is not defined as a formal render engine. However there is a simple alternative approach you might be able to use. See the following questions: Render on more than one computer? Cycles Netrender Returning ...


5

Under Cloth Cache, if you enable Disk Cache, blender will create a folder holding the cloth simulation data, without this no simulation data can be used while rendering. This cache folder needs to be copied to each computer that will be rendering. The name of the folder should be the blend filename prefixed with blendcache_ and be in the same folder as the ...


4

Yes, but all the networked machines will need access to the cache. Another way to do a network render: All machines need access to the blendfile and any external data (textures, fluid sim caches, etc.) Make all the machines output the rendered images to a shared directory Disable Overwrite in Render Settings > Output for all machines. The Overwrite option ...


4

No that isn't an available feature. The network render only supports rendering final images. The best you can do to speedup viewport rendering is to use the machine with the best GPU that can be used by cycles. Other options include using border render to only render the part of the scene you are most interested in.


4

That will depend on the task and the level of effort you want to put into it. If you are rendering an animation then you will easily benefit from using multiple machines, each machine can render a single frame at the same time giving a very linear speedup based on number of machines. There is a network render addon included with blender. You may want to ...


4

using Mike Pan answer i have tried to split the code and put the networking part on a different thread see if this helps with the speed import bpy import urllib import mathutils from bpy import context from urllib import request, parse from math import sin, cos, radians from mathutils import Matrix import time import threading class Worker(threading....


4

Sheep-It! is a render farm service that is 100% free and works by using other people's computers to render your projects. It's a great service, and I would definitely recommend it. The only limitation with Sheep-It! is the 500mb project max size. If you don't want to or can't use Sheep-It!, I would recommend using this method: Using a cloud storage client ...


4

Changes have been made in the Blender API. You can retrieve 'compute_device_type' and 'devices' via the cycles addon preferences prefs = bpy.context.user_preferences.addons['cycles'].preferences print(prefs.compute_device_type) for d in prefs.devices: print(d.name)


4

I do not recommend you to render directly to video. If any error occurs during this render, you will lost all the movie. Rendering in frames (stills) you can render in parts, also in case of error, you can render the remaining frames later. Use the video editor to render the stills to the final video.


3

I have used DrQueue before. It isn't the easiest to use, but it is quite powerful, from what I understand. The guy who built the IKEA cabinet render farm used DrQueue, so it's got to be good, right? I believe the Blender Foundation uses DrQueue for at least one of their render farms- but don't quote me on that.


3

The latest version of MobaXterm works great with Blender! I just had to check the "Hardware OpenGL acceleration" in order to make it work (under "Settings" --> "Configuration" --> "X11" tab. I started a SSH connection with compression enabled in order to lower bandwidth usage and started Blender which now works correctly. I am running the latest free ...


3

The issue isn't with Blender. It's with the command you're running. Using the open command is fine for opening applications, but it doesn't allow you to pass arguments to the application (as far as I know). You want to call the executable directly with: /Applications/Blender/blender.app/Contents/MacOS/blender "/Users/schultz/Desktop/Blender/slave.blend" --...


3

a possible workaround would be use dropbox or have the source .blend on the network file servers. Open on different machines and use the Star Frame: & Frame Skip: settings in render panel. Computer#1: Star Frame 1: & Frame Skip: 5 Computer#2: Star Frame 2: & Frame Skip: 5 Computer#3: Star Frame 3: & Frame Skip: 5 Computer#4: Star Frame 4: &...


3

If multiple people aren't working on the same object, you could separate one bit and Link it into the original file. So one guy works on the main file, and other people work on the Linked files. Or you could separate everything and have a master-file that is not edited directly, but simply has all the separate pieces linked in and placed correctly. As far ...


3

I am not good in programming so I did it this way... I install VM whit 8 cores Hight cpu 2,6ghz xeon whit 8gb ram... 50 gb HDD and Windows server 2012 .... Its exatly 274 dollars... On your free trial you have free 300 dollars... What is amazing... After install VM , you can log in using Remote Desktom Connection... Open Internet Explorer, download Blender ...


3

Since neither Blender nor Python deal with threading very well.. one option is to use async-IO to handle the network socket, and hook bpy.app.handlers.scene_update_post to have your script called periodically. When your handler is called, you would both process async events on the network sockets, and perform whatever operations were requested by the ...


3

Convert the vector to list picklablePosition = list(owner.worldPosition) and back unpickledPositionAsList = [UPData[0],UPData[1],UPData[2]] owner.worldPosition = unpickledPositionAsList # worldPosition accepts a 3-tuple list


3

Usually with exr files one wishes to use extra data stored in them (various render passes, depth buffer information, etc.) to apply the finishing touches to the output after the render is complete. Even if this is not what you have in mind, the same process should serve to simply convert the exr files into pngs. Import the sequence into the compositor (you ...


Only top voted, non community-wiki answers of a minimum length are eligible