Network Rendering - Simple Explanations?

RNLVideo wrote on 5/22/2005, 4:31 PM
I finally got network rendering working, but I'm having trouble getting my head around part of it. Can someone help with a simplified explanation of when I should / shouldn't check the "Use Final Render Template" option & if so, what templates should I choose for different final formats (i.e. I've chosen .wmv in "render as", what should I choose in the Final Render Template?

Also, it would be helpful to better understand how network rendering can / can't be used on MPEG-2 renders. I have a project that took 1 hour 25 minutes to render to MPEG-2 on my 3.6Ghz P4. One 1000 frame section of that alone took 30 minutes - would network rendering have helped?

Computer #1 is a 3.6Ghz P4
Computer #2 is a 2.26Ghz P4
Media is all stored on computer #2
Gigabit network connected through a switch

Rick

Comments

Liam_Vegas wrote on 5/22/2005, 4:44 PM
You should not use network rendering using any template based upon high compression scheme (like WMV or MPEG).

If your source files are DV - then things are pretty simple. Network render to DV format - and no need to specify a final render template. Then render THAT final format to MPEG2 or WMV. This has the advantage of keeping your final AVI around and makes fixes/changes far simpler and quicker than re-rendering the entire timeline.

If you really do want NR to generate a final format that is highly compressed (WMV or MPEG2) then that is when you would use a render format of DV (or DV with the YUV codec) followed by a Final render template of your higher compression format.

*note: You cannot actually use NR to render MPEG2 direct - as this requires additional licensing for the MPEG codec that you don't get. When NR is asked to do a render using MPEG2 it will check that each NR Node has a different serial number/license... if they all have the same one then it will fail). In addition... whenever you use a highly compressed format for NR - it will take a double hit as far as compression... first when the individual sections are rendered - and then when these segments are "stitched" together again. This is why you use the "Final Render template" thing.
RNLVideo wrote on 5/22/2005, 4:58 PM
Thanks, Liam - that makes a little more sense. I guess I'll have to do more experimenting with it to see how it will be of benefit to me.

Rick
RNLVideo wrote on 5/22/2005, 8:07 PM
Liam (or others) - to make sure I've got the concept, is this right?

- If simply using network rendering to get another box to render the file, choose format in "render as" (including high compression formats) and don't check "distribute rendering" (correct?)

- Distribute rendering uses the networked computer to render chunks of the project in whatever format was chosen during "render as" (correct?) then stitches them together and induces further compression in that same format if "Use Final Render Template" is NOT checked (correct?)

- When using distribute rendering option and "Use Final Render Template" IS checked, a selectable but different second compression is used to stitch the file together (correct?)

- Best workflow to minimize loss of quality is to "render as" uncompressed avi, DV, YUV, etc and then stitch using the "Use Final Render Template" with the final lossy compression (i.e. .wmv) chosen (correct?).

Not sure why this has been so difficult to grasp, but thanks for the help!

Rick
Liam_Vegas wrote on 5/22/2005, 9:00 PM
- If simply using network rendering to get another box to render the file, choose format in "render as" (including high compression formats) and don't check "distribute rendering" (correct?)
Yes

- Distribute rendering uses the networked computer to render chunks of the project in whatever format was chosen during "render as" (correct?) then stitches them together and induces further compression in that same format if "Use Final Render Template" is NOT checked (correct?)
Yes... if using something like MPEG or WMV which uses high compression...BUT if you render as regular DV (which is compressed) the final stitch will be a bit-for-bit copy just like it would be if you were rendering from DV to DV and the frames were unchanged

- When using distribute rendering option and "Use Final Render Template" IS checked, a selectable but different second compression is used to stitch the file together (correct?)
correct

- Best workflow to minimize loss of quality is to "render as" uncompressed avi, DV, YUV, etc and then stitch using the "Use Final Render Template" with the final lossy compression (i.e. .wmv) chosen (correct?).
Uncompressed AVI (really uncompressed) will be best... but it takes a huge amount of space..... but DV with YUV codec (still bigger than straight DV) will preserve the most amount of information for that final high compression render/encode.

Not sure why this has been so difficult to grasp, but thanks for the help!
It is quite a tortured learning curve for NR... and I cannot claim to be THE expert on the matter... but I am now making a lot of great use of it.
Spot|DSE wrote on 5/22/2005, 9:06 PM
The Cineform intermediate does hold up great in the stitching process. The 4:2:2 YUV is even better. However, unless you have a Decklink card, you'll not gain *much* by using the larger file size of the YUV. The YUV codec renders faster than the Cineform, but not by a lot.
RNLVideo wrote on 5/23/2005, 4:11 AM
Outstanding - thanks for the help!