I've posted several "in progress" reports about splitting Sony TD10 & NX3D1 files with Sony's Content Management Utility software, but I thought I'd condense them here in one thread for anyone interested in using that software to process out full resolution Left & Right clips from 3D MVC files from both of these cameras.

First, one needs to install the CMU software with one of the above mentioned cameras USB plugged in. Thereafter it will operate without the camera plugged in. I thought for a while you needed to always have the camera connected to the computer to operate the software, but that was not correct.

Here is the link to download the software from Sony;
Sony | Micro Site NXCAM & AVCHD (http://pro.sony.com/bbsc/ssr/micro-nxcamsite/resource.downloads.bbsccms-assets-micro-nxcam-downloads-NXCAMContentManagementUtility.shtml)

Once installed, you can pull files directly from the camera card or internal memory. It is also possible to do this from a folder with transferred footage, and here is the biggest caveat - it is mandatory that these folders maintain the original AVCHD folder format when they are copied to the hard drive. These are the Private and AVF_INFO folders intact. CMU won't recognize orphaned M2TS files.

Second most important tip; use the File/ Import Media Files option and not the Register Folders To View option. This opens up a new window where your Stream M2TS files will appear. I've found it easiest to just Select All then Import. CMU does nasty things to your file names so be prepared to lose the generic 0000x.m2ts file name after the import - you can rename everything later.

Once you close that window and are back in the main window, you have to wait a few minutes (patience is a virtue after all) because the software is compiling the Import but doesn't give you any indication that it is doing so, it just appears to be doing nothing but after a short spell your newly imported folder will appear with the current date as it's name - don't ask me why I don't know and haven't figured out yet how to change this - it just appears to be a default. If you are going to import files from more than one folder, you will find the second, third etc imports in that first folder.

Now you have all of your M2TS files in view in the main window and you can select the ones you want to process into discreet Left and Right files. Select your file, right click and choose Create Separate 2D Movies for Left & Right. The software will make those files and append an L and an R to each file to signify it's status. You can batch process all at once by selecting all or use Control click to select randomly. The converted files are M2TS format and placed in the folder you designate in the Preferences dialog - usually the User/My Videos folder.

That's about it for CMU operation. The two most important tips being be sure to transfer your camera card within the AVCHD format with folders intact, do not drill down to the Stream and pull out the files when copying to hard drive from the camera card - and to use the Import Media Files option within CMU.

I have no idea whether CMU will work for any other manufacturer's 3D MVC files like Panasonic or JVC, if anyone has tried this feel free to chime in.

Postscript: I then use Cineform HDLink on PC or ReMaster on Mac to make Cineform MOV files from the M2TS Left & Right files, which are then muxed into Neo3D files for editing in Final Cut Pro with Cineform Firstlight handling the 3D and some picture adjustments. If I need to fine tune a 3D adjustment for specific shots, I tend to use 3D Stereo Toolbox for that.

Thanks for this update, Bruce.

Since the CMU separates L and R files in .m2ts codec, does it avoid the data size mushrooming that I experienced with MVC to AVI Converter?

And to be clear, are you saying that you need to convert the individual eye streams to Cineform .mov BEFORE then muxing those .movs into 3D? Or can the discrete .m2ts streams be muxed directly to 3D Cineform (hopefully saving a step)?

Thanks for this update, Bruce.

Since the CMU separates L and R files in .m2ts codec, does it avoid the data size mushrooming that I experienced with MVC to AVI Converter?

And to be clear, are you saying that you need to convert the individual eye streams to Cineform .mov BEFORE then muxing those .movs into 3D? Or can the discrete .m2ts streams be muxed directly to 3D Cineform (hopefully saving a step)?

Matt, data size does not mushroom, stays very much in the same data size realm. You can mux the individual M2TS directly to Cineform 3D MOV or convert each eye to MOV then Mux to 3D using HDLink/ReMaster and Firstlight. It's a little tricky to do the M2TS/MTS files directly, you have to use the Open command and not the Import command, but it works except that the audio is out of sync with this method. I use the second method to preserve audio sync. Cineform says they are working on the sync issue.

Speaking of sync issues: today I started getting into Cineform for the first time. I had created a rough first edit in Vegas (of my interview with my dad), and wanted to do the fine cutting in FCP, since I know the software so much better. So I output .m2ts streams for L/R eye, for both A and B cameras.

The B camera footage (from the GoPro rig) transcoded to Cineform fine, but whenever I tried to transcode the A camera footage, Cineform ReMaster added 21 frames of black before one of the eyes. The audio was also out of sync.

Now, this was footage from the same timeline (and was originally just the L and R streams from the TD10), and so shouldn't be out of sync. Nor was the Vegas output out of sync; Stereoscopic Player plays them back together fine. But, for whatever reason, Cineform ReMaster always throws it out of sync.

So far, I have not been able to re-sync it in First Light.

Any thoughts?

...whenever I tried to transcode the A camera footage, Cineform ReMaster added 21 frames of black before one of the eyes. The audio was also out of sync.
...
So far, I have not been able to re-sync it in First Light.
Update: it's still putting 21 frames of black in front of one of the streams, but I've figured out how to re-sync it in First Light.

Unfortunately, because of the 21 frames, it ends just before the end of the footage that I output from Vegas, so I'll have to re-output the ending, with handles.

I'm glad there's a work-around, but it's still a much less elegant solution than I would like. Anyone know what I'm doing wrong with Cineform?

Matt, I've had the audio sync issue with ReMaster but not the black frames. Dunno what to say about it but you can contact Cineform support via email and Jake is pretty quick to respond.

CineForm Support : Login (http://support.cineform.com/requests/new)

Thanks for the link, Bruce.

I'm finding this sync/lag in other files, as well. Since the interview is about consciousness studies, I'm learning to animate a 3D brain model in 3D Studio Max, to bring into the shot with my dad, just ahead of him in z-space, so the model can illustrate the structures and pathways, as he talks about them.

I've had a very complicated QC process, especially since I have ported my project over from Vegas to FCP. So I render from 3Ds Max, bring the image sequence into Vegas, export an M2TS, sneaker-net it to my Mac, transcode it to Cineform through ReMaster, mux it with Neo 3D, and then finally test it on the FCP timeline.

But the weak link in the chain has consistently been ReMaster, which keeps changing the duration, and/or lopping off the end of the source vid. I wonder if the results would be better, with a different input codec?

Anyway, I found a more elegant workflow, just using Vegas as QC, but eventually I'll have to get back into FCP (probably via After Effects), so Cineform is likely to be part of that mix, at least until I can save up enough for the Dashwood Tool Box.

I wonder if anyone has any experience working with Bino's 3D plug in, or OS3D, and can advise about their usefulness. I don't mind some back-bending to make the process work, but the more elegant the solution, the better!