Welcome to our community

Be a part of something great, join today!

Log in
  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

EPIC genlock and shutter phase

Jason Goodman

Jason Goodman

Well-known member
Joined
Jun 4, 2009
Messages
324
Reaction score
0
Points
16
Location
New York City & Los Angeles
We were recently contacted by our good friend Charlie Parra at Denecke about some problems one of his customers was having shooting 3D with two EPICs and an SB-T. After trying sync boxes from all manufacturers, it was determined that the problem originated elsewhere, however the questions prompted us to conduct a comprehensive test of "3D sync" for the current (as of yesterday) build of EPIC.

True "3D sync" requires genlock of course, but also shutter phase sync. In addition to the camera's internal clocks being synchronized, shutter timing must also be perfectly synchronized or very noticeable visual artifacts will be present in your 3D imagery.

Tests we conducted in 2009, prior to shooting Julia X and published on REDUSER by Pedro Guimaraes, indicated that RED ONE (and RED ONE MX) never fully achieved perfect shutter phase sync. RED ONE builds later than 21 did allow cameras to be genlocked and 3D worked well up to 120fps at 2K. Many 3D films have been successfully shot with RED ONE. The small shutter phase error was generally not considered a problem and in most cases was not noticeable.

The good news is, all builds of EPIC I have seen including beta builds we used last year on The Amazing Spider-Man have exhibited absolutely dead perfect shutter phase sync. The ASM crew was instructed by RED as to which frame rates, resolutions and time bases to use in 3D for that film, and to my knowledge we did not exceed 72fps in 3D.

As always, with newer builds come new capabilities.

Last week, prompted by Charlie's inquiry, we tested as many combinations of resolution, frame rate and time base as we could think of. We have assembled the most comprehensive list of working 3D modes for the EPIC and I've posted it here for reference.

One issue of particular interest, we noticed instances where the EPIC would indicate GEN, TC and SYNC, but in fact, shutter phase was off. We did the classic test of shooting an LED timecode slate with the shutter set to 1/1000th of a second so we could see partial redraw of the LED numbers. If the numbers are not drawn exactly the same in both eyes, shutter phase is off. Here's the chart:

https://docs.google.com/spreadsheet/pub?key=0AuJAHpbX6ekpdEt4d3BOcUk1ZDFnYlNROVhLdklKalE&output=html
 
So awesome!!!!!!

I know this took some time to put together so........Thanks Jason!!

Will be a great help to everyone shooting 3D with EPIC.

....naturally would be great it the cameras always locked.....
 
Thanks Jason! That really helps alot... So I guess 72fps still appears to be the highest framerate that syncs at 5K res.
Out of all the genlock boxes you've worked with, is the Denecke your first choice?
 
Jason, Thank you for doing this legwork.....

Keep up the hard work and let me know when you re in LA....

Peter
 
Hi Jason

In your chart, does the column labelled 21C3D Test Sync indicate your qualification of the LED drawing? And the column Epic Sync indicate the icon on the UI?
I assume you were reviewing footage not monitor display. And your test set up is two Epics on your right angle beam splitter rig?
 
Thanks Jason for the input. I've been seeing the SB-T in the 850-$900 range online, but then I realized it may not be the latest & greatest 3D mod'd version of the SB-T.
Will call Charlie and verify this.
 
Hi Dave,

The 21c3D Test Sync column indicates that we photographed an LED timecode slate at that timebase and frame rate, with the shutter set to 1/1000th of a second so we could closely analyze the redraw of the numbers, and we determined that the partially drawn numbers looked identical in each eye. The EPIC Sync column indicates the icon on the UI thought it was in sync. We used our two EPIC-X cameras on a 21st Century 3D BX3.5 beam splitter at 0" IA so essentially identical viewpoints for the EPICs. All evaluations were done on recorded footage rendered out of Cine X.
 
Hi Jason

Can you let me know the sw version number precisley? I am surprised you have so many non syncs in the EPIC Sync column. Had you performed a user black shade calibration and are using it?
Lastly, in your rig, does one camera record an upside down image?

thanks
 
Pedro, if you are seeing problems, I would love to hear details.

Jason, its great that you put this together and go through this. I went through all the ones you have marked as Epic Sync = no, with version 3.0.0, but with a dark shading correction performed and enabled. This allows the higher frame rate cases to work.
In fact all of those do work and show green SYNC icon, with the exceptions of the ones in the 25fps project settings. The 120 and 96 fps cases are not valid conditions since they are not multiples of 25. Changing the 120's to 100's and eliminating the 96fps/5K (you cant get to 100) case clears all those up.

As for the slight phase difference, I believe its because one camera is probably upside down, or essentially seeing the world upside down, so the rolling shutters are not scanning the world in the same direction. I dont know for sure if thats true on your rig. I think if you play with the placement of the slate vertically in the scene, you can make the amount of phase difference increase or decrease.
 
Hi Dave,

Sorry I was off the list for a few days. This was all done with version 2.08. We did not do a black shade beforehand. That is good to know if that makes a difference. I will upgrade to 3.0 and can do more testing tomorrow or Thursday. The BX3.5 mirror rig orients the cameras such that both sensors image top down. The image gets flipped horizontally in CineX to minimize rolling shutter 3D artifacts.
 
Dave,

After your previous message, we upgraded the EPICs to 3.0 then proceeded to black shade both cameras. What we found was that in the Time Base of 25, we got proper sync indications on the camera and also in our visual check of the recorded material for 4K 100fps, 4K 75fps, and 4K HD 75fps. These modes did not work properly in our previous tests of 2.08 without black shading. Our document has been updated and the new doc is here:

https://docs.google.com/spreadsheet/pub?key=0AuJAHpbX6ekpdEt4d3BOcUk1ZDFnYlNROVhLdklKalE&output=html

It is possible that 3.0 and blackshading may have enabled additional working 3D modes, but the cameras had to go out on a rental. We will conduct more tests when they return.
 
Hi Jason,

thanks for the update. Thats not a convenient side effect. User black shade calibration is required to get those higher framerates to sync.
 
Dave Winder said:
User black shade calibration is required to get those higher framerates to sync.
Click to expand...

HI Dave, why is that? What is going on inside that black shade, a noise/color thing in my mind, is related to high speed shutters syncing?
 
To add my story to the mix - my two Epics were just out on a shoot where they were shooting one of the days in 3D. First, when trying to black balance the one camera, it failed to the point where I have to now take it back to RED for repair. But with the replacement camera that they had to rent last minute, they shot their 3D. Turns out that despite the apparent genlock when shooting, the footage is in fact out of phase.

I had originally tested the sync of my own two cameras - using a digital stopwatch and watching the frame blending, etc - and had no problems, though I frankly didn't do a comprehensive test of all resolutions and frame rates... I think we just tested 24 and 72. Regardless, I did not re-test that once I'd upgraded to v3 firmware.

Now it looks like two full shows might have been compromised. (Not to mention the extra cost of renting another camera last minute, though that wasn't a 3D issue.)
 
You must log in or register to reply here.
Back
Top