POWER OF RED

[Phil Holland]

Yes, you do need to calibrate it on a flat surface.

 

[Patrick Tresch]

Can a Gyro be as good as a tracker?
I bet it will be difficult to use Gyro data from a shoot with a long lens.
What is the maximal usable focal length with the actual Gyro data we got in camera ?

 

[Mikael Lubtchansky]

Can someone confirm, I now extracted r3d gyro metadata and this is what I found. Basically its not so accurate. What you see is one keyframe for each axis for a 300 frame handheld clip shooting from a small boat in waves. Clearly not any two frames in a row should have the same values... But as it looks it both a bit eratic and stays still for a couple of franes at the tine.

Did a simple data extract tool several years ago but accuracy of the gyro were similar to your findings... even after a "as good as I could do" low and high temperature gyro calibration...

was your test shot on Monstro ?

 

[Björn Benckert]

Did a simple data extract tool several years ago but accuracy of the gyro were similar to your findings... even after a "as good as I could do" low and high temperature gyro calibration...

was your test shot on Monstro ?

Yes it seams like this gyro is not on par with anything out there. Really low temporal and space resolution. I can't even believe this is the raw gyro data, my guess some code between the gyro and the r3d went seriously wrong. I would love to dive deeper into this. Anyone knows where inside of the DSMC2 cameras the gyro is seated? Replacing it with a of the shelf mobile gyro or what DJI and GoPro is stellar in comparison and would do wonders. Those are in the 100s of dollar range... but for Monstro 80k USD I paid...

Any way, will start a separe thread about it. As I wrote above my programer will make a little r3d data extract gismo.

Basically putting 3D cameras into r3d folders. But sadening to see this... I though red was about high resolution, but when it comes to gyro data its basically on Mindcraft / Lego levels... Gyros is apparently at least currently not the "Power of Red"...


Jarred, would be fantastic to hear if this is a dead end or if it can be improved for DSMC2 and if not if you will put at least a half decent gyro in Komodo or dsmc3, if its on the road map. Reframing and extra resolution is great but this would actually make those extra K´s worth while... and yes you can completely eliminate rolling shutter skew with gyro data. Kind of end the discussion of people requesting global shutter, etc, etc...


Untitled by Björn Benckert, on Flickr

 

[joshua csehak]

Also, it'd be even more rad if you guys could enable it in RCXp/SDK as a "stabilization" slider (in all DSMC1+ footage, no less)

OMG that'd be amazing.

Anyone know if Premiere is using this data for the Warp Stabilizer or if I should using a different method to stabilize?

 

[David Rasberry]

Yes it seams like this gyro is not on par with anything out there. Really low temporal and space resolution. I can't even believe this is the raw gyro data, my guess some code between the gyro and the r3d went seriously wrong. I would love to dive deeper into this. Anyone knows where inside of the DSMC2 cameras the gyro is seated? Replacing it with a of the shelf mobile gyro or what DJI and GoPro is stellar in comparison and would do wonders. Those are in the 100s of dollar range... but for Monstro 80k USD I paid...

Any way, will start a separe thread about it. As I wrote above my programer will make a little r3d data extract gismo.

Basically putting 3D cameras into r3d folders. But sadening to see this... I though red was about high resolution, but when it comes to gyro data its basically on Mindcraft / Lego levels... Gyros is apparently at least currently not the "Power of Red"...


Jarred, would be fantastic to hear if this is a dead end or if it can be improved for DSMC2 and if not if you will put at least a half decent gyro in Komodo or dsmc3, if its on the road map. Reframing and extra resolution is great but this would actually make those extra K´s worth while... and yes you can completely eliminate rolling shutter skew with gyro data. Kind of end the discussion of people requesting global shutter, etc, etc...


Untitled by Björn Benckert, on Flickr

You might correct some skew and jello with gyro data, but you can’t recreate picture elements that were not recorded due to rolling shutter line scans. A global shutter exposes every pixel of the frame for the full duration of the frame, scan only occurs during readout of the stored charge states. Gyro won’t help with flash banding for instance.