Build 30hz rotary disc shutter for 5DII, eliminate rolling shutter at DVinfo.net

Paul Martin · May 17th, 2009, 08:18 AM

Hello everybody,

Based on other's experimentations, the 5DII's rolling shutter read speed has been measured at 43200 lines per second. Thus each frame is read in 1/40 seconds.

This means that, shooting 1080p at 1/30s, there is a period of 1/120s when all pixels are integrating simultaneously.

If the camera can be shuttered by means of a rotary disc shutter for all but this 1/120s, the camera will be effectively globally shuttered.

What is required is a disc of as large a diameter as possible, perhaps 8 inches, with a 2 inch deep notch cut out of 90 degrees of it, spinning at exactly 1800 rpm, in a housing in front of the 5DII lens.

The housing should be as thin as possible, but regardless would exclude the use of lenses much wider than 50mm, in all likelihood, due to vignetting.

Ideally the motor used would be quiet, powered by no more than 24vdc, and would run very stably at 1800 rpm. A means of adjusting the phase of the motor so as to sync it with phase of the 5D would be very helpful if not absolutely necessary.

Can anyone recommend a motor and/or controller that would fit this bill?

Bruce James · May 17th, 2009, 11:46 AM

8" blade with notches spinning at 1,800 RPM.

Sounds a lot like a circular saw to me. I hope you're insured ;-)

I think you want something like a stepper motor or a brushless regulated motor like the ones found in video players.

B

Martin Catt · May 17th, 2009, 01:32 PM

Try four notches in the disc and cut the RPM down to 900.

Synchronizing and maintaining sync with the camera's electronics might be problematical. --IF-- there is a video out on the camera, you could extract a sync pulse out of the video signal and use that to coordinate things so a slot is in front of the lens at the correct time. After that, I'd guess a phase-locked-loop keying off a position sensor on the disc's shaft could keep the spin rate within specification.

On the surface, it looks like an either/or situation. You're basically making a rotating shutter that with the exception of an opaque, slotted disc, is just the same as a DOF adapter. Why go to the smaller, lighter 5DII with the larger sensor for better DOF, then go back and clutter it up with a spinning disc rig? No advantage over a smaller sensor HD camera with a DOF adapter.

Martin

Paul Martin · May 17th, 2009, 03:07 PM

The whole point of building a rotary shutter for the 5D2 is to eliminate its rolling shutter artifact.

Too small a disc spinning too slowly would introduce its own rolling shutter artifact as it would take too long for the leading and trailing edges of the shutter to traverse the face of the sensor.

Because of the nature of the the 5D's rolling shutter, a 90 degree rotary disc shutter is the largest possible shutter for 1080p. A Cinemascope 2.39:1 1920x804 crop would allow the use of up to a 159 degree shutter.

It's not a question of extra "notches". I shouldn't have even used that word.

At 900 rpm the shutter would be a butterfly shutter and would have two 45 degree "notches", instead of the single 90 degree "notch" of an 1800 rpm shutter.

I would prefer not to use a 900 rpm butterfly shutter as it would necessarily have a worse rolling shutter artifact than an 1800 rpm shutter of the same diameter.

Maybe at 8-inches in diameter, 900 rpm would still be fast enough to avoid any major rolling shutter issues, so a butterfly shutter could be a great compromise if 1800 rpm was proving too difficult yet 900 rpm was not.

I think an 8-inch diameter is a good idea because it's a little more than 200mm, so it'd be conceivable to come up with a housing design that could be used in front of lenses as large as 72mm or so. That's without having to locate the motor off center and use a band, like in the Letus Ultimate. Compatibility with up to 77mm front filter threads would be ideal, but might require a 9-inch shutter.

Quote:

Originally Posted by Martin Catt

Why go to the smaller, lighter 5DII with the larger sensor for better DOF, then go back and clutter it up with a spinning disc rig? No advantage over a smaller sensor HD camera with a DOF adapter.

Are you kidding me?

You make it sound like the only reason the 5DII is any good is because it's small and light. Those certainly are nice qualities, but wouldn't care if the 5DII weighed 50 lbs. There isn't a single 1/3"-1/2" sensor camera and DoF adapter combo that can touch it in terms of dynamic range, color, and light sensitivity. Not to mention codec, barring the use of a Ki Pro or Nanoflash, neither of which are even out yet.

Paul Martin · May 17th, 2009, 03:32 PM

Quote:

Originally Posted by Martin Catt

Synchronizing and maintaining sync with the camera's electronics might be problematical. --IF-- there is a video out on the camera, you could extract a sync pulse out of the video signal and use that to coordinate things so a slot is in front of the lens at the correct time. After that, I'd guess a phase-locked-loop keying off a position sensor on the disc's shaft could keep the spin rate within specification.

These are great ideas.

Dan Vance of the Vancecam (VanceCam VC25P Camcorder) built a circuit (VC25P Shutter Drive Circuit) to sync his spinning rotary shutter to the sync pulses of his video camera.

Yeah, he's done all this before.

And the 5DII does output a composite NTSC signal, so it'd should work with Vance's circuit.

The only real difference between the rotary shutter Vance built and the one I'd like to build is the size. His was for a 1/2" sensor and it was still pretty big (see first link). A bigger sensor requires a larger diameter disc. The 5DII sensor is around 5x the diameter of the sensor used in the Vancecam. But the fundamentals are the same.

Bob Hart · May 18th, 2009, 08:32 AM

It is all well said above, all thoughts I have entertained when I suggested a single groundglass disk with segments of different groundglass grades selectable by trimming the spindle motor timing.

For a simpler "global shutter" arrangement, I might be tempted to try to synchronise an electronic shutter from a welder's helmet if it can be made to cycle fast enough. This would be an extreme duty cycle for it so it might not last as long as in the welding application.

There might be an issue with IR contamination.

Matthew Nayman · May 23rd, 2009, 08:36 AM

Dynamic range and colour is wonderful, but no manual controls and AVCHD is not. Buy a Scarlet.

May 17th, 2009, 08:18 AM	#1
Paul Martin Regular Crew Join Date: Dec 2002 Location: San Francisco, CA Posts: 47	Build 30hz rotary disc shutter for 5DII, eliminate rolling shutter Hello everybody, Based on other's experimentations, the 5DII's rolling shutter read speed has been measured at 43200 lines per second. Thus each frame is read in 1/40 seconds. This means that, shooting 1080p at 1/30s, there is a period of 1/120s when all pixels are integrating simultaneously. If the camera can be shuttered by means of a rotary disc shutter for all but this 1/120s, the camera will be effectively globally shuttered. What is required is a disc of as large a diameter as possible, perhaps 8 inches, with a 2 inch deep notch cut out of 90 degrees of it, spinning at exactly 1800 rpm, in a housing in front of the 5DII lens. The housing should be as thin as possible, but regardless would exclude the use of lenses much wider than 50mm, in all likelihood, due to vignetting. Ideally the motor used would be quiet, powered by no more than 24vdc, and would run very stably at 1800 rpm. A means of adjusting the phase of the motor so as to sync it with phase of the 5D would be very helpful if not absolutely necessary. Can anyone recommend a motor and/or controller that would fit this bill?

May 17th, 2009, 11:46 AM	#2
Bruce James New Boot Join Date: Apr 2007 Location: New Milton,UK Posts: 19	8" blade with notches spinning at 1,800 RPM. Sounds a lot like a circular saw to me. I hope you're insured ;-) I think you want something like a stepper motor or a brushless regulated motor like the ones found in video players. B __________________ -- http://www.pomegranate.ltd.uk

May 17th, 2009, 01:32 PM	#3
Martin Catt Major Player Join Date: Jun 2007 Location: The Colony TX Posts: 327	Try four notches in the disc and cut the RPM down to 900. Synchronizing and maintaining sync with the camera's electronics might be problematical. --IF-- there is a video out on the camera, you could extract a sync pulse out of the video signal and use that to coordinate things so a slot is in front of the lens at the correct time. After that, I'd guess a phase-locked-loop keying off a position sensor on the disc's shaft could keep the spin rate within specification. On the surface, it looks like an either/or situation. You're basically making a rotating shutter that with the exception of an opaque, slotted disc, is just the same as a DOF adapter. Why go to the smaller, lighter 5DII with the larger sensor for better DOF, then go back and clutter it up with a spinning disc rig? No advantage over a smaller sensor HD camera with a DOF adapter. Martin __________________ Canon XF300, Canon 5DMkII, Canon XL2, Rolls MX422 mixer, Zoom H4N, AT899 lavs, AT2020's, Azden SGM 1X shotgun, Manfrotto 501 head on 351 tripod

May 18th, 2009, 08:32 AM	#6
Bob Hart Inner Circle Join Date: Dec 2003 Location: PERTH. W.A. AUSTRALIA. Posts: 4,477	It is all well said above, all thoughts I have entertained when I suggested a single groundglass disk with segments of different groundglass grades selectable by trimming the spindle motor timing. For a simpler "global shutter" arrangement, I might be tempted to try to synchronise an electronic shutter from a welder's helmet if it can be made to cycle fast enough. This would be an extreme duty cycle for it so it might not last as long as in the welding application. There might be an issue with IR contamination.

May 23rd, 2009, 08:36 AM	#7
Matthew Nayman Major Player Join Date: Feb 2005 Location: Toronto Posts: 917	Dynamic range and colour is wonderful, but no manual controls and AVCHD is not. Buy a Scarlet.