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 :)

Tessive Time Filter - Holy Crap!

Andrew Rieger

Andrew Rieger

Well-known member
Joined
Oct 31, 2009
Messages
1,024
Reaction score
0
Points
0
Location
Pasadena, CA
Knowing Reduser, there is probably already a thread about this product but if you have not heard of it before, you should check it out. It basically adds a global shutter to the front of any cmos camera and helps to eliminate rolling shutter. I imagine this will be pretty popular with productions using a lot of shoulder mounted cameras and panning. What do you guys think? this looks like one of the most exciting products I have seen in a long time!

http://www.freshdv.com/2011/06/cinegear2011-tessive.html

http://www.tessive.com/home/time-filter

http://www.tessive.com/home/demo-footage
 
Really cool idea and it seems useful for certain applications, but $14,000 and a two-stop light loss? Plus you have additional glass in front of the lens, and an additional box hanging off the back. Not sure if that's worth it in average scenarios. It would be sweet if RED could build something like this right into the camera, but rather than a large filter in front of the lens, it would be a small filter right in front of the sensor.

Thanks for sharing, Andrew.
 
Gotta ask. Is rolling shutter really still that big of an issue. I'd much rather use post solutions to address it and preserve the cash as well as the 2 stop loss. We shot our feature on build 16 and even then we didn't have any issues - noise in the blue channel was our dilemma back then.

cheers
 
I spoke to the guys at CineGear. They are really friendly and awesome - and the footage they showed really did look better. Likely if Red wanted to offer it as an in-camera option, they'd be very interested in playing ball with RED.

I chatted with them about other things they could potentially do and suggested they offer an option to work with the cam at 48fps and alternate the ND every frame, "HDRx"-style. It would probably work perfectly with RED's existing HDRx workflow too!

Interesting to see what special effects you could do as well. Anyway, fun fun fun!

Bruce Allen
www.boacinema.com
 
I did not think about it but I wonder if this would work with HDRX? Ya, 2 stops lost is pretty big but with the Red's sensitivity, I don't think it would be much of an issue.
 
On second thought, I don't think this would work with HDRx. In hdrx, do you have control over the shutter speed of the A and X tracks? The first primary exposure would need to have a combined time of 1/24 of a second. Is this possible?

I am also confused about the stop loss. You set the camera to 1/24 which is already letting in more light than a traditional 180 shutter and since this filter essentially acts as the shutter, why is there any stop loss? Is this 2 stops of light lost from the 1/24 shutter or around 2 stops lost from the traditional 1/48 shutter? I'm confused.
 
Interesting. Does it do anything for sharp strobe bands? Over to Graeme!
 
Andrew Rieger said:
On second thought, I don't think this would work with HDRx. In hdrx, do you have control over the shutter speed of the A and X tracks? The first primary exposure would need to have a combined time of 1/24 of a second. Is this possible?
Click to expand...

Workflow would be:
Run cam at 48fps, shutter wide open.
Set Tessive filter to turn ND on every 2nd frame.
Then manually tell RedCineX that it is a HDR track (right click on the track).

ericyoung said:
Interesting. Does it do anything for sharp strobe bands? Over to Graeme!
Click to expand...
It would make them smoother - they would be more like a gradient. EG more like a film shutter.

Bruce Allen
www.boacinema.com
 
Bruce Allen said:
...It would make them smoother - they would be more like a gradient. EG more like a film shutter.

Bruce Allen
www.boacinema.com
Click to expand...

Thought it might but had problems visualising it.

Sounds like it would be worth hiring one for specific shots. Certainly wouldn't buy one at that price though.
 
Bruce Allen said:
Workflow would be:
Run cam at 48fps, shutter wide open.
Set Tessive filter to turn ND on every 2nd frame.
Then manually tell RedCineX that it is a HDR track (right click on the track).
Click to expand...

Bruce, I might be wrong, but this would probably be less effective than EPIC's HDRx approach. HDRx captures two images AT THE EXACT SAME MOMENT, only with two different shutter speeds. What you're suggesting is trying to combine images that are captured in sequence (1/48th of a second apart), which would make the images not quite line up if there is motion in the frame.
 
Thanks for the kind words, everyone. The response to our Time Filter has been awesome. We're working quickly to get these placed where you can rent them, and we also are working with some people for system demonstrations.

We certainly understand that this is a very new and unusual product, so we're doing our best to help match the needs of artists with our technology. If there's something you'd like to know about the system, we're very happy to answer questions.

Folks here have noted the drawbacks of the system: light loss and price. This is a first-generation product. When we initially had the idea for a temporal antialiasing filter, nobody could tell us with any certainty if there would be any market for such a thing. We've found that there most definitely is, and now that some real cinematographers are getting their hands on it (rather than our engineering staff, who are well meaning but not pros) the results are even more spectacular.

The primary goal of the Time Filter is to correct temporal antialiasing (our website explains the details of this), but a side-effect is that it also corrects rolling shutter. Frankly, the rolling shutter in the RED ONE we use hasn't ever bothered me all that much, and it wasn't a design goal to fix it, but it does. We added in a normal 180-degree shutter into the filter control so that you can use it to correct rolling shutter without using the temporal antialiasing curve. The cameras that would most benefit from rolling shutter correction are DSLRs, and unfortunately we cannot synchronize with them.

We are planning more advanced integration with the EPIC, which has some nice synchronization options for over- and under-cranking.

So thank you again, everyone. Everyone we met at Cine Gear was great, and it was a blast.

Tony Davis (the guy in the white lab coat...not my normal outfit.)
www.tessive.com
 
Welcome Tony!

Can you shoot and post a few test movs or avi movies (some format we can advance through one frame at a time) with camera flashguns, and emergency strobes? Would be great if it solved/reduced the one artifact that can be a deal killer when shooting with rolling shutter digital cameras compared with film.
 
ericyoung said:
Welcome Tony!

Can you shoot and post a few test movs or avi movies (some format we can advance through one frame at a time) with camera flashguns, and emergency strobes? Would be great if it solved/reduced the one artifact that can be a deal killer when shooting with rolling shutter digital cameras compared with film.
Click to expand...

Will do. I'll grab my Canon flash and do some quick tests. It'll probably be a couple of days, but that sounds like fun. I'll post the results to our website and let you know when they're ready.

Tony
www.tessive.com
 
Craig Parkes said:
A question - would this also reduce light flicker from out of phase lighting? I can only imagine it would.
Click to expand...

Yes. 50 or 60 Hz lighting produces a fundamental frequency of 100 or 120 Hz (unless it's broken, in which case it'll be half that and even worse.) In these frequencies, the Time Filter causes the camera's response to be effectively zero. If you've ever chased these flickers around by changing the shutter angle (for instance, 144 degrees for 60 Hz lighting) then you know that the shutter has influence on how this works. The normal square-wave shutter of all cameras has an infinite sequence of zero-nodes in the high frequency response, but they're very sharp. Changing the shutter angle causes these zero-nodes to sweep up and down the frequency dial, so you can fiddle with the shutter angle until the frequency offending you is zeroed out. Of course, you just get the fundamental, not all the harmonics.

The Time Filter shutter drops to zero response around 40 Hz, and stays there all the way up to infinite frequency. So there's no dialing it in or sweeping it around. It just kills all these offending frequencies and their harmonics and they stay dead.

Wow, that was a really long answer to say, yeah, it fixes that.

Tony
www.tessive.com
 
Sync on HDSLR

Sync on HDSLR

Quote: [A frequent follow up question is: "Will the Time Filter work with my DSLR, like the Canon 5D Mark II, etc." We are very sorry to say that it won't right now. We need some sort of capability to synchronize with the camera's frame acquisition, and these cameras simply don't provide any way to do that (yet.) We are also excited about the future of these less expensive, lightweight cameras, and would love to be able to provide a solution for them. Our search for an answer continues.]

There might be several simple ways to sync your optical shutter to any HDSLR.

1) You can put an induction coil near the sensor and pick up the readout frequency.

2) You can put a small resistance between the battery terminal and the camera to produce a voltage, there should be some variability of the power use on the shooting frequency that can be amplified and run through a filter.

3) You can pick up EMI/RFI coming from the camera that varies with the shooting frame rate that can be amplified and filtered.

4) Solder onto some point in the camera circuits that gives a signal related to the frame rate.

5) There might be some frame rate ripple on a plug-in-power microphone bias(?)

6) Likewise if there is a headphone output, those may have frame rate ripple in them as well.

7) Stick a adhesive foil strip inside the lens mount area, or on the outside of the camera and pick up the frame rate by capacitive coupling.

8) Is there a relationship between the HDMI or other monitor output VHz and the sensor fps, if so you may be able to genlock after adjustment for that ratio if they use the same crystal clock in the camera(?)...

So it may just be as simple as putting a pickup on the outside of the camera to find its shooting frequency, then do a PLL to a "jam sync" in the crystal of your control box.

You will need a phase control to time the optical shutter to the start of the readout.

Quote: [By the way, the Time Filter has many modes of operation, one of which is to create a simple 180 degree shutter. If you set your camera to 360 degrees, then the Time Filter to 180 degrees, it will totally eliminate rolling shutter without applying the antialiasing filter if you don't want that for a particular shot. Of course, in antialiasing mode, rolling shutter artifacts are eliminated also.]

For cameras with small VBLANK to readout ratios, your optical shutter should not "totally eliminate rolling shutter" it will put a fade on the slope, which is a reduction of noticeably but not an elimination of the skew, you should re-word your faq. The skew is blurred, not totally removed (when using the "360" degree bell curve (where total light pass equals 180 exposure, if you have it half the time black then its a 180 bell curve with about 90 degree peak), rather than a 1 degree global strobe which is too fast for your shutter anyway(?)). The bell curve graph on the site shows that minimum transmission of the LCD filter is not zero, so there would be less than total extinction.

Quote: [360-degree shutter (the normal configuration with the Time Filter)]

I don't think there are any cameras with 360 degree shutter since the reset and read out lines in the sensor are not the same, so at most you get something like 1 line less than total making the maximum angle a bit smaller than 360, as 359.x?

An electro-mechanical version of this type of filter would have no "light loss" because the two pola filters would not be needed. It would be hard to wip-pan though with a large spinning filter... (the idea of these filters is to cut the light by one stop anyway, what I am talking about is the loss in addition to the one stop loss). I had this idea some years ago in several variations, PLZT filters of that time had some diffraction issues so the newer LCD give less "fog" to the results.

One problem I have noticed with 3D LCD shutter glasses is that when dark the density is somewhat un-even. I wonder what the density variation is for your filter over its area at max density. For longer lenses such un-even density would not be a major issue, but for wide angle lenses the crossed pola filters may not be even center to edge? Do anti-rattle filter clamps in the matte box frames cause issues with the liquid crystals, as usually pressure of such can show up as rainbows or un-even density? How do very cold or hot conditions affect the liquid crystals density and switching?
 
Dan Hudgins said:
There might be several simple ways to sync your optical shutter to any HDSLR.
Click to expand...

The practical question then becomes: Is a $13k item worth it when you could just rent a RED for the same price. :D
 
You must log in or register to reply here.
Back
Top