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 :)

Rolling shutter test with #232

Well, it isn't a show-stopper for me. . . but it certainly is a big disappointment. I'm committed to the RED, so that's what I'll be buying (my number should come up in late February), but MAN, those rolling shutter artifacts are just killers. . . forget skew. . . I can live with skew. . . it's the strobes. . .

Alas. . .

Stephen
 
Jim, can you see the problem with the strobing by looking at the LCD (as you're taping), or is it only apparent in post?
 
Strobing is quite visible on your monitoring device as you shoot... and then again with a quick review of the take as well...
 
The EX1 will likely be my second camera since I just sold my HVX. Hell, first, until RED ships. It's tack sharp, most likely better than RED in resolution when RED is 2K windowed, and you can be a one-man band with it.
Click to expand...

i don´t think like this.
first, the hvx200 is a camera based on using perfect settings. if you get these, the difference to the EX is not so big as in many discussions named.
the red image will always look better than the EX. resolution is one thing, but the DR is much better, and that causes the quality of a good image, and the DOF and the sensitivity of the chip and and and. resolution is not my main argument for defining a perfect camera.
the EX is, seeing the prize, a very very good tool making pictures in HD. but itßs a camera you need a filmadapter for, i don´t like the native image of it, then you have a very videolikish picture, the red is very much closer to film. with an adapter, the EX makes very very nice footage, no question.
 
Let me explain my thinking about the RED and EX1 this way.

Sometimes you need light weight run and gun video on projects, sometimes film.

I've already had discussions with clients that simply don't want the overhead that RED requires. Not just the overhead in post with grading, but the overhead in crew to really pull off effective RED workflow on set.

For instance, is the RED best tool for a talking head? Photographing in a sensitive environment where you want to blend in a bit?

Other times you want film production values. That's where the RED shines. It's that whole genre thing Gibby is always bringing up. The difference is, I feel that RED isn't the genre cross-over than other people think it is. At least not without massive 3rd party hardware support that isn't here yet. Opinions will vary, no doubt.

As the EX1 picture... like other CineAltas it has massive picture control. It has more dynamic range and true resolution than the HVX.

And, to pull back to topic, if you shake pan both the RED and EX1 while flashing strobes at them, they'll match... :)
 
And, to pull back to topic, if you shake pan both the RED and EX1 while flashing strobes at them, they'll match...
Click to expand...


that´s right.

red owners (now and in the future) have one great character:

they have a very good sense of humor.
 
I also had a night-time storm scene I wanted to shoot with my Red One that has lightning strkes going off.

I've done this on many shorts before with a CCD camera and my SLR camera flash.. you might miss a flash or two, but when you get them they are fully on.

For me this is a big concern too.
 
Jim Arthurs said:
And, to pull back to topic, if you shake pan both the RED and EX1 while flashing strobes at them, they'll match... :)
Click to expand...

You'd probably need to sync jam the cameras for the strobing to match. :tongue:

Barry Green wrote a good article about rolling shutters on DVXuser: http://dvxuser.com/jason/CMOS-CCD/

Take a look at the wobble caused by vibration from the wind and from a helicopter. I didn't think the skew would bother me, but the wobbling from vibration is giving me second thoughts. But I seem to remember that Peter Jackson did some helicopter shots for his film, and Jarred did some car shots, and they were okay.
 
Floris Liesker said:
Why wont an LCD shutter work?
http://www.lctecdisplays.com/FOS-TN.asp

Ok it's polarized glass so you lose a stop. But it does work.
Click to expand...

Hmmm... I talked with an engineer at another LCD manufacturer with a similar product, what he told me was that that application won't work because the light entering the filter is not collimated. Most LCD shutters are geared toward highly collimated light such as lasers. The shutter has a very different transmission depending on the angle that the light is entering.

Plus correct me if I'm wrong but one 48th of a second is around 20 ms. That shutter you linked to has an attack time of 2ms but a decay time of 15ms. So that performance might not be enough to protect the sensor during readout, (but maybe it's just enough.)

It looks like a ferroelectric LCD shutters, may have much higher performance.
http://www.displaytech.com/photonics_shutters.html

Worth looking into, but I think the mechanical option is looking pretty good as well. Especially because there would be no polarization and light loss.

IBloom
 
Okay, I've edited my original post that started this thread to include some results from the compression/expansion effects of tilting the camera.

We'll be glad to do this all again (and with some improvements in the testing design) whenever a firmware lists an improvement in this area...

Regards,
 
would be very interesting to compare the lights of a policecar shot with red and the sony EX. the footage of the EX is unusable, but what´s with red?
same, better? slightly better?
 
Simply look at my first post for a link to a clip shot with strobes... the effect will be the same with the same style police car strobes as well...
 
I agree with Jim.
The RED and EX1 are advances in the evolution of this technology.

All cameras have their good and bad. It's really about the ratio of the amount of time these problems will be viewable. Having low noise, high rez, image latitide is needed 100% of the time. If I get an occasional rolling shutter artifact, I'll live with it.
But, of course, I will do my best to avoid amplifying the issue.

The RED ONE is quite the advanced camera. The image quality is nothing short of breathtaking.
 
Jim Arthurs said:
Okay, I've edited my original post that started this thread to include some results from the compression/expansion effects of tilting the camera.
Click to expand...

I gotta say, I'd much prefer the look of expansion/contraction on horizontal motion and panning than angular skew. Combined with normal motion blur, I don't think that it would be nearly as noticeable as skew. It may be more pronounced when the reset swipe has to travel the entire length of the sensor horizontally, as opposed to vertically as in this test. And it still doesn't fix anything else like the wobblies on fast random motion or vibration, or the strobe sync issues. But I personally think that it would help mask some of the shortcomings of the rolling shutter if the reset orientation was shifted to travel from side to side.
 
The only real solution I can see is to freeze the sensor during reads. LCD Shutters won't really help as the data is not frozen and still will have to be read in a raster fashion. Allow the sensor to read for the open shutter time (to get the appropriate blur) then have a global freeze that freezes each pixel for raster reading. Once read--on to the next frame.

Here is a company that did just that for machine vision: http://download.micron.com/pdf/articles/truesnaparticle.pdf--A VERY INTERESTING READ WITH PHOTOS.

While I doubt there sensor quality is anything compared to mysterium (it is mostly used for machine vision at high speeds not high dynamic range), they seem to have the global shutter thing down and maybe there is some technology to license.

At the very least, it describes the approach.

Another link to their tru-snap technology page:

http://www.micron.com/innovations/truesnap/
 
every sensor or stock type has stronger and weaker characteristics.

If you look at todays sensor design, there are sensors who have less or none rolling shutter/smear, FIT-CCD to name one.

However, using several FIT-CCDs introduces other problems into a cameras design (as prisms for a 3 chip design, which in itself are pretty expensive to manufacture). But if you look at the todays typical high-end CCD-FIT cameras (which would be sony hdcam), they mostly have 2/3 Sensorsize and aren´t abel to go over 60fps.

(Funny sidenote: Many DPs complained about the "shuttery" hard look back then when 1080p digital cinematography started in ~99/00, as they weren´t used to the level of perfection the FIT-CCDs offer in that aspect and missed the film-style rolling shutter)

Red 2K @ S16 @ 100fps is available and might be a good choice for whip-pans in VFX / Motiontrack situations.

Another "classic" solution when shooting 100fps (or 96) for the example with the rescue cars and their flashlight:
In 100P you get 4 frames, now combine these 4 frames in post - there are several packages which offer this functionality built in (discreet combustion).

Several CCD-FIT basing cameras (as Panavision Genesis or Sony 750/790/900R/F23) also allow the 360 degree shutter look, which isn´t available with red, but, to be honest, so what?

There will always be cameras suited better for certain situations, we often used film for overcranking on HDCAM-features, other use the weinberg/vision for ultra high-speed and red has other strenghts right now, as resolution, dof, and nice overcranking in 2K.

The problems with rolling shutter artefacts are mainly in ultra-fast VFX (greenscreen, motion tracking, combination of both) and these scenarios should be planned well ahead in any production, and if problematic it might be a good idea to use another camera for these usually very few special shots.

Indeed, several shows (battlestar galactica as on example) which are shot on hdcam add 35mm film cameras for those special highspeed-shoots.
 
ibloom said:
What's the difference between halting the light reaching the sensor and freezing it during a read.

IBloom
Click to expand...

Assumming there is enough of a lag between the time the sensor stops receiving light and the sensor reads all black, assuming the sensor stays fully charged for enough time to read it--then there would be no difference, but I really don't know if this is the case.

Honestly, I really don't know much about sensor technology, I'm just going on my technology experience in general and a bit of google searching.

I was assuming that once the shutter closed--the sensor would quickly, and unevenly, decay to read all black--so the only choice was to read it while it was open and therefore you would be subject to the same skew.

My main reason for posting the link I did was to bring it to the attention of others on the board with more experience that might not of heard of it--to fuel a solution so to speak. Sometimes all a genius needs is a spec of an idea that will lead them to an "a-ha!" moment that will drive them to a solution.
 
You must log in or register to reply here.
Back
Top