Strobe ?

[Rodrigo Lizana]

One of the major issues I´ve noticed when shooting with HD cameras is the strobe effect when panning. I never knew if it was due to the absence of an optical shutter or the all-in-focus characteristic of 2/3" chip size (which makes vertical lines in the background [specially bright ones] to remain in focus and strobe harder) or something else. Japanese manufacturers said the electronic shutter worked exactly like the optical one and there was no difference when comparing theirs with the optical one of film camera. People at Arri told me that they have noticed this effect when they tested some HD cameras along with film ones before the development of the D-20.
I´m wondering what´s the experience/opinion of the Red team in this subject ?.

[LighthouseMEdia]

Are you shooting in a progressive frame rate? There are two things you could be seeing, a progressive strobing or an interlace flicker, caused in ares of fine detail in interlaced footage. I'm guessing your seeing the progressive frame strobing though and sometimes if your not used to it, it can require an eye adjustment. There are several other factors you cuould research that would help minimize the strobing sush as watching your pan speeds.

[Stephen Williams]

One of the major issues I´ve noticed when shooting with HD cameras is the strobe effect when panning. I never knew if it was due to the absence of an optical shutter or the all-in-focus characteristic of 2/3" chip size (which makes vertical lines in the background [specially bright ones] to remain in focus and strobe harder) or something else. Japanese manufacturers said the electronic shutter worked exactly like the optical one and there was no difference when comparing theirs with the optical one of film camera. People at Arri told me that they have noticed this effect when they tested some HD cameras along with film ones before the development of the D-20.
I´m wondering what´s the experience/opinion of the Red team in this subject ?.

Hi,

IMHO the D20 strobes like other video cameras. I have only played with it at IBC several years running, but that was my initial reaction. I commented on CML and did not get flamed, so I guess that is accurate. Arri looked at my website and made no comment!

Stephen

[Billy Summers]

With film, I often notice a stutter effect when panning at 24fps, but not at 30fps. When panning with any d-cine camera (F900,Varicam,Genesis etc.) it does look video-ish to me (IMHO), which I hate! :angry03:

Maybe the RED will eliminate this terrible side effect of shooting digital, I don't know? :unsure:

[Poi Boy]

a one chip progressive digital cinema camera should not have any more strobing than a film cam panning at the same speed.
-A

[Mark Pugh]

a one chip progressive digital cinema camera should not have any more strobing than a film cam panning at the same speed.
-A
you are saying a 3-chip camera strobes more?

[Billy Summers]

a one chip progressive digital cinema camera should not have any more strobing than a film cam panning at the same speed.
-A

It shouldn't! But when you look at it, it does!

a one chip? You mean a cmos?

I see film stutter and digital strobe (or ghost?). What do you see?

[Emmanuel Cambier]

Funny you guys bring that up, cause last night I was watching "Marabunta" a bad C serie TV movie shot in HD, and I was really annoyed at this stuttering wich was really videoish.
The whole time I was praying for the Red One to be exempt of such disease, wich would make the whole thing useless.
Of course, I can't imagine the Red team settling for that kind of crap.

Emmanuel

[Poi Boy]

I guess that what I see is that stutter or strobing are more or less equally annoying. So if you are panning at 24fps with film or video you have to plan or adjust accordingly. I think the biggest culprit in making something look "videoish" is the deep dof that is inherent in three chip cameras.
Aloha
-A

[Graeme Nattress]

Strobing is a factor of edge sharpness and DOF. HD and other video cameras have lots of added fake sharpness which adds to strobing, and find it had to get shallow DOF which really helps avoid strobing. After that, it's down to how fast you pan.

Graeme

[Larry McKee]

If you pan with a subject, the studder, or strobe in this case, is much less noticeable. If you need to pan a shot with no subject to follow, the studder/strobe is much less apparent if the pan can last for about 7 seconds for any object to cross the screen. In other words, if you are panning to the left and there is a tree on the far left side of the frame, if it takes about 7 seconds for the tree to pass through the frame and disappear on the far right side of the frame, the studder/strobe will not draw attention to itself.

Hope that helps.

[Rodrigo Lizana]

HD and other video cameras have lots of added fake sharpness which adds to strobing


Thanks Graeme, I´d never have related electronic detail with strobe. I did with dof...

[Rodrigo Lizana]

and sometimes if your not used to it, it can require an eye adjustment. There are several other factors you cuould research that would help minimize the strobing sush as watching your pan speeds.

My observation was beyond basic technics. The strobe effect is something noticeable at "legal" panning speeds with or without a "subject" in the foreground. It helps to have something to use screen space and block the objects that are strobing but that doesn´t mean the strobe is gone.

Graeme posted an answer that´s new to me.

[Laco Zamba]

Don't forget about "FILM projected for many years with the 72fps light strobes ( 24fps x 3 bladed shutter)"

donatello http://www.reduser.net/forum/showthread.php?t=426

[Gunleik Groven]

Panning speed is really a big part of the strobing equation when you shoot progressive.

If the HD cameras adding of sharpening adds to the problem, I can see the point, but to get a nice pan, you'll have to pan real slow and smooth, follow a subject (doesn't take away the strobing, but sure makes it less painfull) or pan real fast.

I see this in film, too.

It's easier with interlaced images.

Gunleik

[Mark Pugh]

gotta say, in the "milkgirls" clip, the girl's movement looks... well... there is no panning going on... it looks like a fast shutter kind of effect. if you look at it frame-by-frame, there doesn't seem to be much motion-blur on the girl's movement. I'm wondering what the settings as far as shutter speed were for that test.. it would be nice to know because I for one don't think it looks that great, I'm afraid. It's worried me a bit since ther first time I saw the clip

[Walter R]

gotta say, in the "milkgirls" clip, the girl's movement looks... well... there is no panning going on... it looks like a fast shutter kind of effect. if you look at it frame-by-frame, there doesn't seem to be much motion-blur on the girl's movement. I'm wondering what the settings as far as shutter speed were for that test.. it would be nice to know because I for one don't think it looks that great, I'm afraid. It's worried me a bit since ther first time I saw the clip

When the "milkgirls" clip was first available, I pulled it into Final Cut to view it on our Sony PVM-20L5 monitor, test some color correction, etc. As I was playing it back again and again, I also noticed a little bit of storbing in their motion. Curious to see what it would look like with some sharpening, I used Final Cut's sharpen filter to harden up the edges. Now that was stroby! I don't consider that a very scientific test, but it certainly makes the case for avoiding any sharpening on high motion or fast pan footage.

You can also see this effect from in camera processing quite easily, if you take a Sony F900 or other camera that can shoot progressive and shoot the same pans and higher motion shots with the detail on and then with the detail off.

I think (and hope) your guess about high shutter speed is probably right, since I'd imagine they were anxious to show the shallow depth of field and wanted to keep the lens close to wide open. Depending on the amount of ND they had in front of the lens, if it was a bright day perhaps the shutter was a little higher than what most people would use on that shot so that they could keep the aperature open. But that its just a guess.

I've also played with some single CMOS cameras with rolling shutters and have found that sometimes I really needed to play with the shutter settings for each type of shot to get exactly what I wanted. We don't know what the actual shutter will be for RED yet, but with rolling shutters I found I needed to go with a higher shutter than I wanted to keep the effect of rolling shutter at bay in shots with higher speed motion in the frame. This particularly true for motion that is verticle or if the thing in motion is tall in the frame. Now there are certainly times you want that high shutter speed look on fast motion, but I hope we'll be making that choice for artistic rather than technical reasons.

Finally, just in case someone feels like I'm attacking Red (a common response here sometimes) I have very high hopes for the camera and do have a reservation. Like everyone else I'm very excited about the resolution, shallow depth of field and the potential for very great dynamic range ... I'm just very curious to see how motion and shutter play out in the footage.

[Nick Shaw]

Stepping frame by frame through the milk-girls clip, the motion blur does look more like what you would expect on film with a 90 degree shutter angle than a 180 degree one. Maybe this is the case. Can anyone from RED comment what the shutter angle (equivalent) was on that shot?

[Finner]

I definately noticed judder with the milk girls stuff. I took it as it was early on in the proccess and that red would fix it. I did not hear any complaints from anyone at the LA showing so I assumed it had been fixed. If there is motion issues like this it will be a huge problem. Anyone want to comment on what the LA footage demo looked like?

Can anyone with red comment on this?

[Emmanuel Cambier]

I feel bad for mentioning it "only now" that other people start talking about it,
but only two days ago I was checking the "milkshot" again for that very same reason.
There seemed to have been an issue then, but like others I haven't got a clue why.
Any comment would be apreciated of course.

Emmanuel

[Michael Ragen]

I remember reading on the HVX forums that Panasonic said the most "filmic" setting for the Varicam was around a 220 degree shutter angle. I usually ride my HVX setting between 180 and 215 depending on the shot. Helps you gather a little more light too.

[Rodrigo Lizana]

I definately noticed judder with the milk girls stuff.

I´ve noticed it too. Since I was seeing the clip on an old G4 Dual 500 I though it was a lack of processor speed. But now that everybody else see the same thing I´m concerned...

[tj williams]

24fps strobes when panned too fast in either film or video. Poi is right on here.
ASC manual refers to this problem as it relates to film cameras and makes recommendations for operators about pan rates.

[Rodrigo Lizana]

The fact and what´s being discussed here is that HD cameras @ 24fps and 180 degree shutter strobe harder than 35mm film at the same settings when panning at the same speed.

[Finner]

The fact and what´s being discussed here is that HD cameras @ 24fps and 180 degree shutter strobe harder than 35mm film at the same settings when panning at the same speed.

Sure this is what started being discussed but this kind of strobe is first year film student 101 stuff and I would expect anyone with a small amount of camera experience to know this.

What has come to light in this thread and is of real importance is if there is a studder problem with the red camera that a bunch of us have seen and no one has mentioned that it has been fixed or did not see this kind of problem in the LA showing. If the camera has a jutter problem like a bunch of us have seen it will be unusable. Which would be important for all reservation holders and much more important then basic camera knowledge.

[Finner]

Let me also mention that I am not jumping to conclusions that this is a problem that has not been addressed by red. I believe they probably have already seen and fixed this problem It would just be nice to hear that it has been fixed by red or that it was only a problem with early footage.

[tj williams]

Additional judder/stutter like effects in movement have been seen in 24P since it's beginning as opposed to film. I think much of this is because we are looking at 24 frames per second in the recording viewfinder and also the playback. This same effect is visible in the film camera viewfinder, however When looking at film in a theatre since 1937 we are seeing each frame projected twice by the projectors butterfly shutter, so in a sense we are seeing 48 fps. This technology was developed to add smoothness to what were then called the "flicks"

Another fact that several posters have alluded to is that sharpening made it worse, and looks like 90 degree shutter. For those not aware usual 180 degree film camera shutter is an exposure of 1/48 second.

Now if you hand hold your slr and shoot at 1/48th the stills will be quite blurry this effect allows visuals at 24 to have a more acceptable motion.

I'm thinking the important aspect of this question is what are the shutter settings on the RED. can we lengthen the shutter opening? for that matter for sports/stills etc. can we shorten it?

Because shutters in digital cameras are electronic, not a mechanical disk, it should be possible to approach 1/24th sec. exposure at 24fps which would add even more blur (than film) to each frame and therefore make the motion even smoother.

Of course when converted to film and shown in a theatre the motion artifacts in milk girls will be reduced by the butterfly shutter. When shown on HDTV the footage will be converted to 1080I with a field rate of 60/fps. We should keep in mind that the HD24P format was intended as an origination format not a distribution format.

[Jeff Kilgroe]

When shown on HDTV the footage will be converted to 1080I with a field rate of 60/fps. We should keep in mind that the HD24P format was intended as an origination format not a distribution format.


Ah, but HD 1080P/24 is now an official distribution and display format for both HD-DVD and BluRay.

I'm thinking that to relieve some of the juddering or odd motion we may see, will require a longer exposure vs. film camera. Or at least that's the only thing I can come up with. OTOH, I think it will depend more on delivery formats. Possibly some new thinking and techniques will need to be applied to D-Cinema shooting. I'm also hoping that RED has evolved a lot since the milk girls shoot and this won't be much of an issue.

[Finner]

TJ

The lab I use e-mails me 24p dailes of film that I have shot for quick reference and it does not jutter anything like the milk girls stuff. Nor is this a standard problem like what I have seen before with HD 24p footage, it is different. I am just really surprised that no one on the red team like Grahame or Stuart or Rob has posted anything to clear all this up?

[Rodrigo Lizana]

The lab I use e-mails me 24p dailes of film that I have shot for quick reference and it does not jutter anything like the milk girls stuff.

Quite right, this is different even compared with 90 degree shutter.
As Finner said it has nothing to do with accepted pan speeds, everybody knows that, it has to do with getting the right product. I´m afraid and concerned that this kind of strobe could remain on the production models.
BTW, "bringing this shit to light" doesn´t mean we want to run the magazine on reverse, it just means we want the best for the camera and the whole project !. I also want to know what´s Red opinion on the subject.

Regards

[Rodrigo Lizana]

I´m wondering how does an on-chip shutter works...it just add darkness to a select area of the chip ?. If that´s the case I believe the reason might be there. The moving-mechanic optical shutter might left a remaining effect while it pass and let the light back into the film thus blurring the image. This is just guessing btw. Someone said in here that the D-20 had the same strobe issue as any HD camera so that might be proving my theory wrong...Red guys ?

[Graeme Nattress]

The most likely explanation is that the shutter speed was "off" on that test footage. Remember Frankie is rather prototype. I don't think we'll have this kind of issue on the new protos.

Graeme

[tj williams]

Also maybe part of the problem in milk girls is the playback... on 1/2 size it seems less motion artifacted on my computer than full size? I'm unfamaliar with the way the torrent download works, Could that also be part of the problem?

Finner: I agree that film transferred to 24P is less obnoxious generally than digital cameras original in 24P I've always felt that was because film is naturally softer edged.

Certainly if the original footage played back on the projector at nab looks like the motion artifacts in milk girls the RED folk have a serious problem.

[Graeme Nattress]

Artifacts in any footage you've seen so far are irellevent, because Frankie architecture is not shipping camera design.

Graeme

[Finner]

Thanks Graeme.

I would still like to hear from someone at the LA showing as to what the milk girls and other moving shots looked like.

Anyone who was there your comments would be great.

[Rodrigo Lizana]

Artifacts in any footage you've seen so far are irellevent, because Frankie architecture is not shipping camera design.

Graeme

We are the rebelion inside the rebelion...We want new footage !! :ninja:

[tj williams]

If current workflow at NAB -40 is good it shouldn't take all that long off task to show us some current 1080/24P footage replacement. The current footage is disturbing!

[Emmanuel Cambier]

We're disturb… we're disturb… we're the most… dis… turb, etc
sung on West Side story's "Officer Krupkee" tune.

I wouldn't mind the new footage though:tongue:

Or just a more firm confirmation that all is fine in Red's world.

Emmanuel

[Rodrigo Lizana]

I would still like to hear from someone at the LA showing as to what the milk girls and other moving shots looked like.

If the shutter was off when the scene was recorded, then the milk girls footage should have looked the same.

[Graeme Nattress]

By "off" I don't mean turned off, I mean somehow "wrong".

Graeme

[Sean]

Mike Curtis of HDforIndies mentioned a distinct strobe he saw in this footage when it was projected waaaay back when. I wondered why no one really took up the discussion further. Maybe simply because the footage is not at all from the final shipping version of the camera. Still, the footage is all we can toy with for a few more weeks (impatient lot that we are).

[tj williams]

Graeme, How about a sequence called milk techs so you wouldn't have to get models etc. just go outside and blast a few seconds for a torrent download!

Seems a problematic practice leaving, footage with problems on your web site?

If there is not a few minutes to redo it's, probably A good idea to take it off the site and wait for NAB 4K for us all to see it. Better than more people mistaking these images for shipping quality images.

[shaftbond]

wow, you guys are ferocious!

i believe i read somewhere they aren't going to release new footage until around (probably after) NAB. (I don't feel like searching for it though). I know you guys are concerned about the "strobe" on the alpha footage, but you have to remember that any other company wouldn't have shown you any footage from a prototype. The purpose of releasing the footage wasn't to say "Hey, this is what the final footage coming out of the RED is going to look like," but more "Hey, our early test footage looks THIS good, can you imagine what it will look like coming out of the final build?" It is rather odd that the footage has been out for a while and now everyone is concerned? Graeme said the shutter wasn't working correctly. Take his word for it. The RED team has done nothing but impress up to this point.

if any of you had put down a nonrefundable deposit, i could see your point. but, if you see the footage at NAB (or whenever it is released on the internet) and aren't happy with it, you get a full refund (maybe even 110% if Jim wasn't saying that tongue in cheek). We should be hearing what the final pricing will be on the RED and all the accessories pretty soon. Get excited about that!

[Emmanuel Cambier]

Hey shaftbond!

I get your point and I'm 99,99% certain this is a non issue and that The camera will break new grounds in this regard as well.

But you have to keep in mind that strobing problems exist in Digital camera that are shipping and in programs that are broadcasted "now".
So when you see that little stobing in the milk shot, I believe it's totally legitimate to ask what's the deal on that.
I believe this is our "job" as "reservist" to inquire about such matters, my only regrets is that we are starting this thread a bit late, I feel lame for a rebel those days.

Once again I am almost certain this is a non-issue, but I want this camera so bad, that waiting for a refund is not something I can consider heartlightly.

But I'm practicing your mantra : Lets get excited!

In Red We Trust

Emmanuel

[Rodrigo Lizana]

Like Emmanuel says, we all want the best for the camera. It´s "our" camera too. If this problem was happening to the F23 for example I wouldn´t give a &%®€®†$&$.

[Jim Arthurs]

I remember Jim J discussing the shutter with Obin on a different forum... start here and read down...

http://www.dvinfo.net/conf/showpost.php?p=552563&postcount=42

[shaftbond]

Hi Emmanuel-

I don't want to imply we shouldn't ask questions. I love these forums and I love people asking "why?" and "why not?" and making suggestions. Obviously the strobing is something to pay attention to when we are able to see the newer footage. I agree with you completely. I just feel that with a thread this long and multiple replies by Graeme, we are getting as much of an answer as we are going to get for now :) I'm not saying we should forget it.

For what it's worth, I was at the LA screening and I didn't notice the strobing...

Cheers-
Todd

[Graeme Nattress]

For now, yes. You know we read what you write!

Graeme

[Stokestack]

Interesting how we all assumed (or hoped) that the milkgirls strobing was somehow due to our systems or some other anomaly. Initially I stepped through it frame by frame to diagnose it, but it wasn't really clear what the problem was.

I just forgot to ask about it. Glad somebody else did.