GH2 GH2 24p or weird strobe effect

so....
some people claim this sample definitely proves the there IS something wrong,
and some claim it's natural for 24p..
oh, and some (like me) are on the fence.. (my testings look fine to me, but i didn't do this exact pan on a similar location, so, who knows?..)

what will be the end of this ?????
 
roei z said:
so....
some people claim this sample definitely proves the there IS something wrong,
and some claim it's natural for 24p..
oh, and some (like me) are on the fence.. (my testings look fine to me, but i didn't do this exact pan on a similar location, so, who knows?..)

what will be the end of this ?????
Click to expand...

To me the 'end of this' will be if you show a film, and the average, even perhaps the 'student' of film, runs out of the theater complaining that your film gave them near epileptic fits because of the 'strobing' or other visual artifacts...

Then you can worry...

I don't have the time at the moment, but I'll see if I can think of more 'definitive' tests to see what effects various speeds of objects have on the GH-1 capture capability at 24 fps, as I don't have a GH-2...

I'll try to give the setup in terms of 'object crosses entire frame in X seconds, etc.' with the camera mounted on a tripod to eliminate any randomness, and of course try to keep the motion smooth as well... Because I wonder if 'ghosting' could also
be due to sort of a swift 'double' exposure type effect...

In film film cameras the shutter, and how it crosses the film plane does affect certain types of artifacts that can be produced with 'things in motion'. Ancient large format press cameras (still camera) had a slit shutter that would travel across the film plane, in such a way that the fast moving object, say the circle of a car tire, was distorted to look like an ellipse, and tilted, such that it actually gave an 'impression' of speed.. More modern 35mm still cameras wold also have a slit, and depending one would get distortions as well... The shutter on a movie camera would cross in a 'diagonal' fashion, etc. Any of these would perhaps be covered by blur, but if one looked closely enough some of the non-blur artifacts would show through...

Digital cameras are no different in capturing 'still' images of moving objects. Just some of the artifacts that are created, are a bit different than those created with film, because of how the sensor is dealt with during the 'read out' phase, as well as the quantization of the 'continuous' image. Film film has a certain natural 'randomness' to the quantization process, in that film grain varies in both size, 'that is how intense a light it will respond to', and spatial relationships one with the other, that is where image 'intensity data' will be fixed, which is unlike current regularly spaced/placed pixel based sensors of the digital devices.
 
I just played your clip. There is ABSOLUTELY NOTHING WRONG with it. Welcome to the world of 24P judder. What many people don't realize is that high contrast boundaries accentuate the judder effect. And your black vertical bars against a white background is a worst case scenario.

Do not, repeat, do not return your camera. A replacement will do exactly the same thing.
 
Paoloperelli, thanks for loading that clip - it gives us something concrete to discuss (no pun intended).

I agree with Matthew P - on my machine that clip looks normal for 24P, with no signs of 'doubling'. Here's a comparison of your still image versus one that I captured from frame 800 of the clip you loaded. The obvious difference between the two reinforces my suspicion that there's a some kind of problem with software decoding/playback of GH2 footage on some machines.

GH2_24P.jpg
 

Attachments

  • GH2_24P.jpg
    GH2_24P.jpg
    64.3 KB · Views: 0
Once (only once) I shot a GH2 clip that would jump back half a dozen frames (or a bit more) every third of a second or so.

Decoder issues.
 
Uwe Lansing said:
I have worked with the D90, 7D and 550D/t2i so far. But i´ve never seen such a strobe/judder... or whatever effect like this here with the gh2. I have noticed it at the very early footage from Japan: http://www.dvxuser.com/V6/showthrea...ssion-thread&p=2131365&highlight=#post2131365
Click to expand...

as said here, this is just native 24p, not the implemented canon or GH1, so it has this look.
the GH2 sharpness makes it looks more precise. i guess you can make the footage look like canon 24p if you degrade it in post.
 
Uwe Lansing said:
I have worked with the D90, 7D and 550D/t2i so far. But i´ve never seen such a strobe/judder... or whatever effect like this here with the gh2. I have noticed it at the very early footage from Japan: http://www.dvxuser.com/V6/showthrea...ssion-thread&p=2131365&highlight=#post2131365
Click to expand...

Then what is it? The only difference in the cameras you've mentioned is that they have FAR less detail, which I think is the sole cause.

There are only a couple of things that can cause judder, and none of them hinder the GH2. I think it just captures images so well that each frame looks like its own STILL image, rather than part of digital motion (no trailing etc). So really, one could say that it acts more like a film camera in the motion dep. than a video one.

There's only one thing that can cause judder, and that's uneven time between frames. The GH2 most certainly does not suffer from this, and when the footage is played back on a 72hz monitor it looks perfectly even. If you don't have access to a 72hz monitor (most support 48hz though), then simply conform 23.976 to 30.000 in whatever software you use (I suggest After Effects), and you'll see that there's absolutely NO judder at all, because the GH2 captures motion perfectly, and sped up to 30p it will match your monitor's refresh rate, proving that the captured files are perfectly fine.

Now, with judder out of the picture (like what I did there? ;), we must address strobe. Strobe can only really be caused by panning too quickly, especially with high contrast areas. The cameras you mentioned probably don't suffer from this as badly because they're so soft, so the contrast is not there on some edges.

Shutter speed seems absolutely fine, so that's also not a cause of the 'strobe' that people are 'seeing' (or more likely paying attention to).


roei z said:
i guess you can make the footage look like canon 24p if you degrade it in post.
Click to expand...

+1 Absolutely. Even a simple blur works.


Can someone please record a video with the GH2 and a camera mounted on top? Pretty please? :) That way it would be easy to say "NO, there is nothing wrong with the GH2 and motion". Which seems to be the case... people just can't find anything else to complain about with this camera, so they pick normal 24p to complain about. Me included... though I've now concluded that there's nothing wrong with it. Show me some sharp footage from another high end camera that doesn't show the same judder (mounted on top of a GH2), and then there's an argument for the GH2 having bad motion rendition.
 
as already discussed before, the really expensive RED etc,
specifically developed a dedicated motion blur that make the sharp digital native 24p look more fluid.
with the gh2 we can handle this in post.
i see it as getting a more natural raw sample, and have all my options to do what i want.
really, no need to return a camera and get the exact same one. afraid of the native 24p? go back to canon.
 
I did a test with my 5DII and GH2. I have done video with the 5DII (and 7D) for several years and just recently added the GH2. This is a slider move in studio with objects and color charts to see what motion looks like from both cameras.

5DII is set to my flat settings 24p
GH2 is set to Cinema 24p with flat -2 settings

All I did was use the 2 way color corrector and grey balanced them to the same 3rd dark grey on the Color checker passport.

Both went through Log and Transfer via FCP 7 and the output is an iPad sized file from my Compressor preset as an .mp4 file that I have used successfully for the last year. I set both cameras from 1/30, 1/40, 1/50. 1/60. Lit with just 2 LEDs for quick setup.

The GH2 is certainly sharper in static frames and still looks sharper in motion. While I do see a type of strobic effect, both cameras seem somewhat consistent. File is not 1080 but if everything looks fine here then for my work and clients, it's all I need.

I'm interested in people's viewpoints from those with cameras they think act weird and from those who are happy with their GH2.

link to zip of video file
 
Graham,

I am afraid the pictures you're comparing are from different moments of the clip I uploaded, meaning that
-the less twiched is from frame 800 when panning was slow
-the one I got with all that twiching thing is further down the clip when panning is reasonably fast

Besides I have decoded my .mts files with Cineform, well.. same thing. To me software is out of the equation here.

I have attached two more frame grabs, one is in motion the other is a still. See the poles, they are twiching not only blurring when panning. It looks like there is more stuff in the picture when panning!

I have seen the 5d vs GH2 panning comparaison and they are doing pretty much the same, GH2 is possibly sharper assuming the lens used was the same, but to my eyes they are not fluid.. both of them.

Sorry guys, I might not return the camera but simply sell it on to who might like to take pictures with it.

Soon enough I will post some footage I pulled out a DVX100, 24p. That is what film should look like, no matter the refreshing rates, shutter, compression and all that stuff.

thanks a lot for helping me out
 

Attachments

  • port2.jpg
    port2.jpg
    78.9 KB · Views: 0
  • port.jpg
    port.jpg
    57.3 KB · Views: 0
paoloperelli said:
Soon enough I will post some footage I pulled out a DVX100, 24p. That is what film should look like, no matter the refreshing rates, shutter, compression and all that stuff.

thanks a lot for helping me out
Click to expand...

interested to see the differences.. still, might be something to do with SD footage VS much sharper HD.
 
> the one I got with all that twiching thing is further down the clip when panning is reasonably fast

Which frame number?
 
The banding in shots with the sky at times is something i have never had with any cam or canon dslr is for me the one big dissapointment plus not having the kit to edit 24P which i cant blame the GH2 for.
 
"huh!! at last, now it's all over for good." Really? I don't think so. I believe that it's strange that these 2 cameras come out and both have a handful of users making these claims. Just something to think about. I don't think there's a conspiracy or anything. Just a few messed up cameras. Maybe just one? I would encourage Roei to fill me in on what was up with my original clips before he makes such a bold claim.
 
I did a test on this sort of issue the other day. I used a chain link fence as my reference structure, and a house in the background behind the fence, light blue, with pure white trim, and bright sunlight.

I walked along the fence, ad a 'reasonable' pace, so the chainlink pattern would be moving relatively fast to the background. I was at f/16 @ 1/60, 24 fps, at ISO 100. Pretty bright, pretty deep DoF, etc.

Upon looking at the footage captured in single step mode, via CS5 Premiere Pro, and 'native' AVCHD, I saw what I beleve are the encoding 'optimizations' for the GH-1 and AVCHD...

At which point, my thoughts turned to... I know why I don't have any time to shoot actual narrative short films... which is my present desire... I'm on my lunch break, and the Wife has my weekends
filled with things that are on her agenda, furthermore, I don't have a bevy of hot chickie babes that I can tap into to really put my camera through its paces... or even a bunch of buddies to actually
collaborate with to make any sort of film, artsyfartsy or not...

What I want to know is, what are other people's excuses for not actually making a film...
but rather spend many wasted hours 'testing' and finding the most absurd artifacts that can be known to the lowend digital camera user?

For over a 100 years filmmakers have shot film, with each era having a set of 'wyrdnesses' that the then technology did not compensate for, whether it was ISO 10 film, and having to shoot on a rotating
outdoor set to 'follow the sun', to early Technicolor which had to be so brightly lit that the crew would wear sunglasses. late 60's 'push processed' film that had grain the size of golfballs so the
filmmakers could break out of the studio, and its constraints, etc...
 
that is non sense.

In the past people had to comply with limitations but that was the standard.

These days I cannot shoot professionally having that kind of fault.. It would be nothing but true disaster.

And yes rigby, I will return my machine as the end of it. There is something weird going on the software side.

Progressive means that there are sequential photo stills, bottom line. Here I have doubles and how can that be the end of it?
 
Rigby said:
"huh!! at last, now it's all over for good." Really? I don't think so. I believe that it's strange that these 2 cameras come out and both have a handful of users making these claims. Just something to think about. I don't think there's a conspiracy or anything. Just a few messed up cameras. Maybe just one? I would encourage Roei to fill me in on what was up with my original clips before he makes such a bold claim.
Click to expand...

listen, back then i watched your clip imported to my NLE and it looked just as my clips, regular 24p... (+ panning on contrasty skies 'effect').
all i'm saying is most chances you'll get the same from a replaced camera..
and after viewing the last 5D & GH2 pan test, it looks the same as your clip, no?
so.. look like enough evidence (unless this 5D has the same defect as the gh2 this guy has, and probably on many other Panny and Canon)
i don't know a bout you, but i started shooting a feature with it, so i personally am at peace with myself on this.
so i am NOT defending anything- do whatever you can to get another gh2, shoot the same pan. if you are 100% sure your first gh2 has this defect, than it's your call.

i saw at least 10-15 tests these last threads, including other cameras, and i don't see a big difference in you footage, really.
didn't mean to sound SO bold, but it's kind of a gut feeling. perhaps you should ask Barry on a last call, i'm just a little fish in this crazy tank.
 
You must log in or register to reply here.
Back
Top