surely if this issue is so prevalent and obvious, it should be pretty easy to point to any number of vimeo videos with this problem. there must be heaps of 1080p24 videos with panning where it would be visible.

i look back at all the 24H video i've taken and i can't see a problem at all, all pans are nice and smooth. the only big thread i read about this and after 9 pages it turned out to be a playback problem.

there are so many variables here:
shutter speed
panning rate
in-camera adjustments
as well as playback variables:
playing back on computer? is it the raw mts or transcoded? what decoder (ffdshow/ffmpeg, internal vlc, mpc-hc w/dxva, coreavc, ...)?, what renderer (evr, evr sync, haali, madvr, vmr-9, opengl, ...)? tried using ReClock?
output device (lcr, crt, projector)? refresh rate of output device (50/60hz, 23.976hz, 120hz)?

i might even go and shoot a quick pan tomorrow and upload the raw mts so we can all be looking at the same source footage as a comparison.

i am genuinely curious as to what this issue is, mostly because i don't have any other raw footage to compare it to so i'm worried there is a problem here that i don't perceive but would be visible for others.

Without one-to-one comparisons, it will be hard to say. It seems every new video camera goes through the same question regarding overall motion and judder.

Grab another video camera and set the frame rate and shutter to the exact same settings. Next, carefully setup your shots so they are as close as possible in capturing the same motion and speed across the sensor.

@Dalefpf:

Ok, since I'm a bit ignorant on what people are calling strobing I'll post some examples that I've come accross.


I remember the following video from a few months back by Gunther Machu. This was a side by side test with a hacked GH-1 vs GH-2. To me...there was no difference in motion between the two. I see Blackout's request for him to film a more crowded area like a beach etc., but... a motion difference is not evident in this video (IMO).:



Or... does strobing mean what PC CHongor mentions in this very recent video (where there is noise and pixelation in the dark areas)?:



Or.... is strobing the effect we see in the following video @ 1:47 - 1:55? This is the type of strobing I am very much familiar with. This happens in every kind of video camera when it comes to using the wrong shutter speed while viewing monitors or around flourescent/street lights etc.:




The following is a video test done by Mr. Barry Green displaying the effects of LED lights used with ALL CMOS cameras (including Red, F3, FS100 etc., etc.). It just so happens he is using a GH-2 for this test. I'll quote Barry's statement in this test:

"This was a sample video to show what effect a dimmable LED light will show, when used with a rolling-shutter CMOS camera (in this case a GH2). The more you dim the light, the more prevalent these scrolling bands become.

Certain lights (like the CN126) can be dimmed without these scrolling bars showing up, but some lights (like the YN160) cannot. The YN160 dims in discrete steps, it has a +/- button set that you press to move the light to the next dim stage. The CN126 has an analog wheel that you turn to progressively increase or decrease brightness.

The YN160 works great and strobe-free at 100% brightness, but at any dim level the scrolling bars show up. You can try to chase them away by changing the shutter speed, but if you were using two of these lights you would probably soon find it impossible to get a strobe-free look. Unless you used them always at 100% brightness.

Note -- this issue only affects CMOS camcorders; I tried the same test with the HMC150 CCD camera and there was no issue whatsoever. But any CMOS camcorder with a rolling shutter will show these same scrolling bars."



Now this last video is something I would probably consider the closest to what you guys are talking about. It is a video with lots of dolly movements and one could really see a very ugly "strobe-like" effect in many of the shots. I've never seen this effect in my camera...but if this is what you are referring to then yes...that is a problem. This is one of only a few videos from the GH-2 I've seen it in. Then again...there are a lot of other things wrong in this video (Vimeo compression is one). I also remember early on seeing it in other cameras (especially GH-1) but folks called it "pulsing effect.":






As you can see these examples are different from one another. Now you can hopefully understand why I asked what I asked in my last post. With that said...I'm not even sure if everyone here is referring to the same thing. So...when it comes to the subject of this thread it goes back to you Dalefpf. Can you point out in any of the linked videos above what you are referring to... or maybe provide your own example? I really want to learn more. Thanks.

@Dalefpf

I think that it would only be fair if you did allow us to see the footage that you are seeing the problem with so that others can chime in with their individual observations. It's not about raining on anyone's parade - there are lots of experienced people here as well, wouldn't it be helpful to open up the dialogue with direct obervation of the footage in question?

@Elenion

After more panning tests with a pro kine tripod & head, I still can't tell if it's something in the camera or just the nature of 24 fps viewed on a 60p display. Shooting at 1080i60 or 720p60 are both MUCH better at smoothness, but so is shooting at 24H 80% and conforming back to 30. Unfortunately I don't have any other 24p DSLR footage to compare to.

The GH2 strobing reminds me of screening rushes at IMAX in 24 fps back in the 90s. The giant screen really exaggerates the strobing problem. IMAX experimented with a 48 fps format to try to solve the it, but the stock costs were prohibitive. And of course Doug Trumball had Showscan running at 48 fps and now talks about 60 and 120 fps as future standards. 120 would be great because 24, 30 and 60 could all be derived from it without artifacts and of course it would produce fabulous slow motion .

Strobing is also pretty visible sometimes in regular 24 fps theatrical movies when viewed on on a big screen.

For now I'm satisfied with the motion quality of both the conformed 24H 80% and 720p60. Both options are slightly problematic though - the 24H-80% has no sound, and it pushes the codec very hard so there are often "card too slow" errors (with kae's 60m settings). And the 720p though very good IQ looks a little soft when up-resed to 1080.

As a third option, I'm hoping that someone will pick up the ball to get the 1080 interleaved>progressive patch working properly, to give us a full-HD option without strobing.

...um...is it me or does it seem strange that all of these claims of judder don't have any accompanying videos for us to look at? I sincerily have no idea what Dalefpf is talking about. I've heard of these judder claims before with the GH-1. But I've yet to see someone show an example of this "judder" (that is supposedly different from any every other 24p camera).

Now if one is referring to the "tearing" that we see in panning etc.,....then yeah...I always wondered why that happens (in GH-1, GH-2 and really....most CMOS footage I've seen). I don't see that kind of "tearing" in Apple movie trailers that originated on film. But I have to say...the GH-2 looks no different from any other HD camera (when it comes to motion).

Dalefpf please post an example of what you are talking about. I've watched a ton of GH-2 videos with lots and lots of motion in them and understanding that your "terrible" opinion, in regards to its motion cadence, is purley subjective mine is nowhere as extreme. Seeing is believing.

EDIT: What also comes to mind is how some folks with higher end cameras (like Red, AF100 etc.) have been seamlessy intermixing their footage with this camera. If there was that much of a difference it would (or should) have reared its ugly head by now....don't you think? But now that the hack is in full swing then......

Oh, just to clarify, he used a 23.976 timeline and a 24 timeline. Okay, I'm out.

Go for all GH23s or whatever u call it. 24p is stable.

Try hdtv natively supporting 24p display. That is 23.976 fps display. BTW I don't own such tv.

How about this for a test - rig up several cameras to the same tripod head and outfit them with either the same lens or lenses that will give the same equivalent view on a given sensor. So you could have GH1, GH2, 5DmkII, whatever else, and then shoot a series of panning shots. Do it at various speeds, and if you perceive a difference, show us a split screen with the GH2 on one side and the camera you feel has better motion rendition on the other.

I think the perception of motion is a tough topic for many, only some of us really notice whether the cadence feels right or not. You even hear people saying things like 'most viewers can't tell the difference between 24p and 30p' - well to me it's the most obvious thing in the world. Some people leave motion smoothing interpolation enabled on their 120hz/240hz televisions whereas I find it utterly unwatchable. Even with it turned off, I can see a difference in how my 24" Dell LCD renders 24p versus my 40" Samsung LCD, even bluray at 24hz (and I prefer how it looks on the Dell, unfortunately).

It wouldn't explain the discrepancies some are seeing in GH2 vs. GH1 motion, but as for digital image acquisition in general, some have theorized it's the difference between electronic and mechanical shutters that accounts for part of what makes film look like film and digital look like digital. You even see this in RED's "magic motion" feature on the Epics; they seem to acknowledge there's more than one way for digital sensors to render motion, and have given people some options. Who knows what the GH2 defaults to?

@JDN

It could very well be post workflow, I don't know. That would be great, in fact, as it would be fixable. I use FCP7 log&transfer to ProRes422 for most projects. I also use 5DtoRGB to convert footage, though this is still in the testing phase. Should I be considering a different workflow?

@Dalefpf

This 24P artifact bugs me but I put it down to small stuff. The posterization/banding and CMOS jello were much bigger issues to my mind. The first seem to be greatly improved by the firmware hack and the second unchanged (and endemic to most DSLRs). Still, I think the GH2 is a fantastic camera. Every camera system includes some kind of tradeoffs. For me, the GH2 is the sweet spot. I'd like to try to dial in the look as much as possible, but I certainly wouldn't stop using it just because of slightly weird motion on pans. In most types of video work, if your audience is thinking about the pans and judder, you have a much bigger problem...

@dalefpf try two things: first find in this topic my suggestion on panning rates and follow it. Do you still see strobing motion in your footage? Do a review on a powerful enough computer (plays fine on a 2009 MBP, also on a 2.66 Quad core duo Windows 7 desktop) If you are on Windows, use WMP, not VLC for a playback. DO not judge based on the viewfinder.
Secondly, do all of the above but turn off the autofocus or put it into a single mode, rather than c mode.
Report.

@Dalefpf

I can see it clearly as well. When recording in 24H the LCD looks like it's updating at about 12-16 fps. When I first got the camera I thought I was looking at the old silent film rate of 16 fps. I'm suspicious that Panasonic is doing some tricks to make the 24h output look more like "cinema with afterimages" as it says in the manual.

I just did a bunch of tests panning in 24H vs. 1080i, 720p and MJPG modes. The 24h panning is definitely a lot more strobed. This is not just the difference between 30 fps and 24 fps - I've been dealing with that for years. This is something else.

@Vitaly
What's the current status of the "1080 progressive" patch in ptool? is it worth trying this to get a viable high-res alternative to 24H?