ATV vibration causing HDV problem?

[Ken Eberhard]

I'm not sure how to take that posting. "The rep said" (What, the break away features of parts?) "was due to allowing the camera to function better under high vibration, such as found on boats, motorcycles, etc." That's a strange statement. Maybe it's just not coming across right. While at the head of the posting it says they where having all kinds of image problems. Not very reassuring, it does show an awareness of the problem I am having though. At least it seems so. And if so, it doesn't make me feel they fixed the problem on that unit either. With keying problems and imaging problems under minor vibrations. It all makes me move further away from the format. Don't get me wrong. I love the image too. I understand a lot of folks loving this camera, the Sony, to make that statement clear. If I was still shooting musical performances with muti cams. It would be great I bet... I'd be altogether on the other side of the fence.

[Charles Papert]

Once I pictured what you were doing with that ATV, Ken, I must say I was astonished--it had never occurred to me before that a one-man band driver and operator set-up was even possible.

Regarding the wind issue, yes, it is a problem with vehicle mounts. Some use gyros to help with this (I saw a Glidecam at NAB a few years ago with gyros on it--interesting, but I'm willing to bet that no Glidecam operator out there is going to be spring for the money for those). Wind screening is a great way to go, although for a forward-looking lens position not all that easy. The optimal setup for a physical wind break when the wind is comingn at the lens is to set up two permeable (such as a double net) blocks in a V formation behind the camera, with the open end of the V facing the camera (the camera almost inside the V). This will create a little bit of a calm zone where the camera flies. Not perfect, but better than not having one at all. I usually end up on the rear of an ATV looking backwards, and have had a windblock rigged right behind the rig, which is very effective.

[Steven White]

I set up a thread to test your theory Ken.

http://www.dvinfo.net/conf/showthread.php?t=50094

-Steve

[Ken Eberhard]

I'll take a picture of the mount later today and post it... Maybe I should send it to your referenced site too. But then again, should I let my idea out.. oops to late.. besides, sharing knowledge is the only way to be in this business. I've fired engineers before for keeping things to themselves trying to make themselves indispensable, doesn't work with me. We are all learning... With my mount I have forward, sideways, right and left perspectives and rear... just a spin of the sled's post... I'll post my photo late PM most likely.. Ken E.

[Charles Papert]

Looking forward to seeing it. And I wouldn't worry too much about "letting it out"--I think most people would prefer to split the duties of operating and driving...!

[Mark Grant]

To me this just looks like perfectly ordinary interlace artifacts: I honestly don't see what the problem is.

The only difference I can think of between the Z1 DV and PD170 DV is that on the Z1 you're scaling down the DV footage when you letterbox, whereas on the PD170 you'd have native 4:3. That will make interlace artifacts even worse.

[Ken Eberhard]

Okay Charles, here's the picts of my GlideCam rig.

http://homepage.mac.com/kene3/GlideCam.html

Ken E.

[Charles Papert]

Nice job, Ken!

You can really see how much flex is occuring throughout the suspension system--even though your mounting point is biased backwards maybe 10 degrees, by the time you get to the gimbal pin at the front of the arm you can see that it has shifted to a forward bias of another 10 degrees. I would guess that you are having to hold the rig in place otherwise it would want to fly forwards, yes?

This actually is one of the as-yet unsolved issues with vehicle mounts; unlike the body mounted rig which allows the operator to dictate the floating position of the rig through his hip position, with a hard mount you are truly at the mercy of the topography. I'm sure Ken will back me up that as you traverse a grade with the vehicle, the rig "wants" to shoot out in the downhill direction. Same also with corners; the centripetal acceleration requires a good grip to rein in the rig.

In any event, Ken, I hope your format issues work out and you are able to keep shooting--best of luck with all of it!

[Ken Eberhard]

It does want to move out from me some. But I don't find it a big problem. I can actually let go of it for a moment and it will float, not for long. I think I should be using the V16 though. I can see that this light weight unit is getting stressed with the extra weight I have on it. I've got the extra plates bolted under the sled. This is really a great thing to shoot with. I could also use the lower monitor because the on camera monitor is out of sight. I point and shoot mostly. The wide angle makes it work out. I get pretty good at it after awhile. I've done some pretty bumpy, hilly terrane and gotten great footage. You may notice I have the mount tilted up some to compensate for the bias. I can swing my arm out and tilt the camera down to get a shot of the wheel too. Or across the front of my body, and even a side profile of myself riding along.

Lots of fun down here in Baja with this.. Ken E.

[Carroll Lam]

Ken has an interesting, and for him serious, problem. I certainly can "feel his pain" 8-)

As I will discuss following, I don't believe it is an "format" (HDV compression) problem. I believe it is a vibration isolation problem.

Based on looking at Ken's first clip and the tests and comments he has made I am of the strong opinion he's dealing with a vibration-band-limited problem. Ken has stated that the problem is reduced at higher rpm.

And I also believe the problem he is seeing is physical (mechanical vibration somewhere in the optical chain) and not mathematical (due to the HDV compression algorithm). To my viewing the video clip shows the imagery vibrating pretty well in consonance with the ATV's.

Even though Ken stated that he could see no vibration at the lens I don't believe one can see the level of vibration causing the image "jitter" shown in the clip.

We obviously know that the problem starts with the ATV vibration. The fact that Ken sees it with the FX1 and not the PD170 can be due to at least factors - the two times higher vertical resolution of the FX1 and a potentially greater sensitivity to vibration for the FX1, or a combination of both. Also, the resonant frequency of th mechanical unit (the Glidecam plus camcorder) can be affected by the mass of the camcorder. The lower weight of the PD170 could shift the resonant frequency to a point not affected as much by the ATV at idle.

If we assume ( a BIG assumption in this situation) that Ken's camcorder is representative of all FX1/Z1U's, that is it doesn't have some anomaly causing, then the question is whether the ATV vibration can adequately isolated from the FX1 to permit good imaging.

In one posting Ken stated "Are you driving with the camera handheld, I'm sure there is no problem." To me this implies that Ken believes that a handheld FX1 would not be affected by the ATV vibration.

Is this true, Ken? If not, would you do such a test - handhold the camera while someone runs the ATV through the lower rpm range? Let someone else control the throttle so you have less coupling to the ATV.

If such handholding will eliminate the problem, then I would assert there is non-handheld solution possible.

Glidecam-type devices, by their very nature are "low pass filters". That is they allow for slow movements of the camcorder while removing higher frequency motion. But any such filter has a "corner frequency", the frequency at which the filter starts reducing higher motion frequencies. I believe in this situation the Glidecam is not rolling the higher frequencies off fast enough to eliminate the camcorder vibration.

Although not the same situation as Ken's I have posted a 3 min clip on my website that provides some footage of a car-mounted Z1U. I took the original video as a test of a windshield suction cup mount arrangement which can be seen in the clip. I made no particular attempt to minimize the vibration from the car to the camcorder. As can be seen in the clip the camera is resting on a white polystyrene block that isn't particularly soft. I have consider retesting with my Cine Saddle (bean bag) to see if that improves isolation but haven't had a chance to do that yet.

The clip is here: http://www.clamcamvideo.com/gallery.html#mount

I've tried to include samples covering fairly slow speed on a broken up paved road to fairly fast (50 mph) on a fairly smooth paved highway. At about 1:38 there's a scene of a roadside desert plant taken with the vehicle stopped but with the engine idling. At 2:02 there's a section of the pavement that is essentially flagstones. This paving ends at 2:30. The vehicle used - a Scion Xb - has some engine-induced vibration at idle that is reduced at higher rpm. I believe most of the vibration at speed comes from the road surface.

I have at least one alternative to the Glidecam that Ken might try to solve the problem but I'll save that discussion until we hear back from Ken.

Carroll Lam

[Ken Eberhard]

Hi Carroll, Your clip and comments are very welcome. Your WMV file works pretty good, but it stalls frequently even with my 1 meg download speed. But as I examine your clip I see several places that seem to have the same problem I am having. One that seems to show it best is just over half way in you pull over to the right in a large pull out, after the bloom shot. As you start to pull back on the road look at the rock wall to the right. It seems to go fuzzy and I seem to see a doubling of the top wall line. I know your expecting some vibration distortion from your mount. But this may be more than that. It is very similar to my problem. Also, right at the start in the woods, the branches look like they go into a soft focus. Look at the logs as you pull back, there is one spot they look fuzzy too. This is another characteristic of the problem. Hi detail goes blurry, but on close examination it appears the camera can't get it straight, it has nothing to do with focus. Early in the video you hit some pretty rough stuff and the camera jumps a lot. I see traces of the problem there too. One test you might do, if you have one, use a DV camera, not the Z1 in DV mode, but another camera with the same mount and see if you get a much improved picture. My bet is you will.

And yes, it is a specific vibration frequency range that kicks the problem off. Sony also thought it was a mechanical problem in the lens assembly and they replaced it completely.

I'm going to add one more detail to this discussion. I have a 55 inch Mitsubishi HD monitor here also and I plugged the FX-1 directly into it via component video and the problem is there also. It does not seem to be an artifact of down converting.

And yes to the handheld on the ATV. It does help get rid of the problem. The camera works great until it is vibrated. The best I got was with my GlideCam vest on while riding on the back with someone else driving. But the problem is so prevalent even then that I didn't even want to consider cutting around it in the edit.

This test is perfect for the problem. I'm seeing unusable footage in this clip. If we can get the same test with a normal DV and compare, we all might come to some conclusions.

Thank you all for putting up with me... Ken E.

[Ken Eberhard]

I might add. The jumpy characteristic of Carroll's clip is also the nature of the problem I am having. This kind of artifact can be the cause of the steady shot. Did you have it on or off? It looks to me like you have it off, because that's about as good as I can get mine to look in the rough. Turning it on only makes mine worse. This jumpy footage I'm seeing maybe from the WMV frame rate or compression though. Is that jumpy look on the master too? The feeling that part of the picture is hanging there for a moment before it goes on?

Thanks again.. Ken E.

[Carroll Lam]

Originally Posted by Ken Eberhard

Your WMV file works pretty good, but it stalls frequently even with my 1 meg download speed.

Not surprising. The average bit rate of that particular clip is 1.5 Mbps.

But as I examine your clip I see several places that seem to have the same problem I am having. One that seems to show it best is just over half way in you pull over to the right in a large pull out, after the bloom shot. As you start to pull back on the road look at the rock wall to the right. It seems to go fuzzy and I seem to see a doubling of the top wall line.

I'm pretty sure the Scion's hood goes through a maximum engine-induced vibration as its rpm increases so that is probably where it most closely matches your problem. You'll note from the pictures the setup the camcorder is basically retained in place by the cantilever mount from the suction cup but its weight is mostly supported by the stiff polystyrene block resting on the Scion's hood. Thus, the camcorder essentially experiences full hood vibrations.

One test you might do, if you have one, use a DV camera, not the Z1 in DV mode, but another camera with the same mount and see if you get a much improved picture. My bet is you will.

I'm sure that would be the case. I've done a lot of similar stuff with DV cams.

I have a 55 inch Mitsubishi HD monitor here also and I plugged the FX-1 directly into it via component video and the problem is there also.

Assuming you were in HDV mode this is another bit of evidence that the problem is mechanical, not mathematical. The component outputs are no way involved with the HDV compression process.

And yes to the handheld on the ATV. It does help get rid of the problem. The camera works great until it is vibrated.

I figured as much. Better vibration isolation means better image quality.

So....here are my conclusions based on the facts to date:

1. A couple of specific HDV camcorders (your FX1/my Z1U) are more susceptible to vibration than similar DV camcorders.

2. A properly setup Glidecam will not adequately isolate the problem vibration frequencies from the HDV camcorder.

3. Isolating the FX1 from the vibration souce (via the human body) is adequate to eliminate the problem.

4. The problem is vibrational motion of the image chain and not based on MPEG2 compression artifacts.

Some things we do not know:

a. Is the problem just due to the expected higher sensitivity of HD cameras to vibration?

b. Would another brand HDV camcorder (e.g., JVC) exhibit the same vibration sensitivity?

c. Would an HD camera (e.g. a Varicam or Pana HDX200 ) exhibit the same vibration sensitivity?

d. Can the current Glidecam setup be "tuned" to eliminate the problem vibration frequencies.

e. Are there other methods of vibration isolation that would eliminate the problem?

Questions a, b, and c can only be answered by someone with access to those units and are willing to do the testing.

Question d needs input from some good Steadicam/Glidecam "tuning" expert (Charles Papert?). Obviously the suspended weight and spring tensions need to be adjusted to try to move its rolloff frequency low enough to eliminate the problem vibration frequencies.

Question e would require tests. Two possible alternative methods of isolation (besides the human body) are some form of "bungee cam" or a Kenyon gyro suspension mount.

Obviously you have an investment in the Glidecam setup and alternative mounts may be prohibitive. Further work at tuning the Glidecam setup may be your least disruptive alternative.

Or, you may not wish to pursue the problem further. As you've stated you are going to put the FX1 up for sale.

Thank you all for putting up with me...

Well, the problem you discovered (and I validated) is of importance to those of us that use our cameras in sometimes extreme conditions and want to move into the "HD age". Your input has triggered a whole host of qualitative and quantitative tests by a variety of video experts on this forum resulting in very useful information and better understanding of the HDV format and its limitations.

We obviously haven't learned everthing there is know...yet. 8-)

Carroll Lam

[Ken Eberhard]

Yes, the problem is with the Sony's HDV cameras... They seem to have two limitations. One is keying and the other is moving motorized vehicles. No big deal if you know that going in... I'm not that put out by it. I just wish it had been put up front.. Like in small print someplace that says "This camera format may not work well when exposed to minor vibrations." Or "in vibrating places other camera formats may work." That way you can go in with your eyes open to testing. Someone suggested that NASCAR is using them in their cars. I somehow doubt it. Is that a fact? And I don't think anyone is going to come up with a mount that will work. As I've stated. I have done hand held and vest mounted GlideCam.. Think about that. The GldieCam mounted on my vest, total separation via my body, plus the GlideCam.. That footage was better, but not usable without a lot of cutting. So I don't think anyone is going to do much better than that. I also put an air tube between two plywood plates and tied them together with straps so that the top board and the bottom board had no solid connections. Then mounted the GlideCam on the top board... No go. A bungee cam might work, but what a beast to control. And I can't think of how to do that on a quad. I bet the panasonic DVCPRO HD will work. In the photos on the link of Glide mounts. I saw some pretty nice cameras there. Any of them HD? I can't be the the first person to try HD of any format on a moving vehicle with minor vibration. I really do think it's the HDV encoding. And I'm still of the opinion the Sony first goes to HDV and converts it's to DV via it's real time down converter before it goes to tape as DV. No one has proven otherwise yet to my knowledge. That would explain the problem crossing over to the DV mode and explain away the idea it's a mechanical problem.

Now that I am sure it is not just my camera. And I can't fix the problem by buying another HDV and simply replacing this one. I am actually considering keeping it. Yes, I love the camera too. I may need muti cams in music production if I get back into that down here. So for now, I'll use it. I can always sell it later and the production value now is greater than the money I might lose by selling it later on down the line.

I think Panasonic is on the right track at this point. Make a reasonably priced 100 MBS camera. Although for me, it will have to be another year or so when the P2 cards come down in price and the capacity goes up.

Again, thanks for putting up with me.. Ken E.

[Ken Eberhard]

One more posting of a movie clip...

http://homepage.mac.com/kene3/HDVtestSequence.html