Shooting observations and HV20 techniques at DVinfo.net

Mike Dulay · June 24th, 2007, 10:58 PM

Just some tips I've (re)learned ... hoping this will inspire you to share your own.

Rolling Shutter/Jelly Effect
After months of shooting exclusively 24p, I had to try 60i again. This is when I ran into the rolling shutter/jelly effect. Having tested the very same camera under the same conditions in 24p vs 60i I noticed a few things.

I'm driving an SUV on a concrete top road (no asphalt). Dashboard is hard plastic. I set the camera on the dash and drive at 60mph. If you listen closely the gaps in the pavement hit the car 1-2times a second. Tip of the camera noticably bounces.
1) 60i-TV500-OIS on ... the video was jiggly while on the dashboard. BUT when I took the camera in my hand the jiggle went away. I guess my body isolated the camera from the car's vibration.
2) 24p-TV48-OIS on ... no jelly effect, just obvious bumps for sharp bumps on the road.

Seems interlaced high shutter speed combined with frequent up/down motion is prone to rolling shutter. Isolate camera from vibration and it goes away.

Moving Shots
The camera is more sensitive for up/down tilt than side to side. While using a manfrotto 506B I accidentally nudged the camera back too far and it sprung back. This gave a shimmer like the OIS bounced up and down. Side to side jolts don't seem to bother it. With a DIY figrig, this bounce is reduced but it's still no "steadicam shot". My best moving shots were done on elevators/escalators/moving walkway.

Spasmic Hand syndrome
The longer I've been at this hobby the more camera shake I develop. Go figure. I'm able to salvage bad HV20 footage (especially 24p) by (1) downscaling to 1280x720 (2) slow motion. It seems that downscaling has a light 'deshaking' effect. Slowing down the motion to .6 (max) makes headache inducing shots more tolerable.

Chris Barcellos · June 25th, 2007, 12:00 AM

Quote:

Originally Posted by Mike Dulay

Spasmic Hand syndrome
The longer I've been at this hobby the more camera shake I develop. Go figure. I'm able to salvage bad HV20 footage (especially 24p) by (1) downscaling to 1280x720 (2) slow motion. It seems that downscaling has a light 'deshaking' effect. Slowing down the motion to .6 (max) makes headache inducing shots more tolerable.

This camera is tiny, and from that perspective it takes less "energy" to overcome the its intertia. That why I think we see more hand shake. During a shoot last weekend, it had it mounted with a makeshift DVD player as a monitor with its battery, with a XLR adapter, all attached to self made mounting system. It was getting heavy at the end of a 20 hour day, but shake wasn't as big an issue as I thought it would be...

Elmer Lang · June 25th, 2007, 01:53 PM

For me, shake is partly a factor of how it's held. I almost never hold it with my hand in the strap and if I do it's the opposite way, but usually I hold it thumb on top, hand on side and fingers underneath. One can loosen the strap and push the hand down through there (instead of up the usual way), but I rarely do.

I have the DM-50 on top, but if I'm serious--well, more serious--I use the Zoom H4 to as a recorder with my better mics. What I don't like about that is that the audio is not on tape, just a little card and though you save to drive, what if the drive breaks. So I also save to cd, what if that breaks. Right now, tape just seems like the safest storage--whoa, maybe I should export audio to tape!

Anil Dasari · June 25th, 2007, 02:15 PM

Quote:

Originally Posted by Elmer Lang

For me, shake is partly a factor of how it's held. I almost never hold it with my hand in the strap and if I do it's the opposite way, but usually I hold it thumb on top, hand on side and fingers underneath.

That's a good idea. I hold the camera in a similar fashion, and I always get very steady shots (except when I'm walking, ofcourse).

Elmer Lang · June 26th, 2007, 02:25 PM

Quote:

Originally Posted by Anil Dasari

That's a good idea. I hold the camera in a similar fashion, and I always get very steady shots (except when I'm walking, ofcourse).

Hi Anil,

Yeah, it's pretty steady, and you can still ergonomically raise it pretty high. And if you want way high (and who would really) you can switch so fingers are on top.

best,
elmer

Duane Steiner · June 27th, 2007, 06:42 PM

Quote:

Originally Posted by Mike Dulay

Just some tips I've (re)learned ... hoping this will inspire you to share your own.

Rolling Shutter/Jelly Effect
After months of shooting exclusively 24p, I had to try 60i again. This is when I ran into the rolling shutter/jelly effect. Having tested the very same camera under the same conditions in 24p vs 60i I noticed a few things.

I'm driving an SUV on a concrete top road (no asphalt). Dashboard is hard plastic. I set the camera on the dash and drive at 60mph. If you listen closely the gaps in the pavement hit the car 1-2times a second. Tip of the camera noticably bounces.
1) 60i-TV500-OIS on ... the video was jiggly while on the dashboard. BUT when I took the camera in my hand the jiggle went away. I guess my body isolated the camera from the car's vibration.
2) 24p-TV48-OIS on ... no jelly effect, just obvious bumps for sharp bumps on the road.

Seems interlaced high shutter speed combined with frequent up/down motion is prone to rolling shutter. Isolate camera from vibration and it goes away.

I am having no luck using the HV20 in a in-car mounted situation. The OIS can not handle the vibration. But when I hand hold it, it is not bad. I have only tried 60i-OIS on in 'Auto' mode. Do you think that the 24p OIS on-manual shutter, will be better?

Mike Dulay · June 27th, 2007, 07:18 PM

Duane, from the driving I did there wasn't a noticeable rolling shutter while I drove but the sharp bumps will cause a shudder.

Duane Steiner · June 27th, 2007, 07:43 PM

I still don't understand the rolling shutter effect. This is the camera on the mount in my car (60i OIS on Auto): http://www.socalspeedzone.com/incartest/incartest.html

Andrew Plumb · June 27th, 2007, 09:28 PM

Duane, how are you mounting the camera in the car? Got some pictures of the setup?

I ask because watching that video, the dash is bouncing around in the picture as much as everything else. My first guess (coming from a control systems engineering background) is that the OIS is compounding the problems - it's reacting to every jolt the whole car is subjected to. You want the optics to be locked w.r.t. the car's interior frame of reference.

Try explicitly turning off OIS. If you can get the car interior to appear stationary w.r.t. the camera, then play with 60i versus 24p to see if rolling shutter is an issue.

Only one way to find out! :-)

Andrew.

Mike Dulay · June 28th, 2007, 08:11 AM

Duane, I didn't notice any rolling shutter in your footage. Rolling shutter effect makes everything look like you're looking through a piece of jello ... straight objects leaning left and right in alternating pattern. What you have is camera vibration, the OIS isn't able to compensate for that kind of vibration. Like Andrew suggested, turn off the OIS. If possible, get the camera mounted somewhere it moves with the entire frame or isolate it from the vibration somehow.

I haven't tried race style driving and your suspension is probably stiff so I can't suggest exactly where to mount it. When you rev do you shake with the car as well? Maybe there's some literature on how the people at Top Gear/Fifth gear mount their in car equipment?

Andrew Plumb · June 28th, 2007, 09:18 AM

There may also be some secondary complications coming from auto-focus mechanisms as well. You'll probably have to turn off IAF because the camera's looking through the car's windshield.

You may have to engage Manual Focus if the smaller vibrations are being interpreted as an out-of-focus picture. Probably set it for infinity (hold down the Focus button for 2 seconds) if the horizon is what you're interested in capturing, or just set it for some fixed point in front of the car.

Andrew Plumb · June 28th, 2007, 11:40 AM

If you're of the DIY mind-set you might want try adapting this bicycle mounted steadicam design for in-car use. It would absorb the bigger vibrations, and combined with the OIS might make for a more fluid, floating video.

Duane Steiner · June 28th, 2007, 01:01 PM

Thanks for all the ideas. That bike set up looks crazy, but might be a good starting point to make something to isolate the vibrations. Here are pictures of the two mounts: http://www.socalspeedzone.com/photos...nts/index.html

The thing that is frustrating is that both my SD cameras work fine on the mounts (Panasonic GS250 and an older JVC).

I will try a few more camera settings to see if there is a solution, but from what I have read on another forum, it just seems like the HV20 OIS does not work well in high vibration settings.

Mike Dulay · June 28th, 2007, 02:09 PM

Duane, have you tried putting rubber washers on the headrest clamps? That metal bar might work better if you had some a folded towel to keep the vibrations from the seat off the mount.

As for the windshield mount, someone else in the forum pointed the windshield tends to vibrate on its own. I would have thought the suction cups would shield it, maybe the ones touching the dash are transferring motion also?

Duane Steiner · June 28th, 2007, 04:16 PM

Mike-I have just a thin washer for the mount. Going to try and find a thick one or some other type of rubber/foam that will isolate it. As for the suction mount, when I used it only suspended from the window, it bounced around to much. The ones on the dash stopped the bounce, but added more vibration. I am determined to find a solution!

June 24th, 2007, 10:58 PM	#1
Mike Dulay Major Player Join Date: Mar 2007 Location: Somerville, NJ Posts: 304	Shooting observations and HV20 techniques Just some tips I've (re)learned ... hoping this will inspire you to share your own. Rolling Shutter/Jelly Effect After months of shooting exclusively 24p, I had to try 60i again. This is when I ran into the rolling shutter/jelly effect. Having tested the very same camera under the same conditions in 24p vs 60i I noticed a few things. I'm driving an SUV on a concrete top road (no asphalt). Dashboard is hard plastic. I set the camera on the dash and drive at 60mph. If you listen closely the gaps in the pavement hit the car 1-2times a second. Tip of the camera noticably bounces. 1) 60i-TV500-OIS on ... the video was jiggly while on the dashboard. BUT when I took the camera in my hand the jiggle went away. I guess my body isolated the camera from the car's vibration. 2) 24p-TV48-OIS on ... no jelly effect, just obvious bumps for sharp bumps on the road. Seems interlaced high shutter speed combined with frequent up/down motion is prone to rolling shutter. Isolate camera from vibration and it goes away. Moving Shots The camera is more sensitive for up/down tilt than side to side. While using a manfrotto 506B I accidentally nudged the camera back too far and it sprung back. This gave a shimmer like the OIS bounced up and down. Side to side jolts don't seem to bother it. With a DIY figrig, this bounce is reduced but it's still no "steadicam shot". My best moving shots were done on elevators/escalators/moving walkway. Spasmic Hand syndrome The longer I've been at this hobby the more camera shake I develop. Go figure. I'm able to salvage bad HV20 footage (especially 24p) by (1) downscaling to 1280x720 (2) slow motion. It seems that downscaling has a light 'deshaking' effect. Slowing down the motion to .6 (max) makes headache inducing shots more tolerable. __________________ DIY, 35mm, HV20: http://www.primitivebuteffective.net

June 27th, 2007, 07:18 PM	#7
Mike Dulay Major Player Join Date: Mar 2007 Location: Somerville, NJ Posts: 304	Duane, from the driving I did there wasn't a noticeable rolling shutter while I drove but the sharp bumps will cause a shudder. __________________ DIY, 35mm, HV20: http://www.primitivebuteffective.net

June 27th, 2007, 09:28 PM	#9
Andrew Plumb Regular Crew Join Date: Jan 2007 Location: Ottawa, Ontario, Canada Posts: 94	Duane, how are you mounting the camera in the car? Got some pictures of the setup? I ask because watching that video, the dash is bouncing around in the picture as much as everything else. My first guess (coming from a control systems engineering background) is that the OIS is compounding the problems - it's reacting to every jolt the whole car is subjected to. You want the optics to be locked w.r.t. the car's interior frame of reference. Try explicitly turning off OIS. If you can get the car interior to appear stationary w.r.t. the camera, then play with 60i versus 24p to see if rolling shutter is an issue. Only one way to find out! :-) Andrew. __________________ "The future is already here. It's just not very evenly distributed." - William Gibson

June 28th, 2007, 08:11 AM	#10
Mike Dulay Major Player Join Date: Mar 2007 Location: Somerville, NJ Posts: 304	Duane, I didn't notice any rolling shutter in your footage. Rolling shutter effect makes everything look like you're looking through a piece of jello ... straight objects leaning left and right in alternating pattern. What you have is camera vibration, the OIS isn't able to compensate for that kind of vibration. Like Andrew suggested, turn off the OIS. If possible, get the camera mounted somewhere it moves with the entire frame or isolate it from the vibration somehow. I haven't tried race style driving and your suspension is probably stiff so I can't suggest exactly where to mount it. When you rev do you shake with the car as well? Maybe there's some literature on how the people at Top Gear/Fifth gear mount their in car equipment? __________________ DIY, 35mm, HV20: http://www.primitivebuteffective.net

June 28th, 2007, 09:18 AM	#11
Andrew Plumb Regular Crew Join Date: Jan 2007 Location: Ottawa, Ontario, Canada Posts: 94	There may also be some secondary complications coming from auto-focus mechanisms as well. You'll probably have to turn off IAF because the camera's looking through the car's windshield. You may have to engage Manual Focus if the smaller vibrations are being interpreted as an out-of-focus picture. Probably set it for infinity (hold down the Focus button for 2 seconds) if the horizon is what you're interested in capturing, or just set it for some fixed point in front of the car. __________________ "The future is already here. It's just not very evenly distributed." - William Gibson

June 25th, 2007, 01:53 PM	#3
Elmer Lang Regular Crew Join Date: Dec 2002 Location: NY, NY Posts: 319	For me, shake is partly a factor of how it's held. I almost never hold it with my hand in the strap and if I do it's the opposite way, but usually I hold it thumb on top, hand on side and fingers underneath. One can loosen the strap and push the hand down through there (instead of up the usual way), but I rarely do. I have the DM-50 on top, but if I'm serious--well, more serious--I use the Zoom H4 to as a recorder with my better mics. What I don't like about that is that the audio is not on tape, just a little card and though you save to drive, what if the drive breaks. So I also save to cd, what if that breaks. Right now, tape just seems like the safest storage--whoa, maybe I should export audio to tape!

June 27th, 2007, 07:43 PM	#8
Duane Steiner Regular Crew Join Date: Feb 2006 Location: CA Posts: 195	I still don't understand the rolling shutter effect. This is the camera on the mount in my car (60i OIS on Auto): http://www.socalspeedzone.com/incartest/incartest.html

June 28th, 2007, 11:40 AM	#12
Andrew Plumb Regular Crew Join Date: Jan 2007 Location: Ottawa, Ontario, Canada Posts: 94	If you're of the DIY mind-set you might want try adapting this bicycle mounted steadicam design for in-car use. It would absorb the bigger vibrations, and combined with the OIS might make for a more fluid, floating video. __________________ "The future is already here. It's just not very evenly distributed." - William Gibson

June 28th, 2007, 01:01 PM	#13
Duane Steiner Regular Crew Join Date: Feb 2006 Location: CA Posts: 195	Thanks for all the ideas. That bike set up looks crazy, but might be a good starting point to make something to isolate the vibrations. Here are pictures of the two mounts: http://www.socalspeedzone.com/photos...nts/index.html The thing that is frustrating is that both my SD cameras work fine on the mounts (Panasonic GS250 and an older JVC). I will try a few more camera settings to see if there is a solution, but from what I have read on another forum, it just seems like the HV20 OIS does not work well in high vibration settings.

June 28th, 2007, 02:09 PM	#14
Mike Dulay Major Player Join Date: Mar 2007 Location: Somerville, NJ Posts: 304	Duane, have you tried putting rubber washers on the headrest clamps? That metal bar might work better if you had some a folded towel to keep the vibrations from the seat off the mount. As for the windshield mount, someone else in the forum pointed the windshield tends to vibrate on its own. I would have thought the suction cups would shield it, maybe the ones touching the dash are transferring motion also? __________________ DIY, 35mm, HV20: http://www.primitivebuteffective.net

June 28th, 2007, 04:16 PM	#15
Duane Steiner Regular Crew Join Date: Feb 2006 Location: CA Posts: 195	Mike-I have just a thin washer for the mount. Going to try and find a thick one or some other type of rubber/foam that will isolate it. As for the suction mount, when I used it only suspended from the window, it bounced around to much. The ones on the dash stopped the bounce, but added more vibration. I am determined to find a solution!