Motion trails on LCD monitor at DVinfo.net

Chad Terpstra · June 2nd, 2006, 02:46 PM

I have been using my new HD-100U for nearly a month now and while I’ve been delighted with it, I am having a lot of problems with motion trails (echoes on moving objects). It seems to happen on soft or out-of-focus moving objects. I found this thread: http://www.dvinfo.net/conf/showthrea...ghlight=trails and have decided I should probably send my camera in to be looked at.

I’ve noticed that they happen the most outrageously in 60pSDHDV mode, but also in 24pHDV. I haven’t noticed it in SD60i or 24p yet and have only wandered into HDV30p on a few occasions. It doesn't happen all the time, but when it does, it's very annoying. It also manifests as strange banding.

The problem is harder to see on just still frames, so here are some short clips. Note that motion smoothing is OFF.

Quick test around my office - HDV24p
http://video.terpstar.com/JVCTests/24pHDV-H264.mov

Clip from a recent short film test. – HDV24p
http://video.terpstar.com/JVCTests/2...Scene-H264.mov

Two clips from a Slow-mo test I did w/ my DP. Had to stay indoors because it was raining… These were shot SD60pHDV with a shutter of 1/120th of a second. Iris wide open. Conformed to 24p and put into DVCPro NTSC format.

http://video.terpstar.com/JVCTests/6...V_120_DV-1.mov
http://video.terpstar.com/JVCTests/6...V_120_DV-2.mov

It's a great camera other than this particular issue. I’d be really interested to get some opinions of these clips.

Barry Green · June 2nd, 2006, 03:46 PM

Are you using a mini35 or other type of lens adapter?

When we shot with the mini35 we got this type of ghosting through much of the night footage; I thought it was probably attributable to the mini35/ground glass overloading the HDV codec. Your clips are showing exactly the same effect (especially that third clip, DV-1).

Here's the kind of thing we got:
http://www.icexpo.com/HD100/HD100-weirdness.wmv

Chad Terpstra · June 2nd, 2006, 10:02 PM

Wow.... And I thought my stuff was bad. No, I don't have a GG at all. But it does seem to happen most with the iris wide open (which makes things softer). What's wrong with this codec??? It's so frustrating. I know that was a pre-release model that you guys used. Can anyone confirm that subsequent models have improved?

The 4:2:2 DVCProHD codec of the HVX is starting to look pretty appealing about now in spite of the lower resolution. I'm hoping when I send it in, they'll tell me it is a defective unit and everything will be good... If not, that's pretty disappointing.

Barry Green · June 3rd, 2006, 01:04 AM

Well, that was the question. I never made a big deal of it specifically because a) it was a prerelease model, and b) I figured it was compounded by the moving ground glass, so if you're not using a mini35 I figured it wouldn't be much of an issue (borne out by the fact that I haven't seen many examples of it).

Yours is the third example I've seen of this, but the second example (one of Nate's "downtown" .m2t's) was really quite mild. Yours is the only one I've seen that was as significant as the stuff we got with the mini35.

I am surprised you didn't have a moving ground glass system on your clips, I figured that was what was behind this phenomenon.

MPEG-2 does employ motion prediction, so I figure it may be due to that? Something getting "left behind", and not clearing up until the next I-frame? I can't think of anything else that would cause something to perpetuate from frame to frame like this.

Scott Jaco · June 3rd, 2006, 01:48 AM

Quote:

Originally Posted by Chad Terpstra

I am having a lot of problems with motion trails (echoes on moving objects).

Is the "motion smoothing" function turned off?

Barry Green · June 3rd, 2006, 01:58 AM

It was definitely turned off on the footage we shot with the mini35.

Chad, how about your footage? Motion smoothing on or off?

Chad Terpstra · June 3rd, 2006, 09:55 AM

Quote:

Originally Posted by Chad Terpstra

The problem is harder to see on just still frames, so here are some short clips. Note that motion smoothing is OFF.

It's not available for HDV60p anyway... Could maybe be something wrong w/ my lens that it reacts the same way a GG would/ Perhaps a film or something? I haven't noticed anything like that. Though sometimes the noise pattern doesn't seem to refresh the way it should. (it sometimes appears as if a static GG were in front of the lens or something. I've seen it on Nate Weaver's clips as well). But that has to do more with the fact that it's a noisey camera overall. I think.

What else could it be? Do you agree it's worth sending in for service?

Barry Green · June 3rd, 2006, 01:58 PM

Well, the fact we've seen something similar on three different cameras implies that it's not necessarily a service item, more likely it's something that should be verified by others first.

Can you create a step-by-step scenario that makes it happen every time? If so, post the procedure here and let's get a few others to try it and see if it happens on theirs. If it does, there's nothing service could do for you, but if nobody else gets it on theirs, then definitely send yours in for service.

Chad Terpstra · June 4th, 2006, 01:33 PM

No problem, Barry. Here's about as simple of a test that I can think of. Point the camera at a white wall, open up the iris all the way and adjust to expose the wall as roughly light gray. For my room I just opened it up and that was about as bright as it would get at 1/60 or 1/48 shutter.

Bring the wall into focus, and wave your hand around in front of the camera like a madman. If you're using my camera you'll see lots of motion trails or echoes from previous frames.

Here are my two tests in 60pHDV and 24pHDV. I didn't get a chance to try this at different F-stops, but I'd love to hear that other people don't get this problem. Looking forward to the results!!

http://video.terpstar.com/morevids/TRAILTEST24p.m2t
http://video.terpstar.com/morevids/TRAILTEST60p.m2t

Chad Terpstra · June 4th, 2006, 09:21 PM

Just wanted to say that I tried the test again at F2.8 and F5.6 and both displayed the same problem. I'm most concerned by the HDV60p as it is very distracting when slowed down to 24p. Any kind of imperfection like this reminds the viewer that they're watching a movie and not "witnessing a reality" so to speak. It's a distraction and it's not too pretty either.

I’ve decided to send the camera in based on this last capture. For the trails but also because of the bad case of SSE. I have seen SSE before at 0db, but not so bad that it ruined any shots. It also seems to come and go based on how long the camera’s been on. This was a cold camera test.

http://video.terpstar.com/morevids/T...p-5.6Fstop.mov
Codec: Apple DV/DVCPro NTSC
Size: 10MB

One other question: How long does recalibration usually take? We’re planning on shooting our short in two weeks and absolutely must have it back before then. Would I be better off waiting until after that project and dealing with these issues or sending tomorrow overnight and hoping it would come back in time? Thanks in advance.

Also wondering what folks such as Tim D, Nate Weaver, and Paolo Ciccone among others have to say about the trails issue.

Chad Terpstra · June 6th, 2006, 06:04 AM

I really need some help with this issue. This is a potential defect w/ the camera that we should try to understand. Anyone????

Daniel Patton · June 6th, 2006, 10:47 PM

Quote:

Originally Posted by Barry Green

MPEG-2 does employ motion prediction, so I figure it may be due to that? Something getting "left behind", and not clearing up until the next I-frame? I can't think of anything else that would cause something to perpetuate from frame to frame like this.

Just reading this and was wondering... to eliminate the possibility of it being introduced by the MPEG-2 compression, could you not go out component direct to disk? I'm not saying it's a simple test for just anyone (as I don't have a Wafian recorder) but doing this could locate or eliminate the possibility of it being the codec.

Quote:

Originally Posted by Chad Terpstra

The 4:2:2 DVCProHD codec of the HVX is starting to look pretty appealing about now in spite of the lower resolution.

And Chad, if you think that this issue is "disappointing" you have a long road ahead of you by switching to the HVX. After using both cameras and personally editing the footage, I would recommend that you spend more time with both cameras before deciding if the lower resolution of the HVX is something you can live with. My immediate impression is that you would be far more unhappy with the HVX after pressing your nose to an LCD with the HD100. ;)

Chad Terpstra · June 7th, 2006, 12:10 PM

Truly I don't think it's likely that I'll switch. I love the form of the JVC too much to turn back to the "handheld" cameras. Just working with your hands on the lens is the tipping point... And then of course there's the resolution. I was merely saying that the codec of the HVX is far superior based on my observations with this problem.

I don't feel I'm being too picky with this problem. It's actually more visible during playback than when paused. If it were only visible on still frames, I would be fine with it.

I tried thoroughly cleaning the front and back of the lens and it seems that the problem is somewhat lessened by this, but not fixed.

So what about the test that Barry proposed? Anyone try focusing on a wall and moving your hand or face near the lens? The more blurry, the more echoes and banding you'll see (if using my camera anyway). On shots with sharp details throughout the frame, the camera performs very nicely. It’s only with motion blurs or fast moving out-of-focus elements that it has problems.

Tim Dashwood · June 29th, 2006, 06:08 PM

Hi Chad,

I'm just catching up to this thread. It looks like plain old Mpeg compression artifacting to me... one of the things we always try to avoid with the HDV format.

Now here's the weird thing. I can see the "echo trail" on my LCD computer monitor when viewing the file frame by frame, but when I watch the file via the SRDVD-100U upconverted to 1080i and connected to a 17" HD monitor I just don't see the trails. I've watched it 10 times now (the waving hand 720P24) and I just can't see it.

Chad Terpstra · June 29th, 2006, 06:32 PM

That sounds like great new to me. Now that you mention it, perhaps there is something to be said for the monitor/TV one watches the footage on. I always thought it was curious why the echoes weren't visible when down-converted to SD and played on a CRT TV.

After reading your post I tried something different. I have two LCD screens: A Dell 24" (which I use as the primary screen) and my powerbook's 15." So instead of watching a problematic clip on my Dell monitor, I tried on the smaller Apple screen at full screen and voilá! No trails visible during playback! How strange is that? When I drag the clip over to the Dell, they're back and clear as day. It's almost like the Dell were lingering on those frames and leaving them up somehow in a way that highlights them... Or the refresh rate is not great or something. Either way it's a night and day difference.

I probably should have noticed this before, but my Apple screen is fading and I don't play footage on it very much. But it makes a lot more sense to me now why not many others were being disturbed by these horrendous-looking trails in their footage. -They're only visible on certain displays.

Thanks for looking into this for me, Tim. Like I said before I can live with these imperfections as long as they don't get in the way of normal playback and this more or less proves they don't (assuming that your display has a good refresh rate).

June 2nd, 2006, 02:46 PM	#1
Chad Terpstra Major Player Join Date: Jun 2005 Location: Grand Rapids, MI Posts: 320	Motion trails on LCD monitor I have been using my new HD-100U for nearly a month now and while I’ve been delighted with it, I am having a lot of problems with motion trails (echoes on moving objects). It seems to happen on soft or out-of-focus moving objects. I found this thread: http://www.dvinfo.net/conf/showthrea...ghlight=trails and have decided I should probably send my camera in to be looked at. I’ve noticed that they happen the most outrageously in 60pSDHDV mode, but also in 24pHDV. I haven’t noticed it in SD60i or 24p yet and have only wandered into HDV30p on a few occasions. It doesn't happen all the time, but when it does, it's very annoying. It also manifests as strange banding. The problem is harder to see on just still frames, so here are some short clips. Note that motion smoothing is OFF. Quick test around my office - HDV24p http://video.terpstar.com/JVCTests/24pHDV-H264.mov Clip from a recent short film test. – HDV24p http://video.terpstar.com/JVCTests/2...Scene-H264.mov Two clips from a Slow-mo test I did w/ my DP. Had to stay indoors because it was raining… These were shot SD60pHDV with a shutter of 1/120th of a second. Iris wide open. Conformed to 24p and put into DVCPro NTSC format. http://video.terpstar.com/JVCTests/6...V_120_DV-1.mov http://video.terpstar.com/JVCTests/6...V_120_DV-2.mov It's a great camera other than this particular issue. I’d be really interested to get some opinions of these clips.

June 29th, 2006, 06:08 PM	#14
Tim Dashwood Wrangler Join Date: Aug 2005 Location: Toronto, ON, Canada Posts: 3,637	Hi Chad, I'm just catching up to this thread. It looks like plain old Mpeg compression artifacting to me... one of the things we always try to avoid with the HDV format. Now here's the weird thing. I can see the "echo trail" on my LCD computer monitor when viewing the file frame by frame, but when I watch the file via the SRDVD-100U upconverted to 1080i and connected to a 17" HD monitor I just don't see the trails. I've watched it 10 times now (the waving hand 720P24) and I just can't see it. __________________ Tim Dashwood

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search

June 2nd, 2006, 03:46 PM	#2
Barry Green Barry Wan Kenobi Join Date: Jul 2003 Location: North Carolina Posts: 3,863	Are you using a mini35 or other type of lens adapter? When we shot with the mini35 we got this type of ghosting through much of the night footage; I thought it was probably attributable to the mini35/ground glass overloading the HDV codec. Your clips are showing exactly the same effect (especially that third clip, DV-1). Here's the kind of thing we got: http://www.icexpo.com/HD100/HD100-weirdness.wmv

June 2nd, 2006, 10:02 PM	#3
Chad Terpstra Major Player Join Date: Jun 2005 Location: Grand Rapids, MI Posts: 320	Wow.... And I thought my stuff was bad. No, I don't have a GG at all. But it does seem to happen most with the iris wide open (which makes things softer). What's wrong with this codec??? It's so frustrating. I know that was a pre-release model that you guys used. Can anyone confirm that subsequent models have improved? The 4:2:2 DVCProHD codec of the HVX is starting to look pretty appealing about now in spite of the lower resolution. I'm hoping when I send it in, they'll tell me it is a defective unit and everything will be good... If not, that's pretty disappointing.

June 3rd, 2006, 01:04 AM	#4
Barry Green Barry Wan Kenobi Join Date: Jul 2003 Location: North Carolina Posts: 3,863	Well, that was the question. I never made a big deal of it specifically because a) it was a prerelease model, and b) I figured it was compounded by the moving ground glass, so if you're not using a mini35 I figured it wouldn't be much of an issue (borne out by the fact that I haven't seen many examples of it). Yours is the third example I've seen of this, but the second example (one of Nate's "downtown" .m2t's) was really quite mild. Yours is the only one I've seen that was as significant as the stuff we got with the mini35. I am surprised you didn't have a moving ground glass system on your clips, I figured that was what was behind this phenomenon. MPEG-2 does employ motion prediction, so I figure it may be due to that? Something getting "left behind", and not clearing up until the next I-frame? I can't think of anything else that would cause something to perpetuate from frame to frame like this.

June 3rd, 2006, 01:58 AM	#6
Barry Green Barry Wan Kenobi Join Date: Jul 2003 Location: North Carolina Posts: 3,863	It was definitely turned off on the footage we shot with the mini35. Chad, how about your footage? Motion smoothing on or off?

June 3rd, 2006, 01:58 PM	#8
Barry Green Barry Wan Kenobi Join Date: Jul 2003 Location: North Carolina Posts: 3,863	Well, the fact we've seen something similar on three different cameras implies that it's not necessarily a service item, more likely it's something that should be verified by others first. Can you create a step-by-step scenario that makes it happen every time? If so, post the procedure here and let's get a few others to try it and see if it happens on theirs. If it does, there's nothing service could do for you, but if nobody else gets it on theirs, then definitely send yours in for service.

June 4th, 2006, 01:33 PM	#9
Chad Terpstra Major Player Join Date: Jun 2005 Location: Grand Rapids, MI Posts: 320	No problem, Barry. Here's about as simple of a test that I can think of. Point the camera at a white wall, open up the iris all the way and adjust to expose the wall as roughly light gray. For my room I just opened it up and that was about as bright as it would get at 1/60 or 1/48 shutter. Bring the wall into focus, and wave your hand around in front of the camera like a madman. If you're using my camera you'll see lots of motion trails or echoes from previous frames. Here are my two tests in 60pHDV and 24pHDV. I didn't get a chance to try this at different F-stops, but I'd love to hear that other people don't get this problem. Looking forward to the results!! http://video.terpstar.com/morevids/TRAILTEST24p.m2t http://video.terpstar.com/morevids/TRAILTEST60p.m2t

June 4th, 2006, 09:21 PM	#10
Chad Terpstra Major Player Join Date: Jun 2005 Location: Grand Rapids, MI Posts: 320	Just wanted to say that I tried the test again at F2.8 and F5.6 and both displayed the same problem. I'm most concerned by the HDV60p as it is very distracting when slowed down to 24p. Any kind of imperfection like this reminds the viewer that they're watching a movie and not "witnessing a reality" so to speak. It's a distraction and it's not too pretty either. I’ve decided to send the camera in based on this last capture. For the trails but also because of the bad case of SSE. I have seen SSE before at 0db, but not so bad that it ruined any shots. It also seems to come and go based on how long the camera’s been on. This was a cold camera test. http://video.terpstar.com/morevids/T...p-5.6Fstop.mov Codec: Apple DV/DVCPro NTSC Size: 10MB One other question: How long does recalibration usually take? We’re planning on shooting our short in two weeks and absolutely must have it back before then. Would I be better off waiting until after that project and dealing with these issues or sending tomorrow overnight and hoping it would come back in time? Thanks in advance. Also wondering what folks such as Tim D, Nate Weaver, and Paolo Ciccone among others have to say about the trails issue.

June 6th, 2006, 06:04 AM	#11
Chad Terpstra Major Player Join Date: Jun 2005 Location: Grand Rapids, MI Posts: 320	I really need some help with this issue. This is a potential defect w/ the camera that we should try to understand. Anyone????

June 7th, 2006, 12:10 PM	#13
Chad Terpstra Major Player Join Date: Jun 2005 Location: Grand Rapids, MI Posts: 320	Truly I don't think it's likely that I'll switch. I love the form of the JVC too much to turn back to the "handheld" cameras. Just working with your hands on the lens is the tipping point... And then of course there's the resolution. I was merely saying that the codec of the HVX is far superior based on my observations with this problem. I don't feel I'm being too picky with this problem. It's actually more visible during playback than when paused. If it were only visible on still frames, I would be fine with it. I tried thoroughly cleaning the front and back of the lens and it seems that the problem is somewhat lessened by this, but not fixed. So what about the test that Barry proposed? Anyone try focusing on a wall and moving your hand or face near the lens? The more blurry, the more echoes and banding you'll see (if using my camera anyway). On shots with sharp details throughout the frame, the camera performs very nicely. It’s only with motion blurs or fast moving out-of-focus elements that it has problems.

June 29th, 2006, 06:32 PM	#15
Chad Terpstra Major Player Join Date: Jun 2005 Location: Grand Rapids, MI Posts: 320	That sounds like great new to me. Now that you mention it, perhaps there is something to be said for the monitor/TV one watches the footage on. I always thought it was curious why the echoes weren't visible when down-converted to SD and played on a CRT TV. After reading your post I tried something different. I have two LCD screens: A Dell 24" (which I use as the primary screen) and my powerbook's 15." So instead of watching a problematic clip on my Dell monitor, I tried on the smaller Apple screen at full screen and voilá! No trails visible during playback! How strange is that? When I drag the clip over to the Dell, they're back and clear as day. It's almost like the Dell were lingering on those frames and leaving them up somehow in a way that highlights them... Or the refresh rate is not great or something. Either way it's a night and day difference. I probably should have noticed this before, but my Apple screen is fading and I don't play footage on it very much. But it makes a lot more sense to me now why not many others were being disturbed by these horrendous-looking trails in their footage. -They're only visible on certain displays. Thanks for looking into this for me, Tim. Like I said before I can live with these imperfections as long as they don't get in the way of normal playback and this more or less proves they don't (assuming that your display has a good refresh rate).