Vibrating image on TV output

[Patrick]

I'm still a relative novice using this equipment, but I'll never get beyond this point if I can't get it to work properly. I use a Mill2/RR-S, and I have a heck of a time trying to capture any number of clips where several of them don't display a "vibration" while being viewed on a TV monitor. There is no vibration to the image on the computer monitor. This vibration can stop and start any number of times in a one minute clip. My source footage is time base corrected Hi8. I capture using MSP at 704 x 480(NTSC), frame rate locked to 29.97, with MJPEG set at the highest quality. I never have any problem with dropped frames showing up during capture. However, I notice that the running time of the clip being displayed during capture is always slower than real time. (I think it was Markus of AVI_IO who has stated that this suggests the system can't handle the data flow.) Even when I capture at the second highest quality setting, I still have the same symptoms and problems. The type of footage I'm dealing with right now is just talking heads. I wouldn't think that there should be any problem with the RR handling this.

There are a zillion different settings and hardware issues that could be brought into this discussion, but before I go through the hassle of listing everything, I wanted to see how common a problem this was for others who use the RR or Marvel.


JTurner, are you still out there somewhere? I know you've discussed this in the past.


[This message has been edited by Patrick (edited 25 November 1999).]

[Ray]

Hi Patrick,
Are you sure this is not an external interference problem. Electric motors, radio gear etc or earth loop hum.

Ray

[Patrick]

No Ray, the problem is not something that happens all the time. The image vibration will start and stop in the same locations every time in the affected clip. I believe it may have something to do with dropped frames and the order of the A/B frames being screwed up. I know I stated that no frames are reported being dropped during capture, but I'm positive they're being dropped during playback. Figuring out how to stop this is the big challenge.

[Grigory]

Hi Patric,

I cannot understand what is vibration in your case, without looking at image. However, when you said that this effect is visible only on TV screen, this makes me thinking that your problem is with field order. Please select one of the cases below:

1. You capture video and then play it without editing. Vibration starts in some points.
2. You play edited video. Some EDITED fragments show vibration, when they contain any moving object. Another fragments are OK.


One common observation: vibration is visible only on scenes with motion. Still scenes do not vibrate. You may also do not see any vibration for very slow motion in a scene. This is an indication of incorrect field order.

If you have different problem, for example a brief stop and then vibrating/flickering picture for a part of a second, this is dropped frame, but you should see it on PC monitor too.

I will continue asking after you make your choice.

Grigory

[Patrick]

Hi Grigory,

This is a problem that can occur with a clip that has been captured and then played back in the MSP capture program, or the MSP editing program, or Windows Media Player. It doesn't appear to have anything to do with movement in the captured clip. The example I was referring to earlier in this thread was just a talking head, and it was shot with my camcorder on a tripod. In an affected clip, the vibration will start at the same point every time, and then disappear later at the same point in the clip. This vibration is very subtle. It basically makes the picture look less sharp or unfocussed.

Here's a situation that may be informative - Let's say that I have an affected one minute clip that starts to vibrate at the 10 second mark and then stops vibrating at the 50 second mark. If I was to put this clip into the MSP editor, and then trim this clip at the 30 second mark, all (previously unaffected) clips on the time line after this clip would have the same vibration. If I was to play back this time line a second time, the vibration would start at the beginning of the first clip and then stop at the 10 second mark.

It does appear that the field order is being reversed at these points. What I don't understand though, is if a frame is being dropped, is that not two fields? If two fields are being dropped, why would this reverse the order?

Here's an earlier thread that discusses this same problem:

http://forums.murc.ws/ubb/Forum2/HTML/001240.html

The fellas in this thread seem to believe that this is a heat related issue and that it's because the capture card is overheating. Well, I don't have a fan mounted right beside and aimed at my RR-S, but I do have two fans blowing in, and two fans blowing out. I also have the case cover off now, but to no avail.

In the project that I'm desperately trying to get done before Friday night, I finally had to recapture all the clips at 352 x 480(NTSC) using the highest quality setting. This seems to have eliminated the problem for now, but I'm not finished yet. Besides, I'm really disappointed that I don't appear to be able to capture at 704 x 480 with this series of shots.

It's difficult to convince a roomful of fellow video club members how good non-linear editing is when I have to turn up at a meeting with all my hair pulled out!

Ideas, anyone?

[Patrick]

A couple more points...

Just in case it was a heat related issue, I installed yet another fan. This one is aimed right at the Mill2/RR-S from three inches away. It has not made any difference.

When I recaptured a couple of troublesome clips at 704 x 480, the vibration reared it's ugly head again. It's interesting that the vibration starts and ends at different points in these recaptured clips than it did originally. It is also interesting (and annoying and frustrating, etc.) that there is NO movement of ANY type occurring at these "trigger" points. There is also no high contrast or large areas of white or anything that I've heard about that can cause dropped frames (if that is indeed the problem).

Ok, I'm begging now...

[SwAmPlAdY]

Howdy Pat. If I remember correctly, 352x480 is still capturing both fields.

I also doubt that it's a field order problem cuz you're gettin them in different places.

In any case, what happens if you capture at 704x480 but bring up the compression, or, leave it at 6:1 but capture with no audio?

Don't spose you have a pci sound card do ya?

SwAmPy

[Grigory]

Hi Patric,

I am afraid I cannot help in your situation.
One note: if you see correct movie on PC screen, then the problem is in field order.
how the card can reverse field order suddenly - I don't know.
Maybe, RRS can drop a single field and attach next captured field regardless of its odd or even number.

This is possible in principle, but only if field sequence control inside the card is bad or unreliable.
RR_G has no such problems, as far as I know. I have no idea how RRS works actually.

Grigory

[SwAmPlAdY]

Hmmm, since the vibrating is NOT happening in the same spot, it means that the field order is ok when captured but has a playback problem.

Maybe heat or bus mastering is causing something during the playback for RR to screw up the field order.

SwAmPy

[mmilburn]

Patrick,

I have had the `jitters` (no i`m not an alcoholic, but that would explain some of my probelms !!.

My system:- AMD K6 2/400 ASUS P5 SS7 mobo, 128 PC100 ram, 8.4gig Maxtor boot and Data. 2x 10gig Maxtors Promise Fastrak, Mill2 (PCI) + RRS with latest drivers. Win 98 MSP 5.2 Full. AVI_io.

That system above was built as my first PC which contained my RRS was not up to the job. SO I THOUGHT. However experince and thanks to quite a few contributers to this excellent forum i have solved my vibrating image problem.

I belive it had nothing to do with with my system but i do admit that the Fastrak has made a differnce. For some unknown reason when editing on the timeline using MSP the programs comes across a `ROGUE` frame. Eg all RRS are field A. So on the timeline you should have all AAAAAAAAAAAAAAAAAAAAA
but for some reason MSP throws a wobbler and you get AAAAAAAAAABBBBBBBBB
So once it comes across a frame with a B field it makes all the others B fields. The Solution: When playing back via preview stop the previewing where the vidoe starts to jitter. Then go to the Insert Title and as if you were going to add a title to the clip. Instead of tying text just hit return for a couple of lines. Then insert this on the OVERLAY track at the point where it jitters. Then play back the clip. For some reason by doing this it seems to give the `ROGUE` frame a kick up the arse and turn it back from field A to B. It DOES work. I still get this problem but use this solution with good results. You can even fine tune it so that you only have one frames of blank text on the overlay. I forget the name of the bloke who gave me the tip but i owe him a drink at least. If not my PC would have been used as a door stop because i was getting really pissed off with it. The RRS is not perfect but it does the job. I must admit i`m no expert.

Cheers

Martin

[jeff b]

Martin, you've hit the nail on the head!

A while back, we had some threads that refered to this problem, calling the rogue frames EVIL FRAMES. The EVIL FRAMES syndrome is something that RRS users stand a pretty good chance of turning up, every once in a while.

Patrick: the overlay of a blank title is one good way to re-render the EVIL FRAME point, and make it play back okay. Another solution is to find the frame, itself. If you stretch out the timeline, and set up VIEW to show each individual frame, you'll find it to be a horizontally striped EVIL FRAME. This one frame can simply be cut.

The reason why this happens, of course, can be attributed to a couple of different things. One reason can be that the tape has dropout due to wear, or just a defect on the tape. With TBC on something like that, the loss of one field in a frame will not always show up on the Matrox MJPEG codec capture as a dropped frame, for whatever reason.

Another possible cause for repeated EVIL FRAMES in a capture can be a combination of either tape defect and/or a large area of the frame having a solid color that the RR might choke and puke on. You said that the tape in question was just talking heads, so I'm guessing that the background may have had an expanse of solid color behind them. I've had bottomed out black, and blown out whites make my RR choke and either drop frames of develop EVIL FRAMES during capture many times with Hi8. I've also had it choke on my green screen, too, if there's very little else in the frame.

At any rate, the vibrating that you refer to is exactly what happens in MSP playback when it hits one of these EVIL FRAMES which, for whatever reason, ends up reversing the field order on playback. It'll just keep playing back in the reverse order to the end of the clip, or until it hits another EVIL FRAME and reverses the order again.

Pretty wierd, huh?

[CarterTG]

When I made the transition from a Mystique/RRS to a G200MarvelTV, my MSP results exhibited what could be described as "vibrations", but I called them jitters at the time. During the transition, I was mix-matching videoclips from both boards. When playing back these clips, parts of it seemed to play fine while crossfades would clearly show jittering on the NTSC screen (no jittering on the VGA).

JT in this forum suggested it was a Field Order issue & he was right. The RRS captures with Field Order A while the G200 likes Field Order B. This means when editing in MSP, right-click any/all clips and make sure its Field Options are set to Field A if the system uses a RRS. I'd also select Deinterlace. This should ensure the entirety of the resulting clip gets rendered in a Field Order friendly to the RRS.

There might also be a Field Option setting in MSP's Create Video dialog box... but I'm not at the editing machine right now to check.


------------------
Carter
------
Abit BH6 v1.02 ATX motherboard
Celeron 433mhz Slot1
128mb PC100 SDRAM
[C:] IBM 10.1gig 7200rpm UDMA33
[D:] IBM 10.1gig 7200rpm UDMA33
[E:] FastTrak66 RAIDed 2x13.5 7200rpm UDMA66
[F:] 24x CD-ROM
[G:] SanDisk USB CompactFlash reader
Matrox Marvel G200 TV AGP
Canopus DVRaptor PCI
SBLive Value PCI
DLink Ethernet PCI
v.90 modem PCI
Win98, MSP 5.2
Canon L2
Canon XL1

[Patrick]

Hey, all these responses are great.

Ordinarily, I would respond to everyone's post individually (and I will eventually), but I think my digestive tract has been invaded by "evil frames". It's all I can do to just type this out at the moment.

When I'm feeling better I'll give you all some feedback on your suggestions. Thanks very much everyone.

[Nick Piercey]

Hi Martin,


I wish I could claim that drink..... but the info I sent you was gleaned from earlier posts on MURC. So that's a lot of drinks you own a lot of people!?!?!?!

Can any one help me?

I'm about to upgrade from a Mystique220/RRS to a G400 Marvel... does the dreaded EVIL FRAME appear on occasions in the Marvel or was it unique to the RRS?

Thanks,


Nick

[mmilburn]

Nick,

If you ever come to the N.E. (Teesside) i`ll quite happily buy you a pint.

I`ve go no suggestions re. The EVIL FRAME with the Marvel. (don`t have one).

Cheers

Martin