When small interaxials just don't cut it! - Page 25 - AVS Forum
Forum Jump: 
Reply
 
Thread Tools
post #721 of 1097 Old 05-27-2012, 10:30 AM
AVS Special Member
 
Wolfgang S.'s Avatar
 
Join Date: Aug 2011
Location: Vienna/Austria
Posts: 1,148
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 12 Post(s)
Liked: 16
Well, I am not a nativ speaker - so sorry my wording is not as precise as it should be. But I am also a little bit disappointed about the wording you choose here.

For sure a camcorder cannot start with a half frame. BUT a second camcorder can start to readout a frame at the point in time, when the first camcorder is at the half of his frame. The result is asynchronity, that will harm the 3D effect.

And I wonder why the asynchronity with consumer camcorde and lanc are ignored here. Only because there is another effect - that can be compensated as we do that since years with every multicam-editing?.

And yes my friend, if you have measured the sync drift with my ste-fra Lanc V3, that is able to do so. That are the figures that I have mentioned about. But have you done so?

Kind regards,
Wolfgang
videotreffpunkt.com
Wolfgang S. is offline  
Sponsored Links
Advertisement
 
post #722 of 1097 Old 05-27-2012, 10:39 AM - Thread Starter
AVS Special Member
 
Frank's Avatar
 
Join Date: Apr 1999
Location: Dept. of Offense
Posts: 5,194
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 4 Post(s)
Liked: 14
Quote:
Originally Posted by Don Landis View Post

I thought we were discussing science. As friends, I trust you believe in what you see.

Meanwhile- I need to get back to pairing and keyframing my clips. But, please consider this- I don't have the problem. When I do, I'll let you know.

It's not just what I see Don.
My friend and I designed a dual camcorder LANC controller and did lots of testing with it including saving data and graphing it with spreadsheets.
Hopefully we'll get back to it at some point.

Did IQ's suddenly drop sharply while I was away?
I enjoy 3D in spite of HDMI 1.4!
Full screen only 3D doesn't cut it!
Frank is offline  
post #723 of 1097 Old 05-27-2012, 11:25 AM
AVS Special Member
 
Wolfgang S.'s Avatar
 
Join Date: Aug 2011
Location: Vienna/Austria
Posts: 1,148
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 12 Post(s)
Liked: 16
Frank, Don, all of us love 3D and shoot since years. The discussion is fun, but it is also learning. I think it would be a nice idea to go back and try to listen to each other. Maybe the discussion would benefit from that, given the different experience that we have.

Kind regards,
Wolfgang
videotreffpunkt.com
Wolfgang S. is offline  
post #724 of 1097 Old 05-27-2012, 11:49 AM - Thread Starter
AVS Special Member
 
Frank's Avatar
 
Join Date: Apr 1999
Location: Dept. of Offense
Posts: 5,194
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 4 Post(s)
Liked: 14
Quote:
Originally Posted by Wolfgang S. View Post

Frank, Don, all of us love 3D and shoot since years. The discussion is fun, but it is also learning. I think it would be a nice idea to go back and try to listen to each other. Maybe the discussion would benefit from that, given the different experience that we have.

Sounds good to me.
I'll start by sharing some of my experience with trying to design a dual camera LANC controller.
We built one using three AVR 8 bit processors.
Two of them were used just for critical timing for communication with each camera. The other was the supervisor.
One of the reasons I decided to design one was I thought I could better compensate for poor sync between cameras then the LANC shepherd could.
During the testing phase I was looking for repeatability in the power-up to see if it would be practical to get a longer run time by starting the camera with the faster clock later then the other.
What we found was there was very little repeatability and the drift between the cameras was considerable.
The cameras in question were Canon Vixia HF-S21s.

Did IQ's suddenly drop sharply while I was away?
I enjoy 3D in spite of HDMI 1.4!
Full screen only 3D doesn't cut it!
Frank is offline  
post #725 of 1097 Old 05-27-2012, 12:13 PM
AVS Special Member
 
Wolfgang S.'s Avatar
 
Join Date: Aug 2011
Location: Vienna/Austria
Posts: 1,148
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 12 Post(s)
Liked: 16
The link that I posted before - I received an error message since I wrotte them an email. And I got the information that Sony has changed the LANC-protokoll. To hold the cameras via lanc in syn was possible with some DV-camcorders, but is not possible any more for HD. At least that told me Werner from the stereoforum.de - and I assume him to be right, since he sells ste-fra lanc.

So maybe your solution was also for DV?

Kind regards,
Wolfgang
videotreffpunkt.com
Wolfgang S. is offline  
post #726 of 1097 Old 05-27-2012, 12:20 PM - Thread Starter
AVS Special Member
 
Frank's Avatar
 
Join Date: Apr 1999
Location: Dept. of Offense
Posts: 5,194
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 4 Post(s)
Liked: 14
Quote:
Originally Posted by Wolfgang S. View Post

The link that I posted before - I received an error message since I wrotte them an email. And I got the information that Sony has changed the LANC-protokoll. To hold the cameras via lanc in syn was possible with some DV-camcorders, but is not possible any more for HD. At least that told me Werner from the stereoforum.de - and I assume him to be right, since he sells ste-fra lanc.

So maybe your solution was also for DV?

That does not surprise me.
I suspected they had changed the protocol.
My solution is much less sophisticated and doesn't work anyway because of the lack of repeatability.

Did IQ's suddenly drop sharply while I was away?
I enjoy 3D in spite of HDMI 1.4!
Full screen only 3D doesn't cut it!
Frank is offline  
post #727 of 1097 Old 05-27-2012, 01:08 PM
Senior Member
 
MLXXX's Avatar
 
Join Date: Jan 2007
Location: Brisbane, Australia
Posts: 238
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 56 Post(s)
Liked: 45
Two different approaches it seems.

In one camp, Don appears to consider it sufficient to time align a pair of hyperstereo Left and Right clips to the nearest frame. This would mean that the Left and Right image content could differ in time by up to half a frame (worst case scenario). If Vegas can align to 1/60th second [as it does if you set a project frame rate of double NTSC, and "quantise frames"], then the Right clip might lag the left by up to 1/120sec [disregarding the NTSC 1000/1001 detuning factor], or lead by up to 1/120th sec. The accuracy of alignment of the content would be ± 8.3mS.

What if the two cameras had a relative drift of 1mS per 5 minutes? It would take about 8 minutes for them to drift by half a frame, at 60 frames per second.

Eight minutes is a long time. A clip that long could be split into two [or more] segments to allow dropping of a frame from the slower camera, at an inconspicuous part of the video.

In the other camp, Frank and Wolfgang appear to be aiming for timing discrepancies of much less than half a frame. If a dual LANC controller can start both cameras at once with a very small timing difference, and if each segment of a shoot is kept short (relative to camera drift), that goal might be realised.

_________

Just to report now on my own little exercise this weekend, using an i50 [aka i25] source for the Left and a p59.94 source for the Right! The shooting session lasted only 3 minutes.

LEFT: Canon HV-20 camcorder with output at 1440x1080i50 [i25] MPEG2. [Output de-interlaced to 1440x1080p50 uncompressed, and then frame rate converted (blend option) to 1920x1080p59.94, uncompressed, using VirtualDub.] This camera produces noticeable barrel distortion at its wide angle setting. [In addition a wide-angle lens was screwed on, though this was still not quite enough to match the wide-angle view of the Right camera.]

RIGHT: Sony TD10 recording at 1920x1080p59.94. [Zoom adjusted to just short of full wide angle.]

INTERAXIAL: about 230mm. Scene contained objects at distances varying from about 2 metres to 5 metres.

VEGAS PRO: The clips included some arm waving and clapping of hands and these movements were able to be time aligned to the nearest 1/60th second, without difficulty. The Left and Right clips were then paired as a stereoscopic subclip, and adjusted automatically with the stereoscopic adjustment video effect.

Burned to one BD-RE as a 1280x720p59.94 3D Blu-ray, and to another BD-RE as a 1920x1080p23.976 3D Blu-ray.

RESULT: The stereo effect was exaggerated as expected.

The 3D hung together timing-wise. The Blu-rays could be paused during playback at various frames without any obvious timing effects, apart from the fact that the 24p Blu-ray tended to show blurring for the arm waving (an artefact of the 60i => 24p conversion).

COMMENT: There appeared to be more blurring and loss of temporal resolution from using 24p rendering, than from using the less than ideal '50i' source for the Left view

There were considerable differences in contrast and colour, and optical geometry of the two cameras. The subjective quality of the 3D experience was quite a bit less than normal. However, the 3D effect certainly worked, and much better than anaglyph.
MLXXX is online now  
post #728 of 1097 Old 05-27-2012, 02:22 PM - Thread Starter
AVS Special Member
 
Frank's Avatar
 
Join Date: Apr 1999
Location: Dept. of Offense
Posts: 5,194
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 4 Post(s)
Liked: 14
Quote:
Originally Posted by MLXXX View Post


In the other camp, Frank and Wolfgang appear to be aiming for timing discrepancies of much less than half a frame. If a dual LANC controller can start both cameras at once with a very small timing difference, and if each segment of a shoot is kept short (relative to camera drift), that goal might be realised.

With me, it's not a goal it is a reality and has been for quite some time. ( now that I have a combination of cameras that work well together.)
It's a necessity since I need to review the left and right clips in 3D without editing them first.

Did IQ's suddenly drop sharply while I was away?
I enjoy 3D in spite of HDMI 1.4!
Full screen only 3D doesn't cut it!
Frank is offline  
post #729 of 1097 Old 05-27-2012, 02:35 PM
AVS Special Member
 
Wolfgang S.'s Avatar
 
Join Date: Aug 2011
Location: Vienna/Austria
Posts: 1,148
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 12 Post(s)
Liked: 16
You cannot be sure if you have 8 minutes or 10 minutes. It also depends on the missync at the start. Then the drift seems to happen at a stable average. That is what I see here. You can measure the sync drift - witht eh ste-fra Lanc you do that in a way, where you take a measurement point every 30 seconds. There is a variance, but after 8-10 minutes you reach about 1 ms. And 1 ms is seen as the maximum of missync one should accept - here that position was taken by Frank, in German it is the number that the producer of tht ste-fra lanc tells you too.

10 minutes or 8 minutes - well, sounds much. But that is not much. You have to adjust the rig to the position you wish to shoot. Adjust the cameras. Then half of this time is gone. After that time you have to reset the camcorder, means that you have to shut it down. And need a new start-point. With my actual pair it happens that the startpoint is above the 1 ms we go for. So you do it some times until you have a valid startpoint. That is borring.

With the other camera pair I had at least 30-60 minutes but the fault optics. So that is something that I look again for.

Kind regards,
Wolfgang
videotreffpunkt.com
Wolfgang S. is offline  
post #730 of 1097 Old 05-28-2012, 07:18 AM
Senior Member
 
MLXXX's Avatar
 
Join Date: Jan 2007
Location: Brisbane, Australia
Posts: 238
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 56 Post(s)
Liked: 45
With all this focus on extreme accuracy of capture synchronisation, I thought I'd describe the following less demanding type of hyper-stereo application. Please forgive me if this subject matter seems too obvious to mention!

There would be situations where an alignment to the nearest frame would probably suffice even for a "formal" hyper-stereo video, e.g.:-

1. The video contains only slow moving content, e.g. a very long stereo base with cameras directed to a distant scene, and;

2. Any pans or zooms are very slow.

I note that manual frame quantisation alignment in post to the nearest frame for a 60p capture really means alignment to 1/120th second (worst case discrepancy) for the underlying movement. For example:


Asynchronous frames at 59.94fps

The time interval between successive frames is 1/59.94 = 16.683mS

Trial alignment 1: Left motion, by inspection,* leads Right motion by about 10mS.

Solution: move Left clip one frame, or 16.683mS, to the right on the time line

Final alignment: Right motion leads Left motion by about 6.7mS

_______

* In practice it could be very difficult to estimate this without artifical help, such as deliberately introducing visible fast and erratic motion at the start or end of the clip. Or if the sound capturing latency is consistent, relying on precise alignment of individual cycles of high frequency sound waves on the project time line.
MLXXX is online now  
post #731 of 1097 Old 05-28-2012, 07:44 AM
AVS Special Member
 
Wolfgang S.'s Avatar
 
Join Date: Aug 2011
Location: Vienna/Austria
Posts: 1,148
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 12 Post(s)
Liked: 16
Sure, you can see it in that way. But do not oversee that the start-point of frames will shift too. Means, you can do such an adjustment at a point in time - but if the sync drift is high you may be out of sync again some timespan later.

But maybe that is enough for consumer purposes.

Kind regards,
Wolfgang
videotreffpunkt.com
Wolfgang S. is offline  
post #732 of 1097 Old 05-28-2012, 09:40 AM
AVS Club Gold
 
Don Landis's Avatar
 
Join Date: Jun 1999
Location: Jacksonville, FL
Posts: 11,285
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 354 Post(s)
Liked: 171
Quote:
Originally Posted by MLXXX View Post

With all this focus on extreme accuracy of capture synchronisation, I thought I'd describe the following less demanding type of hyper-stereo application. Please forgive me if this subject matter seems too obvious to mention!

There would be situations where an alignment to the nearest frame would probably suffice even for a "formal" hyper-stereo video, e.g.:-

1. The video contains only slow moving content, e.g. a very long stereo base with cameras directed to a distant scene, and;

2. Any pans or zooms are very slow.

I note that manual frame quantisation alignment in post to the nearest frame for a 60p capture really means alignment to 1/120th second (worst case discrepancy) for the underlying movement. For example:

Asynchronous frames at 59.94fps

The time interval between successive frames is 1/59.94 = 16.683mS

Trial alignment 1: Left motion, by inspection,* leads Right motion by about 10mS.

Solution: move Left clip one frame, or 16.683mS, to the right on the time line

Final alignment: Right motion leads Left motion by about 6.7mS

_______

* In practice it could be very difficult to estimate this without artifical help, such as deliberately introducing visible fast and erratic motion at the start or end of the clip. Or if the sound capturing latency is consistent, relying on precise alignment of individual cycles of high frequency sound waves on the project time line.

We're on the same page here. Frank has some unique requirements that require precise start points. Those similar requirements are necessary in live switch twin camera 3D production such as the ESPN 3D channel. They use twin SBS cams and live switch the pairs for live event broadcast. The rest of us are content to use a far less costly approach and edit the clips in post for Twin cam viewing. As far as vertical and horizontal sync frequency drift is concerned, I strongly feel this is a belief based on electronic ignorance. But, people will believe in myths all the time. The bottom line is that these sync oscillators do not drift enough to cause any concern. After all if they did none of the TV production would function the way it does. Start point match is a different story and this can be addressed with the right equipment which does exist. But, LANC does not support it and never did in video. Therefore the LANC device we have are at best good for convenience in control of the switches, not to GENLOCK the cameras.
Sony Vegas has one of the slickest tools to accomplish what we need to do wide stereo base editing accurately and consistently. It's the 3D stereoscopic adjust with auto correct and auto crop/zoom. I recognize that if your cameras are severely out of optical alignment, this can cut the HD resolution with the auto crop and zoom feature, but no benefit has ever been achieved without some compromise somewhere. IMO, the tool Sony has here is a god-send to make our task easy. But if one insists on 100% perfection then maybe doing video by today's standards is not worth the effort. What I see here is that in using the Sony Tool to create paired 3D clips that work, is people are not following the proper procedure, clearly described by Sony in their help files. It's not just with this task but I've seen it repeatedly for 3D in many other procedures.
I have to say that I am very pleased with how Sony Vegas has made what I want to do in building my stories for 3D. I also want to thank Frank for making a hard to get Lanc Shepherd device available to me as it has made my shooting much easier. And by using the right work flow in Vegas post, I have been able to make the zoom control work without jarring the vision by key framing the auto correct alignment in the zoom range. Some of these methods are not described in the help files in Vegas but I'd be happy to explain the procedure if anyone wants to learn it.


Frank- last night I had a thought about your sync requirements. First, please understand that I don't believe you will ever be able to sync two cameras close enough to switch both on record at the precise moment in time required using LANC technically it just won't support that ability. But, I know it can be done using GENLOCK capable cameras and a vertical interval switcher. Unfortunately, these cameras and the associated hardware are very expensive. There may be other ways to do this, however. Time base correctors are not indicated here because they are mostly designed to sync tape machines using things like advanced sync control. We need not get into those but what about using a frame store synchronizer? Here, you can connect two consumer camcorders and the frame store synchronizer will mix both cameras by aligning the vertical sync point of one camera with the other. It does this by delaying one of the video signals enough to align the two precisely at the same frame. Now the output of the frame store will contain both left and right pictures. I'm not sure how this will work with horizontal off set but there may be a way to adjust this too. I had one of these years ago but sold it. Besides, it was only SD and you would need one for HD. Often for consumer devices these may be called video mixers.
Don Landis is offline  
post #733 of 1097 Old 05-28-2012, 10:09 AM - Thread Starter
AVS Special Member
 
Frank's Avatar
 
Join Date: Apr 1999
Location: Dept. of Offense
Posts: 5,194
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 4 Post(s)
Liked: 14
Quote:
Originally Posted by Don Landis View Post

As far as vertical and horizontal sync frequency drift is concerned, I strongly feel this is a belief based on electronic ignorance. But, people will believe in myths all the time. The bottom line is that these sync oscillators do not drift enough to cause any concern. After all if they did none of the TV production would function the way it does.

I hope you're not suggesting that I believe this drift exists based on mythology.

The LANC sync pulse is the same as the vertical sync.
I have precisely measured it and the drift is not a myth. It's a scientifically measured fact.
Why do you assume that consumer camera sync is as accurate as professional versions?
I don't get it.

Did IQ's suddenly drop sharply while I was away?
I enjoy 3D in spite of HDMI 1.4!
Full screen only 3D doesn't cut it!
Frank is offline  
post #734 of 1097 Old 05-28-2012, 10:51 AM
AVS Club Gold
 
Don Landis's Avatar
 
Join Date: Jun 1999
Location: Jacksonville, FL
Posts: 11,285
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 354 Post(s)
Liked: 171
Quote:
Originally Posted by Frank View Post

I hope you're not suggesting that I believe this drift exists based on mythology.

The LANC sync pulse is the same as the vertical sync.
I have precisely measured it and the drift is not a myth. It's a scientifically measured fact.
Why do you assume that consumer camera sync is as accurate as professional versions?
I don't get it.

Frank- I'm not saying the consumer cameras clocks don't drift. I even stated that broadcast cameras drift and what conditions I have seen significant mismatch, heat. What I question is the evidence that the drift is significant to cause trouble for our use in pairing 3D. So far none, that's a big fat zero paired clips have enough drift that I can't pair them for good quality 3D that I have experienced. Plus the math on it doesn't calculate to indicate the process won't work. again, I am not judging your specific needs, I'm stating that for the rest of us who want to work in twin camera pairing that the procedure exists and one need not worry over any of this sync frequency drift issue. It's a totally non issue if you follow the work flow as outlined by Sony.

And besides, I don't think the Greeks know anything about this myth. Heck they can't even manage their money!
Don Landis is offline  
post #735 of 1097 Old 05-28-2012, 11:11 AM
AVS Special Member
 
Wolfgang S.'s Avatar
 
Join Date: Aug 2011
Location: Vienna/Austria
Posts: 1,148
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 12 Post(s)
Liked: 16
Don, I think we have written again and again that the workflow as outlined by Sony is fine for us. Nobody questions that. But we have also stated that this does not touch syn drift.

Ok, you say that you know that there is syncdrift - fine. But you say that you think that it is not import. Ok I accept that as your opinion, but it must be possible to say that I do not share that opinion.

I see your point that you have the experience that in your footage that never has been an issue. But I understood that you have not measured your sync drift (ok why should you do so if that is not important for you). MLXXX has a point - you shoot very static objects - in the national parks. Great, I love that. But I shoot dynamic objects, who move. So from that point sync is important to me.

I have learned long time ago that I trust my measurements. You can manage what you measure. And my measurement says that my pair reaches the 1ms missync after 10 minutes. A time-span that sounds not so long to me. That is the - only - reason why I will try to improve that.

Beside that I am happy with your rig design using the IGUS-slides. It is great and thank you for that.

Kind regards,
Wolfgang
videotreffpunkt.com
Wolfgang S. is offline  
post #736 of 1097 Old 05-28-2012, 11:39 AM - Thread Starter
AVS Special Member
 
Frank's Avatar
 
Join Date: Apr 1999
Location: Dept. of Offense
Posts: 5,194
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 4 Post(s)
Liked: 14
How would you guys like it if you could not review your TD-10 videos in 3D without editing them in Vegas first?

I suspect you might not like it.

That's how I feel about the 3D video I shoot the most which is with dual camcorders.

Did IQ's suddenly drop sharply while I was away?
I enjoy 3D in spite of HDMI 1.4!
Full screen only 3D doesn't cut it!
Frank is offline  
post #737 of 1097 Old 05-28-2012, 08:40 PM
AVS Club Gold
 
Don Landis's Avatar
 
Join Date: Jun 1999
Location: Jacksonville, FL
Posts: 11,285
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 354 Post(s)
Liked: 171
Now that I have plenty of hands on experience with the TD10, I mostly shoot outdoors with the 2D monitor as it is brighter. I stated before I trust the camera to handle the 3D nicely. I'm mainly concerned if the shot is framed and focused right than how the 3D looks. So far Sony has not disappointed me in taking care of the 3D automatically. But when I was very new with the TD10, I used the 3D monitor to learn how it reacted to different camera angles and foreground objects. Today, its my backup. Also, I prefer the two monitors so I can visually compare the color balance and exposure when shooting twin cameras. I feel I have more control this way.
Don Landis is offline  
post #738 of 1097 Old 05-28-2012, 08:59 PM
AVS Club Gold
 
Don Landis's Avatar
 
Join Date: Jun 1999
Location: Jacksonville, FL
Posts: 11,285
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 354 Post(s)
Liked: 171
Quote:


But I understood that you have not measured your sync drift (ok why should you do so if that is not important for you)

Wiolfgang- I said I have NO Drift in the frame rate in either camera for the clip length I typically shoot. The maximum is maybe 15 minutes. The frame rate is a match and that is all that matters if I line up the clips in post. They must be lined up manually because the start point is not the same from clip to clip. This is not my procedure. It is what is taught in the Sony classes at NAB on how to edit in Vegas and what they state in the help file. I follow the process and get results I'm looking for, whether its shooting fireworks, water and laser light shows, or static mountains and people walking through the canyons. I don't just shoot National Parks.

I'm through with this subject as I can offer no other comments and at this point I'm just repeating myself. Final word- stick to the method described by Sony, it works.

Besides, I have some other observations on SBS shooting limitations that I'd like to discuss.
Don Landis is offline  
post #739 of 1097 Old 05-29-2012, 01:33 AM
AVS Special Member
 
Wolfgang S.'s Avatar
 
Join Date: Aug 2011
Location: Vienna/Austria
Posts: 1,148
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 12 Post(s)
Liked: 16
Quote:
Originally Posted by Frank View Post

How would you guys like it if you could not review your TD-10 videos in 3D without editing them in Vegas first?

I suspect you might not like it.

That's how I feel about the 3D video I shoot the most which is with dual camcorders.

No, I would not like that. You have to be able to adjust the two events, here I agree with Don that the start point may differer for some frames.

To avoid that you need another behaviour from your camcorder, compared with what I see: you need a very synchrone start-up when you switch the units on, you need a high sychronized start of the capturing and no sync shift. Be happy if your camcorders behave like that, I am looking for that.


Quote:
Originally Posted by Don Landis View Post

Wiolfgang- I said I have NO Drift in the frame rate in either camera for the clip length I typically shoot. The maximum is maybe 15 minutes. The frame rate is a match and that is all that matters if I line up the clips in post.

What do you mean by drift in the framerate? My understanding is that there is not so much a drift in the framerate, but there is a drift when the camera beginns to readout the first line from the CMOS. I think that will differ.

To my opinion Vegas aligns the frames to the 25 or 30 or 50 or 60 fps, whatever you have adjusted in the project settings - so you will not see that in Vegas in the framerate at all. But what you can see is a difference in the content. To see that better I think about a small experiment: to shoot a large clock with a seconds hand with both cameras and compare that about a longer time span. Putting the footage from both cameras in Vegas and see when and how fast a shift takes place, in a side-by-side view.

10 or 15 minutes sound enough - but I think we have to set up the tripod and the rig, switch on the camcorder, adjust the camcorder - that takes some minute. And then the 10 minutes can be over. At least it means that I have to shut down the cameras after the first or second shoot. If I work like that then - yes - then I can stay with my system.

Sure, there are other topics we should discuss. For example, I had the funny situation that one of the camcorder did not see a window when I shoot my house - it was hidden behind a tree. The other camcorder saw that window. After aligning and pairing the clips it was a crazy irritation that took place when looking at the footage.

Kind regards,
Wolfgang
videotreffpunkt.com
Wolfgang S. is offline  
post #740 of 1097 Old 05-29-2012, 04:21 AM
Senior Member
 
MLXXX's Avatar
 
Join Date: Jan 2007
Location: Brisbane, Australia
Posts: 238
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 56 Post(s)
Liked: 45
Quote:
Originally Posted by Wolfgang S. View Post

To my opinion Vegas aligns the frames to the 25 or 30 or 50 or 60 fps, whatever you have adjusted in the project settings - so you will not see that in Vegas in the framerate at all. But what you can see is a difference in the content. To see that better I think about a small experiment: to shoot a large clock with a seconds hand with both cameras and compare that about a longer time span. Putting the footage from both cameras in Vegas and see when and how fast a shift takes place, in a side-by-side view.

Using a sweep seconds hand that moves smoothly, a discrepancy of 1mS will of course be one thousandth of the distance the seconds hand moves in one second, a very small distance indeed! Even a full frame discrepancy at 24fps, a discrepancy of 1/24th sec or 41.66mS, could be hard to see, i.e. 1/24th of a division of one minute on the clock face!

I think for this experiment to give quantifiable results, the two cameras might need to be on a telephoto setting so that only a very small portion of the clock face filled the camera viewfinder.
MLXXX is online now  
post #741 of 1097 Old 05-29-2012, 05:36 AM
AVS Special Member
 
Wolfgang S.'s Avatar
 
Join Date: Aug 2011
Location: Vienna/Austria
Posts: 1,148
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 12 Post(s)
Liked: 16
Well, you are right: it takes some time until we see such a significant shift that it would become visable on a clock. If I take the sync shift of 0.2 ms/minute that I have measured, to come up with a second timedifference would mean to wait 83 minutes, also depending on the start missync. So after 1 and a half hour there would be a chance to see something.

To compare that number of 83 minute: if I would take my sync shift with my other TD10 (0.016ms/min), I would have to wait 1041 minutes or 17,4 hours until I reach one second.

That is another number that shows how small the effects are, that we discuss here. I think about if it is worthwhile really the effort.

It would be really nice to know what missync in ms it is, that is recognized by our brains typically.

Kind regards,
Wolfgang
videotreffpunkt.com
Wolfgang S. is offline  
post #742 of 1097 Old 05-29-2012, 05:41 AM - Thread Starter
AVS Special Member
 
Frank's Avatar
 
Join Date: Apr 1999
Location: Dept. of Offense
Posts: 5,194
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 4 Post(s)
Liked: 14
In addition to my tests with paired TD-10s and I also tested my Sony HDR-CX560s in a dual camera setup.
Both types had two things in common.
1. They often started out of sync by several frames and drifted sync wise more then I found acceptable.
2. They both disabled the composite outputs when LANC is used.

The numerous Canon camcorders I've tested all start within on frame in my experience and the composite outputs are not disabled.

Both of these issues make the Sonys unusable for my purposes.

I used to let my dual camcorders record non stop for hours while recording wildlife. While the first 12 minutes of the recording were always in sync (started at the same time) all of the clips after that were out of sync and required time consuming pairing in Vegas to view.
I wound up creating macros to automatically stop the recordings before the 12 minute mark and restart them in sync again.

Did IQ's suddenly drop sharply while I was away?
I enjoy 3D in spite of HDMI 1.4!
Full screen only 3D doesn't cut it!
Frank is offline  
post #743 of 1097 Old 05-29-2012, 05:54 AM - Thread Starter
AVS Special Member
 
Frank's Avatar
 
Join Date: Apr 1999
Location: Dept. of Offense
Posts: 5,194
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 4 Post(s)
Liked: 14
Quote:
Originally Posted by Wolfgang S. View Post


It would be really nice to know what missync in ms it is, that is recognized by our brains typically.

The answer is it depends.
I shot a lot of video of small animals that move very fast at higher than normal shutter speeds and I could easily see an error of 2ms while watching at normal speed. I'm in my mid 60s and I would think much younger people would notice it even more so.

Did IQ's suddenly drop sharply while I was away?
I enjoy 3D in spite of HDMI 1.4!
Full screen only 3D doesn't cut it!
Frank is offline  
post #744 of 1097 Old 05-29-2012, 07:46 AM
AVS Club Gold
 
Don Landis's Avatar
 
Join Date: Jun 1999
Location: Jacksonville, FL
Posts: 11,285
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 354 Post(s)
Liked: 171
Quote:
What do you mean by drift in the framerate? My understanding is that there is not so much a drift in the framerate, but there is a drift when the camera beginns to readout the first line from the CMOS. I think that will differ.

OK, then the language barrier is why we may not have understood what your complaint is. I thought you meant the frame rate was drifting about 23.976 fps, like plus or minus 2 fps. Sorry I misunderstood your use of "drift".
Here it is quite likely that the start moment of the first trace in the first frame will start at different moments within the first recorded frame. However, I'm afraid you are now asking for something that will never happen with our consumer camcorders as consumer camcorders lack the circuits to achieve this match. The technology does exist, however. It's called "Genlock" I have mentioned this over and over before but maybe I was speaking to an audience that does not know what Genlock is or how it is used in multi camera production, and most important, what is required to use it. Let me explain for anyone who is not familiar with the concept of genlock.
Genlock is a process where a camera in a 2 or more multicamera shoot is designated the master clock. This camera has a sync output jack. It also has a time code output jack. (In TV studios, the master clock is usually a device and not a camera and all cameras will be locked to the device.) In broadcast cameras, there are four such jacks for TC and sync input and Time code ad sync output. Consumer cameras have none of these! The master camera is set for "free run" time code drop frame for those in the NTSC world. Drop frame is required so that the clocks in the camera will frame count to match the real world clocks.

Special note about consumer camcorders- A consumer camcorder with a composite video output jack could be used as a master clock because the video composite signal would contain the necessary sync required for the slaves. But this would present one limitation and that is no time code. This would not be as important in a 3D setup because one could still use the markers, clap board slate for clip match.

Now we set up our first slave camera by putting the time code into free run mode, drop frame, lock to external input. Some cameras this setting also puts the slave camera's clock to reset and lock to an external sync. Others may need a menu setting to switch from internal to external sync. A short cable is now connected between the master and the slave. In a few seconds the slave will lock onto the master and now both cameras are in lock precisely to one oscillator clock phase. The start of a frame or field will be at exactly the same moment down to the microsecond or better. Now we can discon the cable and "genlock" the third camera in our shoot. (Not necessary for a camera pair) Some newbies believe that if you disconnect the cable you will lose genlock but this is not true. Cameras will remain in genlock ( subject to some drift over many , many hours, days even) unless the slave camera's internal memory battery is dead. ( repair issue). The beauty of this system is now you can start and stop any one camera in the genlocked production on it's own and later match up the recorded clips to the exact frame by time code in post. Two cameras that have clips recorded that overlap in time code will obviously also have the same action recorded by two different cameras at EXACTLY the same moment in time, including the start of of each frame or field 1 of the first match frame.

So, what we have are cameras that are genlock capable to achieve what you are bemoaning about. The technology does exist and has worked for many years. I have done well over a thousand TV shows using this process and it became quite routine.

There is another technology that could align your frames to the exact moment you are requesting you need for your work. This is a frame store synchronizer. A bit of history- In the early days TV stations were mostly internal and never needed to mix their video with external sources for live recording and broadcast. But then along came microwave network links and later still satellite links. The problem was that the signals could not be genlocked over such a long distance because TV signals travel at the speed of light and there is a significant delay that far exceeds the nanosecond delay in a cable between two cameras. So the frame store synchronizer was added and what this does is delay one video signal to match up to another in real time so the frame start is in lock. These devices sound expensive and complex but in reality they are readily available. The first really popular frame store device was the Video Toaster. It could take 4 asynchronous video signals, even VCR's and synchronize them to match frame. I'm not sure how this would work for a 3D camera setup to mix the signals but the frame store would match the frame start for you.

So, I have given you two ways you can do what you claim you require to shoot 3D in match frame. Genlock cameras ( the simple but broadcast cost way) and a frame store synchronizer, but this may require some additional hardware to generate the 3D frame packed or SBS for display. Some frame store devices would output the independent videos too after they were frame matched. I don't recall if the Toaster did this but, you will need one that does HD resolution anyway.
Don Landis is offline  
post #745 of 1097 Old 05-29-2012, 08:07 AM
AVS Club Gold
 
Don Landis's Avatar
 
Join Date: Jun 1999
Location: Jacksonville, FL
Posts: 11,285
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 354 Post(s)
Liked: 171
New subject-

Issues with the Lanc Shepherd:
I had hoped it would slave the camcorder to the master for:
1. exposure
2. white balance
3. shutter speed
4. f-stop

I have not looked at LANC to see if these functions are even supported. Maybe they aren't.

But I had hoped that if I adjust a manual exposure in the master that the controller would adjust the slave camera accordingly.

As it stands, I have made it a practice to use the following camcorder setup-
I put both camcorders in auto for all functions including focus most of the time. except for manual exposure. But I need to adjust both cameras exposure with each shot.
In post I see a few clips are not quite matched in color and these need additional color match before pairing. I don't know why the auto white balance didn't correct for this but it didn't.


Does your ste-fra Lanc V3 slave any of these other camera settings to the master camera? Frank- how doers this work with your Canon cameras?
Don Landis is offline  
post #746 of 1097 Old 05-29-2012, 08:15 AM
AVS Club Gold
 
Don Landis's Avatar
 
Join Date: Jun 1999
Location: Jacksonville, FL
Posts: 11,285
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 354 Post(s)
Liked: 171
Quote:
It would be really nice to know what missync in ms it is, that is recognized by our brains typically.

You could measure it in microseconds up to a half field time difference with a wave form monitor and a strobe you flash at both cameras. You will see the blip in wfm trace for one cam from the strobe which has a very short pulse time and the second camera will be different, then the scope can tell you what that time difference is.
Don Landis is offline  
post #747 of 1097 Old 05-29-2012, 08:25 AM - Thread Starter
AVS Special Member
 
Frank's Avatar
 
Join Date: Apr 1999
Location: Dept. of Offense
Posts: 5,194
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 4 Post(s)
Liked: 14
Quote:
Originally Posted by Don Landis View Post

New subject-

Issues with the Lanc Shepherd:
I had hoped it would slave the camcorder to the master for:
1. exposure
2. white balance
3. shutter speed
4. f-stop

I have not looked at LANC to see if these functions are even supported. Maybe they aren't.

But I had hoped that if I adjust a manual exposure in the master that the controller would adjust the slave camera accordingly.

As it stands, I have made it a practice to use the following camcorder setup-
I put both camcorders in auto for all functions including focus most of the time. except for manual exposure. But I need to adjust both cameras exposure with each shot.
In post I see a few clips are not quite matched in color and these need additional color match before pairing. I don't know why the auto white balance didn't correct for this but it didn't.


Does your ste-fra Lanc V3 slave any of these other camera settings to the master camera? Frank- how doers this work with your Canon cameras?

I set both my cameras to auto focus and shutter priority.
The LANC standard as it is does not support control of exposure,shutter speed,white balance,f-stop, as far as I know.

Here's what we need:
A device that attaches to the touch screen and actuates it for you.
With this we can gain control over virtually all camera functions and control multiple cameras that way at the same time.
Picture a wireless remote that communicates with them that allows a single button to perform all the touch functions in succession automatically.
I'll buy several.

Or do I have to try to design it myself?

Did IQ's suddenly drop sharply while I was away?
I enjoy 3D in spite of HDMI 1.4!
Full screen only 3D doesn't cut it!
Frank is offline  
post #748 of 1097 Old 05-29-2012, 09:05 AM
AVS Special Member
 
Wolfgang S.'s Avatar
 
Join Date: Aug 2011
Location: Vienna/Austria
Posts: 1,148
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 12 Post(s)
Liked: 16
Quote:
Does your ste-fra Lanc V3 slave any of these other camera settings to the master camera? Frank- how doers this work with your Canon cameras?

With the stefra lanc v3 you can adjust settings of both cameras at the same time. You have access to the menus. But be warned: with the TD10 my impression is that it is not quite comfortable. From time to time I find myself to prefer to adjust the settings manually on both cameras.

And: Sony has invented this central bottom that allow fast access to the most important parameters, for the TD10 and also other cameras. You do not have access to that bottom with the stefra Lanc, you have to use the menus - what is more comlicated.

I beliefe that it is hard for the producers of controllers like the stefra Lanc to implement all the the changes that are done by Sony.

Kind regards,
Wolfgang
videotreffpunkt.com
Wolfgang S. is offline  
post #749 of 1097 Old 05-29-2012, 09:14 AM
AVS Special Member
 
Wolfgang S.'s Avatar
 
Join Date: Aug 2011
Location: Vienna/Austria
Posts: 1,148
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 12 Post(s)
Liked: 16
Quote:
Originally Posted by Don Landis View Post

OK, then the language barrier is why we may not have understood what your complaint is. I thought you meant the frame rate was drifting about 23.976 fps, like plus or minus 2 fps. Sorry I misunderstood your use of "drift".

I am not so sure how the drift takes place really. From a lot of muticam-editing in Vegas I know that it is no issue to import and align footage from different cameras. In the timeline of Vegas I never have seen different frame positionts - since there is ONE timeline. So what Vegas does is that it allocates the frames of the footage to the correct position in the timeline - what is fine.

I do not know what would happen if the syn has become so large that we come to a missync of a full frame. Maybe Vegas does not care and imports the frame anyway in the correct position anyway. So frame no 100 is always in the timeline as frame no 100.

But since we know that missync takes place I assume that we will see that in the content of the frams. One CMOS starts earlier to be read out line by line, the second CMOS later. I think that this will shift after some time. And if that is true then there is no way to correct that. Then the only way would be to have good synchronizes cameras.

What Frank describes for his Canon sounds great. To have the similar start up point, to have cameras that stay synchron for hours - that is what I would like to have. I think Canon shows that it is not necessary to use Genlocks - Franks cameras are consumer units too, but seems to be more stable and similar in terms of oszillation what results in lower misync. So that would be enough.

Anyway, I have decided to use TD10s, so I will look if the last unit will be better or not. And then I stop that, because it makes no sense to try endless to pair cameras.

Kind regards,
Wolfgang
videotreffpunkt.com
Wolfgang S. is offline  
post #750 of 1097 Old 05-29-2012, 06:07 PM
AVS Special Member
 
Wolfgang S.'s Avatar
 
Join Date: Aug 2011
Location: Vienna/Austria
Posts: 1,148
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 12 Post(s)
Liked: 16
According to that information, one can expect a sync period of about 15 minutes:

Quote:


These work with most camcorders that have a LANC port or AV/R port and provide good sync for periods of up to around 15 minutes at a time. To reset the sync the cameras must be powered off and back on. They work by synchronising the start up of both cameras and then measuring the sync error. The sync won't be perfect, but it will be good enough for most 3D applications. However as there will always be slight variations in the master oscillators in the cameras, over time the sync will start to drift apart. The controller will tell you how far apart the sync is and when it becomes excessive you will need to re start the cameras to bring them back into sync, typically you get between 10 minutes and 20 minutes of useful synchronisation.

Source: http://www.hurricane-rig.com/guides-...c-and-genlock/

Kind regards,
Wolfgang
videotreffpunkt.com
Wolfgang S. is offline  
Reply 3D Source Components

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off