GH1 Stereoscopic Rig Collaboration

[David Cole]

Hi, I take it you are aware of Panasonic H-FT012E?

The lens and the Panny consumer camcorder will be excellent for "user created", non-pro content. The fundamental problem with these solutions for professional applications is that they anamorphically scale the left and right images onto one frame (halving the horizontal resolution). This creates a format known as side-side. One of the biggest problems with side-side is that the scaling is very destructive of the sharp edges needed to provide the brain with the "edge" information that it uses to define objects in space. Combine this with the artifacts that accompany relatively low bitrate DCT-based compression and you have a recipe for headache-vision.

This has been demonstrated on a wide scale with the recent cable and satellite broadcasts of the FIFA World Cup, NASCAR, and PGA golf. Those were side-side scaled then transcoded to a fairly low bitrate for distribution. The effect, at times, made people rip the glasses off of their faces.

 

[David Cole]

David Cole, have you get any footage?

Yes, however, what I've got is with relatively long lenses (Canon 50mm's). It's the only matched set I could get my hands on (thanks, Mike!). I didn't have time to carefully align the rig, which is BAD with effective 100mm lenses.

I'll have an iris-controlling adapter for Nikon 10-24mm's tomorrow. Then, I can really take this thing out for a spin. Should have results to share this weekend.

In the interim, PM me if you want to see the first tests w/ lots of parallax and vertical disparity fix'n in post. The good news is that initial sync is damn good and drift seems to be better by half than my Sony 550v rig (which is great from 0 - 30 min).

 

[David Sykes]

David, has anybody tried powering up two SDM cameras at the same time (e.g. pressing both "power-on" buttons and then applying external power supply to both cameras)? Namely, delayed power signal can be made relatively simple with some additional electronics. If yes, what happened to video synch?

Hi Damir. Not as far as I know. An interesting and simple idea. But, as we both know, we are very busy with so many other things. EWAVR never did confirm his original comment that the movie-capture task could be paused. David

 

[crunchy]

But, as we both know, we are very busy with so many other things. EWAVR never did confirm his original comment that the movie-capture task could be paused.

It seems that there are at least three busy persons if counting EWAVR as well...
Unfortunately, I cannot try the idea yet, since two of my cameras are not working via external power supply (long story). However, when I find the way how to test it, I'll post results (probably not on this forum, since the subject is not exclusively related to GH1).

On the other hand, it is very interesting that, for GH1, video cycle starts AFTER lens check. I just hope that this is not a common case for all the cameras.

 

[crunchy]

It seems that there is a relationship between the position of the focus ring on the M4/3 lenses and camera startup delay (even with AF turned OFF). Difference positions in the focus travel result in different delay durations. With manual lenses - actual sync closely covaries with the composite video blanking sync (I have 25 test runs confirming this).

This is worrying me.... I still cannot get it. Let's think... If composite video out synch is very good and the actual synch is not good, it means that internal video cycle is not correlated to video out cycle.

So, when have you measured composite synch? Just after switching the power on, just before pressing button for "shooting" or during the actual shooting?

I have measured delay between power-on and the first signal at composite video out. It seems that this time is relatively constant. I have got about 300ms delay with relatively small deviation (about 0.25ms, which is excellent). Lens were 14-140mm in AFS and MF modes (it seems that time delay does not depend on the mode).

David S., results on Canon cameras are not very good. I could not achieve reliable synch.

 

[iWATCH3D]

Hi guys, I'm a newb here on this forum, but I am really exited of what you're getting up to. Myself I film with two 550d's and to some extent I am happy with the results, but for medium and faster action shots it just doesn't cut it because of the lack of sync. After reading this thread, I am now hesitating if to buy get GH1 from ebay that are upgradable to GH13 or should I wait until Tester13 cracks GH2. I will be selling by 550d's either way.
Also I would love to see some footage from your most succesful current attempts, is anybody uploading them anywhere?
One other thing is maybe one of forum leaders / testers could provide us all who monitor your progress with some sort of a statement update including details like:
-where has this collaboration brought us until now
-this is as far as all attempts go the best solution
-how exactly one has to proceed from setting up the cameras,starting it, ect to achive the current best sync
-what is needed to achive it ( product - price wise and/or skills - electrical wise ? )
... I think you know where I am going with this. Just an update as at this point of time.

And BTW I saw Phillip Bloom tweeted about Panasonic soon to lend him a GH2 for test and his opinion on it, and as he's got a GH13 and is a fan of it, maybe he might be able to secretly send it to you for a bit of GH213 testing and maybe comparing how it might be better or worse for 3D from the sync point of view.

Thanks again,
Michael

 

[zcream]

Is it really worth worrying about m4/3 lenses if they cause so much trouble ?
If you like, I can compile a list of c-mount lenses that work in full manual mode with an adapter. This problem goes away easily.

 

[David Sykes]

@zcream Do not forget your 57-camera rig !

Is it really worth using C-mount lenses for anything other than close-ups ?

They do not have the coverage and are designed for low-resolution media ?

Also, they are not designed for CCD use.

Technically, the chief ray should be almost perpendicular to the sensor, only a relatively small inclination is suitable.

The 'proper' lenses for the camera are designed with an exit-pupil at infinity.

That means the rays are perpendicular to the sensor.

The craze for C-mount lenses on m4/3 cameras has resulted in ridiculous prices on Ebay for certain such lenses.

The results that I have seen are quite poor and show vignetting or brightness fall-off at the edges.

 

[zcream]

David, its a small world. My rig had to wait as I ordered some regulators and now I have to solder them all.

I have a number of c-mount lenses. Look for Fujinon 25mm 1.4, Computar 25mm 1.3, Navitar 50mm f0.95 and Navitar 75mm 1.3
Sharp and full coverage (the 25mm lenses have very tiny corner darkness but it can be removed in post with a brightness mask on the corners)

On the GF1, you can also use the excellent Switar 26mm 1,1 and Navitar 25mm f0.95.
Some machining needed for most of these lenses.

I disagree with the CCD comment. CCDs use a prism (especially for the 3-ccd cameras). The GH1 is a non-prism NMOS. The film plane was also flat and many of these lenses were fine for that.

The Navitar and the Fujinon lenses are used in megapixel applications with currently available high res line scan cameras.

If anything these offer more contrast and sharpness, and may need to be softened for the film look.

@zcream Do not forget your 57-camera rig !

Is it really worth using C-mount lenses for anything other than close-ups ?

They do not have the coverage and are designed for low-resolution media ?

Also, they are not designed for CCD use.

Technically, the chief ray should be almost perpendicular to the sensor, only a relatively small inclination is suitable.

The 'proper' lenses for the camera are designed with an exit-pupil at infinity.

That means the rays are perpendicular to the sensor.

The craze for C-mount lenses on m4/3 cameras has resulted in ridiculous prices on Ebay for certain such lenses.

The results that I have seen are quite poor and show vignetting or brightness fall-off at the edges.

 

[David Cole]

So, when have you measured composite synch? Just after switching the power on, just before pressing button for "shooting" or during the actual shooting?(

I measured during the entire take, however, the correlation tests (video sync to Wimmer sync) use the sync value from when recording begins. Usually, I wait for the system to settle for 30 seconds after power-cycle - before I begin recording.

I can see no correlation between the sync offset that occurs with M4/3 lenses installed and video sync.

 

[David Cole]

is anybody uploading them anywhere?

If you PM me, I'll share an early test with you. Otherwise, I'll have tests up from a proper set of lenses in a few days (experiencing tropical storm at the moment, which is slowing things down a bit).

sort of a statement update

I'll take a quick crack at this from my point of view. Your milage may vary...

where has this collaboration brought us until now

Switching to manual lenses has resulted in initial sync that is reliability very good, using the previously discussed power-up and start record in parallel methods. Also, drift on my cameras is low enough to be a non-isssue for take lasting 20-30 min. The testing apparatus and methods have been enumerated in previous posts and anyone can try, however, I'm not sure that my results have been duplicated. This is a concern because there may be variation from camera to camera. My two GH1's are very close in serial numbers - which is a desirable condition when trying to match two cameras for a 3D rig. Others may have different results.

this is as far as all attempts go the best solution

I think sync might be solved well enough to move forward. The apparatus is pretty simple for powering up and starting recording. More elaborate (microprocessor controlled) apparatus designs have been discussed and are under development - however - it's not certain if that level of sophistication is needed for all users.

For me, I've pulled my rig off the test bench and I'm going to shoot a bit. Should have some results to post very soon.

how exactly one has to proceed ...

I hate to say this, but, I'm gonna have to suggest that you go back and review this thread. All the details for the simple controller are shared. You can also consult www.beampath3d.com for the materials and basic design. As far as skill, all you need to replicate my setup is VERY basic soldering skills.

I'm interested in moving beyond sync (even while refining the controller) and beginning to look at firmware hack parameter optimizations. I'm interested in optimizing the firmware hack to get the best possible bitrate and the lowest GOP settings possible on a "best case" class 10 card with manual lenses. I'd also like to experiment more with 720P 60 and MPEG. Motion coding is a VERY tricky component of compression that can destroy depth cues in fast-moving negative parallax (stuff that sticks out of the screen) subject matter. (my day job is as co-founder of a company with several patents on digital stereoscopic compression and we're dealing with this in a BIG way right now) If the motion coding demons can be tamed with the hack, the value to stereographers would be quite good.

 

[David Cole]

Is it really worth worrying about m4/3 lenses if they cause so much trouble ?

Not in my opinion. The Panny 7-14mm is a NICE wide-angle option however. Not a lot of manual options that are that wide.

 

[David Cole]

This is a GREAT idea for a cheap field alignment setup...

http://3dvision-blog.com/dual-camera-rig-live-stereoscopic-3d-preview-and-alignment-method/

Uses two $10 USB composite video digitizers. Brilliant!

 

[zcream]

Definitely a good idea. Perhaps we could even ask Peter Weimer to put it in an option that check for drift and warns us if the images are drifting.

Sometimes, its not obvious that there is a drift between the 2 camcorders.

 

[David Sykes]

StereoMovie Player (SMP) will also display dual streams as anaglyph ... and it is free.

I have recently been trying to align two keychain cameras for macro stereo movies in confined spaces.

I have tried using SMP and StereoscopicMultiplexer but both crash the PC, probably because the camera output is uncompressed RGB.

Windows reports the PC has recovered from a serious error.

I have tried this on two PC's, both running XP Pro.

Even though the PC is alleged to have two Enhanced (USB 2.0) controllers, I have never been able to work out how to connect to two controllers using the available USB sockets.

Windows seems to make-up its own mind how to connect real sockets to controllers.

 

[crunchy]

Is it really worth worrying about m4/3 lenses if they cause so much trouble ?

There is another problem. Let's imagine that you want to take shots simultaneously on both GH1(3) by using wired controller. You can start taking movie by wired controller only in MOVIE mode. However, in this mode, when using manual lenses AND fixed shutter speed, auto ISO does not work. It could be a problem in some situations where lighting conditions change rapidly (clouds/sun). On the other hand, when using Panasonic lenses, everything works OK.

 

[crunchy]

Windows seems to make-up its own mind how to connect real sockets to controllers.

It sound similar to my colleague's experience when he wanted to control two Canon cameras some time ago and he was not successful.

 

[zcream]

In 3d, you want nothing to be on auto. So, for a 3d application, that problem does not arise.

 

[David Cole]

Well.... you don't want anything on auto that isn't synchronized, in any case.

 

[crunchy]

In 3d, you want nothing to be on auto. So, for a 3d application, that problem does not arise.

You are right. However, sometimes, you are forced to take some 3D shots where lighting changes dramatically (interesting physics experiments with oxygen or hydrogen, change lighting on a stage and all the situations where lighting is unpredictable).

In any case, I would be glad if Vitaly could add a patch (preferable selectable within menu) which would always trigger shooting video no matter which button is pressed (shutter, external trigger or movie button).

Moreover, I am missing correction of the previous patch which allowed taking external sound AND remote control signals simultaneously (unfortunately, AFAIR that patch caused one bricked GH1).