Thread Tools
Feb 23, 2007, 03:53 PM
Registered User
FredericG's Avatar
Thread OP
Help!

OSD / I-Vision goggle / Stange Issue


My I-Vision goggles arrived today. Viewing a regular video signal looks OK. When I hook them up to my OSD module I see a very strange/annoying issue.

This what the OSD must do: When powering the system the OSD generates a blue screen with the sentence "RPV cockpit'. After a second, the GPS info appears. As there is no GPS connected for this test, it now just mentions "No NMEA". If there is a video input signal, it is displayed else the background is blue.

This all works very well when viewed on a TV or captured with my Pinnacle PCTV box (the pictures you see are taken with this device)

BUT, with the goggles strange things happen: everything looks OK the first time but when rebooting the system, the last picture can be seen trough the blue background and a phantom of the the message "No NMEA" is still (very) visible. When switching to the next screen, a phantom message "RPV cockpit" is still visible.

Switching the goggles off/on does not help but with time the phantom messages disappear. It looks like the pixels get "exhausted". I suppose that in reality the GPS values will just become one blur.


Frederic
Sign up now
to remove ads between posts
Feb 23, 2007, 07:22 PM
Oxford Panic
AndyOne's Avatar
Fredric,

I've noticed strange after-images with my Rimax video goggles. After any high contrast image has been in the same place in the picture for a while there seems to be a persistence of vision thing when switching to another view. This, I think, is an artifact of this kind of LCD display. Video Goggles are far from perfect yet and I expect there will have to be a lot more development before they loose some of their bad habits.

Andy.
Feb 24, 2007, 07:47 AM
Live FPV or Die
VRflyer's Avatar
Andy: I've notice the same effect with the I-Vision, after-image (persistence in French). It look like a shadow image over the picture. It hapen when the camera do not move for few second. I never seen that with my old I-glasses goggle.
Frederic: I don't know what to say. I've notice some time on my I-Vision, the picture stay black after power up, I toggle the power switch several time and the picture appear, it happen when I change video source ,like watching DVD and plug it into the camera receiver after, perhaps the PAL/NTSC detector have problem to lock on the NTSC signal, I don't know. But if I don't change video source, it work well each time I power up the goggle.
Feb 24, 2007, 11:15 AM
Registered User
hihptsi's Avatar
Quote:
Originally Posted by VRflyer
Andy: I've notice the same effect with the I-Vision, after-image (persistence in French). It look like a shadow image over the picture. It hapen when the camera do not move for few second. I never seen that with my old I-glasses goggle.
Frederic: I don't know what to say. I've notice some time on my I-Vision, the picture stay black after power up, I toggle the power switch several time and the picture appear, it happen when I change video source ,like watching DVD and plug it into the camera receiver after, perhaps the PAL/NTSC detector have problem to lock on the NTSC signal, I don't know. But if I don't change video source, it work well each time I power up the goggle.

so in your honest opinion are the i vision glasses worth it?is the picture good.dispite what problems you seem to be having with it..

have a good one
walter
Feb 24, 2007, 11:32 AM
Registered User
FredericG's Avatar
Thread OP
Andy and VRflier, thanks for your comments.

I did some experimenting and the short conclusion is that the image-persistency issue is blown out of proportion by the background generated by the OSD. In the real application where the background is the camera image, all looks fine

When the blue background is on, one can see the last image. A counter that is incremented each second becomes soon a big unreadable blur. At the bottom of the screen there is an arrow moving horizontally that indicates the direction to the pilot. After a while there are arrows everywhere... With the camera background, however, all these affects are gone.

There is something else with the background. While it should be uniformly blue, it is only blue at the borders, in the center it is completely white. Also, after this background has been displayed and it is replaced by the camera view, the screen blinks for some minutes.
I have changed the background color to black. It looks better but still the center is white.

In the case the background is the camera image, the OSD chip keeps the signal from the camera and replaces it where needed to form the characters. This is know as "mixed" mode. In the "full screen" mode, the OSD chip generates the complete screen. There must be something there that upsets the goggle. However this very same signal works well on a regular TV.

Anyway, I will avoid using the mixed mode with the goggle.

Frederic
Feb 24, 2007, 12:14 PM
Happy FPV flyer
Kilrah's Avatar
Are you sure your levels (including sync) and impedances (75Ohm terminations everywhere) are good in your entire video chain? Proper AC coupling all the way?

I've never noticed that sort of problems on these goggles but with Archos A/V players. Proper video gear should have AC coupled inputs and outputs, all with 75Ohm impedance. It happens some devices for some reason have skipped some of that like the AC coupling, giving an all-positive signal (sync is at 0V instead of being negative) that is not really compliant with the usual standards. As the device that recieves the signal is supposed to have an AC-coupled input that doesn't cause problems. According to my scope it seems that's what Archos player do. Now, if the display device also has skipped that, we get into trouble. It seems it's the case with the goggles.
If you plug them together you get an image for a split second, and then the screens slowly turn to black (or white, don't remember) from the center to the edges, as they're not updated by the controller anymore because it has lost sync.

My guess is that it's a bit the same with your setup. For some reason the goggles only get a sync and thus update the LCD's once in a while instead of each frame. That gives the "slow response" feeling and would explain the white center thingy from my observations.
Feb 24, 2007, 01:46 PM
Oxford Panic
AndyOne's Avatar
I don't think it has anything to do with the quality of the video signal or the impedance of the line. It is quite simply down to the quality of the LCDs and optics used. Even though these devices seem expensive to us they are still in their infancy and a lot of development needs to be done to make them better at an affordable price. If you do a search you'll find military quality video goggles but they sell between $10,000 and $15,000. Leave it a few years and what was "military" will become "consumer" but for the time being we have to use what we can get.

Andy.
Feb 24, 2007, 01:47 PM
Registered User
FredericG's Avatar
Thread OP
Thanks for the explanation Kilrah. I don't know too much of analog electronics and I can only partially follow what you are saying.

The OSD chip is a STV5730A on this commercial product: http://www.blackboxcamera.com/stv5730a/applications.htm. (BTW, it seems they now mention my project on their site ) I have used small coax cable (suppose it is 75 Ohm) and original S-video cables that came with video equipment.
A strong indication is perhaps that the problems only occur when it is the OSD chip that generates the complete video signal and not when it is "patching" an existing signal. Could it be there something is not completely correct with this signal?

Does it learn us something that the same problems occur with and without video TX/RX in the chain? Does this guarantee correct AC coupling?

Thanks,
Frederic
Feb 24, 2007, 02:07 PM
Happy FPV flyer
Kilrah's Avatar
Quote:
Originally Posted by AndyOne
I don't think it has anything to do with the quality of the video signal or the impedance of the line. It is quite simply down to the quality of the LCDs and optics used.
I don't agree. My tests with these goggles have shown that they work well, apart that they indeed are a bit picky on the signal. And as I said, that has nothing to do with the LCDs, but with the controller.

Regarding the time thing, the goggles I use for flying are 8 year-old Sony Glasstrons that still are among the top models taking all aspects into account. And they were $3500 8 years ago, for a high-end model. So, there would be no need to spend $15k to get something decent today. I'd expect equivalent quality for $1500 today taking the progress that is made in such a timeframe into account.
That's why I've been very critical about the Z800s.
The little problem is that in these days electronics must be cheap mass-consumption products. Manufacturers just want to be on the safe side and produce/sell cheap stuff rather than risking too much developing a good $1000-1500 device that wouldn't sell. That just wouldn't fit in the market of ipod owners watching movies in the train, which is the current target market, and until we draw thousands of people into FPV that won't count for them either. The technology has been there for a long time, but the market isn't big enough. That's why Sony only produced goggles for a couple of years before dropping them, and why you only get either cheap and average stuff, or professional / military multi-thousand $ stuff today.

Back to the OSD issue:
Having the schematic would help. If you're not that much into analog stuff I guess you used something advised by the chip's manufacturer?
Do you have an oscilloscope handy?
Last edited by Kilrah; Feb 24, 2007 at 02:16 PM.
Feb 24, 2007, 02:22 PM
Registered User
FredericG's Avatar
Thread OP
Quote:
Having the schematic would help. If you're not that much into analog stuff I guess you used something advised by the chip's manufacturer?
Well I did not design or build the PCB, it is an of-the-shelf PCB with 2 RCA connectors and a power connector. One of the reasons I went for this OSD solution was that I did not have to make a PCB.

As this is not essential (I can live without full screen mode), I don't think we should put lots of effort debugging somebodies else design.
On the other hand I am curious. I have a scope, is there something simple (and simple to explain) I could check? I could also ask a colleague at work who has experience with video. What should we look for?

Thanks,
Frederic
Feb 24, 2007, 02:32 PM
Happy FPV flyer
Kilrah's Avatar
I'd start by measuring the signal at the output of the OSD in the different modes, check if the levels / amplitudes match the video source's ones (when it is connected to a display, for example directly to the goggles). If your colleague has experience with video he will know what to look for more precisely, but comparing those would be a good start.
Feb 24, 2007, 05:23 PM
Spanish UAV
eladiomf's Avatar
Quote:
Originally Posted by Kilrah
Are you sure your levels (including sync) and impedances (75Ohm terminations everywhere) are good in your entire video chain? Proper AC coupling all the way?

I've never noticed that sort of problems on these goggles but with Archos A/V players. Proper video gear should have AC coupled inputs and outputs, all with 75Ohm impedance. It happens some devices for some reason have skipped some of that like the AC coupling, giving an all-positive signal (sync is at 0V instead of being negative) that is not really compliant with the usual standards. As the device that recieves the signal is supposed to have an AC-coupled input that doesn't cause problems. According to my scope it seems that's what Archos player do. Now, if the display device also has skipped that, we get into trouble. It seems it's the case with the goggles.
If you plug them together you get an image for a split second, and then the screens slowly turn to black (or white, don't remember) from the center to the edges, as they're not updated by the controller anymore because it has lost sync.

My guess is that it's a bit the same with your setup. For some reason the goggles only get a sync and thus update the LCD's once in a while instead of each frame. That gives the "slow response" feeling and would explain the white center thingy from my observations.

I have this problem when I connect my I-glasses gvd520 to my archos. Is there any solution or modification so that they can work together?

Eladio
Feb 24, 2007, 05:38 PM
Happy FPV flyer
Kilrah's Avatar
I had a relatively quick try at putting something togehter but didn't come to an easy solution, which actually is weird. On the scope, it clearly appears that an AC coupling is missing. The problem is that putting a capacitor in line with the signal doesn't work. I'm getting at the edge of my analog knowledge here, but if I'm not mistaken, and from my logic, to be able to use a simple AC coupling capacitor it needs to be on the device's side of the line termination resistor, or it will do a nasty filter. That would require disassembling the unit and having access to this point, assuming it's actually reachable in the device. Otherwise an active device would be needed.
Maybe I should have a go with my new video splitter. That could maybe improve things if it has an AC-coupling. I don't have an Archos handy right now though.
Last edited by Kilrah; Feb 24, 2007 at 05:43 PM.
Feb 27, 2007, 09:47 AM
Fidler & twidler
empeabee's Avatar
Quote:
Originally Posted by Kilrah
I had a relatively quick try at putting something togehter but didn't come to an easy solution, which actually is weird. On the scope, it clearly appears that an AC coupling is missing. The problem is that putting a capacitor in line with the signal doesn't work. I'm getting at the edge of my analog knowledge here, but if I'm not mistaken, and from my logic, to be able to use a simple AC coupling capacitor it needs to be on the device's side of the line termination resistor, or it will do a nasty filter. That would require disassembling the unit and having access to this point, assuming it's actually reachable in the device. Otherwise an active device would be needed.
Maybe I should have a go with my new video splitter. That could maybe improve things if it has an AC-coupling. I don't have an Archos handy right now though.
You just twinged a 40 year old memory - though it is only a keyword - Black Level Clamping - it was a problem in 50's/60's home built TV sets...
? google ?
Mike
Feb 27, 2007, 10:41 AM
Live FPV or Die
VRflyer's Avatar
Quote:
Originally Posted by eladiomf
I have this problem when I connect my I-glasses gvd520 to my archos. Is there any solution or modification so that they can work together?

Eladio
Me too I have this problem when I plug my I-Vision(GVD520) to my Archos AV500


Quick Reply
Message:

Thread Tools