Hidden service menus?

Ken Hodson · July 26th, 2005, 12:18 PM

Yah, but we are JVC users. Much smarter. :>)

Leo Bodnar · July 26th, 2005, 03:17 PM

Quote:

Originally Posted by Allen Lu

Question on the cable needed for the HD1 to get to the eeprom, is it the same one as the one for JVC's DV cams?

It seems to be unique to HD1/HD10/PD1. The problem is in the connector itself. It is a very fine-pitched and hard to find part. I had to cannibalise one from a broken digital camera. Tried to get the cable from JVC but nobody sells them!

Wayne Morellini · July 27th, 2005, 03:25 AM

Quote:

Originally Posted by Leo Bodnar

I have a feeling that there is no single switch that would enable NTSC/PAL or 24 operation. There could be literally hundreds of memory locations to adjust with time/pixels/frequency parameters. I can comment on the possibility after I will have compared PD1 and HD1 or HD10 EEPROM memory dumps (I have PD1.)

I think most of this stuff (except for gain) would require programming, pity that the component out was not live uncompressed, that would make the camera descent. For you guys in the US, 30fps is not so much of a problem for TV, so gain would be the more important one.

Ken Hodson · July 27th, 2005, 11:53 AM

The PD1 should have uncompressed component out. The SD mode of the HD1/10 are.
My feeling is that the gain ability of the cam is quite limited. Upping the gain will probably incure massive video noise. The auto setting was probably tweaked to be as good as it gets.

Wayne Morellini · July 28th, 2005, 08:14 PM

HD component out would be good.

I understand there is particular problems with noise in the reds, and that the gain was lopped off to 0db (why the 35lux). But I would prefer (and with the newer sensor revision, if it exists) gain and some noise than a black image in some situations.

But I didn't buy a HD1 (I live in a 25p/50i market) has anybody heard of the HD1 replacement that is supposed to be due out soon, does it use the new 1/3rd inch Altasens? Wouldn't it be good if it turned out better than the Sony HC1. You know in the sensor industry there has been some good recent movements. Cirrus Semiconductor has bought both Fill factory and smalsensor, both use two versions of high latitude sampling, Fillfactory with high fill-factor and some other advancements. Altasens has the cheap 1/3inch mentioned, probably using their high efficiency low noise designs. Foveon is rumoured to becoming out with a cinema version for their x3 chip. So, they all have some thing to offer for interesting cameras (particularly if they include some of the features of their competitors).

Xander Christ · July 29th, 2005, 04:25 AM

A while back, I tried using the HD10U for greenscreen work and yeah, the whole image sensor seems riddled with red noise so the green screens looked a little grey and the macroblocking of MPEG-2 didn't help. However, thanks to Leo, he pointed out that the HD1/10 uses CMYW as the filter array on the CCD.

I did some research on that type of array and started looking at algorithms for color conversion (CMY -> RGB -> YUV). CMY has a better spectral response according to all the charts I've seen, but the fact there are multiple transforms to go through to end up with YUV might be causing some issues. I also wonder if it's true that the HD10U uses BT601 colorspace vs. the HD standard of BT709.

Sony makes their own CCDs and CMOSs. JVC has used AltaSens in the past, so chances are JVC will continue to use them since R&D efforts have already been established. As far as an HD1 replacement, I haven't heard anything; the only thing coming out of JVC that gets any mention is the 100U.

Perhaps JVC can't compete with Sony in the consumer market now that HDV is somewhat a toddler and out of infancy. JVC had HDV all wrapped up until the FX1 hit. They've always been kind of a bleeding edge/maverick of a company which is why so many people don't think too highly of them.

I'm a little wary of enabling gain on the HD10U for one reason: MPEG compression compresses better with no noise. If you crank up the gain, chances are the B-frames are gonna get screwed (misprediction due to the movement of noise). But it'll be nice to know at least we can control the gain once we find the bit. I think JVC left out gain control so it doesn't screw up their short-GOP compression (maybe gain breaks it).

Leo Bodnar · July 29th, 2005, 06:39 AM

I have more on my plate at the moment that I can cope with - e.g. I have found how to separate EE and sharpness. And I think I have found gain and knee control settings too. I need a controlled way to measure things that I am altering.

Colour conversion is a tricky beast. There is a good primer on WhGrCyYe matrix used in HD10 here: http://images.videosystems.com/files...sexpertise.pdf

You are right, Xander, there are so many stages in getting YUV that it is difficult to pick the optimal setting for all the stages. At the moment I have some great low light results but at the expense of colour saturation and separation.

Fundamental difference between 3 and 1 CCD cameras is that in 1CCD camera all four (or three) pixel colours are digitised by the same ADC and it is impossible to control ADC parameters (shift and gain) independently (i.e. on the fly). In 3 CCD camera you can pre-adjust the ADCs separately (gain is the first stage of ADC conversion process and is applied to analogue signal before the signal hits the ADC itself, even though it is done inside the ADC chip), therefore excess in one part of the spectrum , e.g. Red for flood lights will still allow Blue CCD chip ADC to use independent higher gain to pick up blue details.

What happens in 1CCD and especially with WGCY matrix, you have only one ADC and only one chance to set the gain to squeeze CCD output into ADC dynamic range. Naturally it is set to avoid overexposure. But because White pixels on WGCY matrix have nearly twice the sensitivity of Cyan, the ADC produces readings for Cyan pixels at only half of its dynamic range and it has later to be doubled in software. Thus noise is doubled too. Now if you consider that Cyan is opposite to Red and it is Cyan pixels that play major role in Red colours separation from the rest (Red = White - Cyan), this might explain bad handling and excessive noise in Reds in JVC cams.

Does it make any sense?

Wayne Morellini · July 29th, 2005, 07:58 AM

Perfectly Leo. Xander thanks as well. Gain like this is not a total loss, in guerrilla Documentary style, some sort of image in a dark place would be better than nothing.

Has anybody every broken one of these apart and applied a SLR lens through a condenser to get a brighter image, like some Canon Lens adaptors do for the XL1s?

Ken Hodson · July 29th, 2005, 12:08 PM

"I have found how to separate EE and sharpness. And I think I have found gain and knee control settings too."
Wow, that is fantastic.

On a downside, untill someone can produce the needed cable for a decent cost, or provide a DIY for dummies, this experimentation may end at Leo. It has been said JVC does not sell the cable, so we need a solution for the masses (well the brave few.)

**Note**
I have reviewed the whole thread and realize it was stated that the plans at http://pagina.vizzavi.pt/~ng93827a/JVC/
are said to work with the HD JVC's. Good news. Does anyone want to add any tip'n'tricks to the info provided at this page?

Xander Christ · July 29th, 2005, 01:23 PM

Wayne - regarding removing the current lens and attaching a 35mm lens to the HD10, a friend of mine is working on such a thing right now. However, I think he's looking to attach a 2/3" lens since the multiplication factor will only be 2x instead of 7x (1/2" lens would be 1.3x, but I don't have one of those to give him).

Ken - the PD1/HD1/HD10's cable assembly from JVC costs $275. If there's enough people wanting to get one, I can perhaps place a quantity order and reduce the cost? I built the cable shown on the website noted above, but trust me, it's not worth the hassle (I'm a novice solderer and had issues). $275 is reasonable. :)

Leo Bodnar · July 29th, 2005, 02:13 PM

Xander, the chip in PD1/HD1/HD10 is ICX426AU which smells to me like a Sony part. They also have a gap in their lineup with that number so I assume it was a custom order.

So you have a cable!? I desparately need a dump of HD1/HD10 EEPROM to move any further. I was even considering buying one for the purpose of downloading it!

Allen Lu · July 29th, 2005, 04:21 PM

Xander,

We got two very similar threads going..

$275 is reasonable. My Paypal is ready to go.

I'm willing to provide the dump as fast as this transaction will go and that my GR HD1 wont blow.

Ok one more question..is the JLIP connected behind that little secret square door next to the iLink port?

Leo, I'd be still willing to donate towards your efforts on finding all the adjustments.

Allen

Leo Bodnar · July 29th, 2005, 05:00 PM

Quote:

Originally Posted by Allen Lu

$275 is reasonable. My Paypal is ready to go.

I'm willing to provide the dump as fast as this transaction will go and that my GR HD1 wont blow.

Ok one more question..is the JLIP connected behind that little secret square door next to the iLink port?

Here is a picture of my connector: http://www.lbodnar.dsl.pipex.com/video/plug.jpg
I found the plug that is the right pitch but slightly larger, so I had to file it a bit then epoxied it to breadboard and soldered three tiny wires. It also has a MAX232 level converter at the other end of the cable.

I can make a few commercial quality ones if I find the right connector that would be the right size. However my current thought is about an inch-sized self-contained unit with Microchip's PIC controller inside, couple of push-buttons and maybe a LED, that can be plugged in the camera in the field for quick settings changes.

Yes, "the thingy" is behind that little secret door!

Wayne Morellini · July 30th, 2005, 07:05 AM

Quote:

Originally Posted by Xander Christ

Wayne - regarding removing the current lens and attaching a 35mm lens to the HD10, a friend of mine is working on such a thing right now. However, I think he's looking to attach a 2/3" lens since the multiplication factor will only be 2x instead of 7x (1/2" lens would be 1.3x, but I don't have one of those to give him). :)

Tell him to use a condenser instead, that way you should get no magnification factor with the right condenser, and some stops more in light (with right lens). That may get him to 9 Lux with a very fast lens, maybe a good SLR could get down to 5 lux etc. There is a Canon adaptor that does this so you can use their 35mm SLR lens on the Canon in this way. Search for my first post on DVinfo (on increasing camera image brightness) and go down towards the end of the thread there should be a link to another thread that talks about that adaptor. There might be an aperture limit to how far you can open up to, where the microlens array distorts the image because the angle of light is too great.

Thanks

Wayne.

Xander Christ · July 30th, 2005, 09:46 PM

To all who are interested:

I can get a group discount for the official JVC service cable if I can order 5 or more. THe cost goes from $275 to $250. Email me off list is you want to order one.

July 26th, 2005, 12:18 PM	#76
Ken Hodson Trustee Join Date: Aug 2003 Location: Vancouver BC Canada Posts: 1,315	Yah, but we are JVC users. Much smarter. :>) __________________ Damnit Jim, I'm a film maker not a sysytems tech.

July 27th, 2005, 11:53 AM	#79
Ken Hodson Trustee Join Date: Aug 2003 Location: Vancouver BC Canada Posts: 1,315	The PD1 should have uncompressed component out. The SD mode of the HD1/10 are. My feeling is that the gain ability of the cam is quite limited. Upping the gain will probably incure massive video noise. The auto setting was probably tweaked to be as good as it gets. __________________ Damnit Jim, I'm a film maker not a sysytems tech.

July 29th, 2005, 12:08 PM	#84
Ken Hodson Trustee Join Date: Aug 2003 Location: Vancouver BC Canada Posts: 1,315	"I have found how to separate EE and sharpness. And I think I have found gain and knee control settings too." Wow, that is fantastic. On a downside, untill someone can produce the needed cable for a decent cost, or provide a DIY for dummies, this experimentation may end at Leo. It has been said JVC does not sell the cable, so we need a solution for the masses (well the brave few.) Note I have reviewed the whole thread and realize it was stated that the plans at http://pagina.vizzavi.pt/~ng93827a/JVC/ are said to work with the HD JVC's. Good news. Does anyone want to add any tip'n'tricks to the info provided at this page? __________________ Damnit Jim, I'm a film maker not a sysytems tech. Last edited by Ken Hodson; July 29th, 2005 at 12:50 PM.

July 30th, 2005, 09:46 PM	#90
Xander Christ Regular Crew Join Date: Jul 2004 Location: Minneapolis, MN Posts: 115	JLIP JIG for PD1/HD1/HD10 To all who are interested: I can get a group discount for the official JVC service cable if I can order 5 or more. THe cost goes from $275 to $250. Email me off list is you want to order one.

July 28th, 2005, 08:14 PM	#80
Wayne Morellini Inner Circle Join Date: May 2003 Location: Australia Posts: 2,762	HD component out would be good. I understand there is particular problems with noise in the reds, and that the gain was lopped off to 0db (why the 35lux). But I would prefer (and with the newer sensor revision, if it exists) gain and some noise than a black image in some situations. But I didn't buy a HD1 (I live in a 25p/50i market) has anybody heard of the HD1 replacement that is supposed to be due out soon, does it use the new 1/3rd inch Altasens? Wouldn't it be good if it turned out better than the Sony HC1. You know in the sensor industry there has been some good recent movements. Cirrus Semiconductor has bought both Fill factory and smalsensor, both use two versions of high latitude sampling, Fillfactory with high fill-factor and some other advancements. Altasens has the cheap 1/3inch mentioned, probably using their high efficiency low noise designs. Foveon is rumoured to becoming out with a cinema version for their x3 chip. So, they all have some thing to offer for interesting cameras (particularly if they include some of the features of their competitors).

July 29th, 2005, 04:25 AM	#81
Xander Christ Regular Crew Join Date: Jul 2004 Location: Minneapolis, MN Posts: 115	A while back, I tried using the HD10U for greenscreen work and yeah, the whole image sensor seems riddled with red noise so the green screens looked a little grey and the macroblocking of MPEG-2 didn't help. However, thanks to Leo, he pointed out that the HD1/10 uses CMYW as the filter array on the CCD. I did some research on that type of array and started looking at algorithms for color conversion (CMY -> RGB -> YUV). CMY has a better spectral response according to all the charts I've seen, but the fact there are multiple transforms to go through to end up with YUV might be causing some issues. I also wonder if it's true that the HD10U uses BT601 colorspace vs. the HD standard of BT709. Sony makes their own CCDs and CMOSs. JVC has used AltaSens in the past, so chances are JVC will continue to use them since R&D efforts have already been established. As far as an HD1 replacement, I haven't heard anything; the only thing coming out of JVC that gets any mention is the 100U. Perhaps JVC can't compete with Sony in the consumer market now that HDV is somewhat a toddler and out of infancy. JVC had HDV all wrapped up until the FX1 hit. They've always been kind of a bleeding edge/maverick of a company which is why so many people don't think too highly of them. I'm a little wary of enabling gain on the HD10U for one reason: MPEG compression compresses better with no noise. If you crank up the gain, chances are the B-frames are gonna get screwed (misprediction due to the movement of noise). But it'll be nice to know at least we can control the gain once we find the bit. I think JVC left out gain control so it doesn't screw up their short-GOP compression (maybe gain breaks it).

July 29th, 2005, 06:39 AM	#82
Leo Bodnar Regular Crew Join Date: Jul 2005 Posts: 26	I have more on my plate at the moment that I can cope with - e.g. I have found how to separate EE and sharpness. And I think I have found gain and knee control settings too. I need a controlled way to measure things that I am altering. Colour conversion is a tricky beast. There is a good primer on WhGrCyYe matrix used in HD10 here: http://images.videosystems.com/files...sexpertise.pdf You are right, Xander, there are so many stages in getting YUV that it is difficult to pick the optimal setting for all the stages. At the moment I have some great low light results but at the expense of colour saturation and separation. Fundamental difference between 3 and 1 CCD cameras is that in 1CCD camera all four (or three) pixel colours are digitised by the same ADC and it is impossible to control ADC parameters (shift and gain) independently (i.e. on the fly). In 3 CCD camera you can pre-adjust the ADCs separately (gain is the first stage of ADC conversion process and is applied to analogue signal before the signal hits the ADC itself, even though it is done inside the ADC chip), therefore excess in one part of the spectrum , e.g. Red for flood lights will still allow Blue CCD chip ADC to use independent higher gain to pick up blue details. What happens in 1CCD and especially with WGCY matrix, you have only one ADC and only one chance to set the gain to squeeze CCD output into ADC dynamic range. Naturally it is set to avoid overexposure. But because White pixels on WGCY matrix have nearly twice the sensitivity of Cyan, the ADC produces readings for Cyan pixels at only half of its dynamic range and it has later to be doubled in software. Thus noise is doubled too. Now if you consider that Cyan is opposite to Red and it is Cyan pixels that play major role in Red colours separation from the rest (Red = White - Cyan), this might explain bad handling and excessive noise in Reds in JVC cams. Does it make any sense?

July 29th, 2005, 07:58 AM	#83
Wayne Morellini Inner Circle Join Date: May 2003 Location: Australia Posts: 2,762	Perfectly Leo. Xander thanks as well. Gain like this is not a total loss, in guerrilla Documentary style, some sort of image in a dark place would be better than nothing. Has anybody every broken one of these apart and applied a SLR lens through a condenser to get a brighter image, like some Canon Lens adaptors do for the XL1s?

July 29th, 2005, 01:23 PM	#85
Xander Christ Regular Crew Join Date: Jul 2004 Location: Minneapolis, MN Posts: 115	Wayne - regarding removing the current lens and attaching a 35mm lens to the HD10, a friend of mine is working on such a thing right now. However, I think he's looking to attach a 2/3" lens since the multiplication factor will only be 2x instead of 7x (1/2" lens would be 1.3x, but I don't have one of those to give him). Ken - the PD1/HD1/HD10's cable assembly from JVC costs $275. If there's enough people wanting to get one, I can perhaps place a quantity order and reduce the cost? I built the cable shown on the website noted above, but trust me, it's not worth the hassle (I'm a novice solderer and had issues). $275 is reasonable. :)

July 29th, 2005, 02:13 PM	#86
Leo Bodnar Regular Crew Join Date: Jul 2005 Posts: 26	Xander, the chip in PD1/HD1/HD10 is ICX426AU which smells to me like a Sony part. They also have a gap in their lineup with that number so I assume it was a custom order. So you have a cable!? I desparately need a dump of HD1/HD10 EEPROM to move any further. I was even considering buying one for the purpose of downloading it!

July 29th, 2005, 04:21 PM	#87
Allen Lu Regular Crew Join Date: Jul 2005 Posts: 65	Xander, We got two very similar threads going.. $275 is reasonable. My Paypal is ready to go. I'm willing to provide the dump as fast as this transaction will go and that my GR HD1 wont blow. Ok one more question..is the JLIP connected behind that little secret square door next to the iLink port? Leo, I'd be still willing to donate towards your efforts on finding all the adjustments. Allen