advice on wierd discoloration... at DVinfo.net

Skip Hunt · February 27th, 2003, 08:57 AM

Hey all,

I posted this over at 2-pop and haven't been able to get a response. I thought maybe someone over here could offer some advice or speculation. I've been getting some wierd discoloration off and on especially when I'm shooting something with a clear blue sky as background. To me it looks as if where the image should just be deep blue, it shifts abrubtly to having a purplish red tinge. I'm not sure if contrasty shots like this are just pushing the limits of DV25 and 1/3in CCDS? Or, if I have a malfunctioning cam on my hands. There seems to be others posting images with this discoloration as well. I thought the blue was just too saturated, but I had my chroma level cranked down to about -5 and the chroma phase at -4. Could one of these settings be causing it? I believe this was shot with WB at 5600, but I did have a A WB preset I was using as well. Or, am I just in denial and need to go ahead and deal with the horror of trying to get Panasonic to replace my cam?

I've uploaded a frame that shows what appears to be a purplish tinge in the top right-hand quadrant of the frame. I've seen this infrequently in other shots with smooth gradiation like a clear blue sky. Does anyone think there's anything wrong with this image? (other than it's not interesting ;-) I'm really wondering if I've been "working around" a cam characteristic with lowered chroma levels, etc. and that may actually be a camera flaw. Can anyone tell me if this kind of purplish tinge is completely normal for a DV shot like this?

http://www.shardt.com/vagabond/Untitled-1.html

thx,

Skip Hunt

Christopher Go · February 28th, 2003, 03:47 AM

Hey Skip, I certainly hope it isn't a problem especially what with the 24p audio sync "issue" a while back. I don't think I understand properly, did you white balance using a white object or card prior to shooting or did you just use an available preset?

Jeff Donald · February 28th, 2003, 06:45 AM

The color temperature of the sky does change. I would shoot a series of tests with a subject that has a consistent color temperature. If the tests show no change in color balance then I would chalk it up to WB and color temperature changing during the shot.

During those times when color temperature changes at a rapid rate you may want to try to manual WB more frequently.

Skip Hunt · February 28th, 2003, 08:34 AM

Chris,

I don't remember for sure. I didn't have anything white handy, so I did a manual WB from my sock for shots I was doing in the shade. For that particular shot, I'm not certain if I was using my A channel, manual WB or the 5600k preset.

Jeff,

I did have some shots of this same tower from another angle that didn't show any purple tinge. The color temp of the sky changing from one end to the other sounds plausible, but it still seems odd it would be represented so abruptly with the purplish color. Seems like it would be gradual.

Skip

Jeff Donald · February 28th, 2003, 10:33 AM

The color temperature of the sky does change. I've a Minolta color meter and the temp of the sky can change several thousand degrees, depending on angle of the sun, cloud conditions, etc. I think you probably encountered a change in color temp and an anomaly in your Auto Tracing White Balance. It may have over corrected or just made an error in White Balance.

Skip Hunt · February 28th, 2003, 01:04 PM

It's possible I suppose. However, my AWB tracking is off. I'll have to do some more testing this weekend if the sky is clear to see if I can reproduce it. I will add that it seems like its most frequently shots like the one I referenced where the intensity goes to extremes. The other shots from the shame shoot that did not exhibit this anomaly, had resonably consistant blue from corner to corner. I believe I was using both the cinematrix and cinegamma as well. I don't usually use both because I noticed strange color anomalies early on with this combo. I was just experimenting with them again because I've read they make an overall better combo. But, I'm beginning to think I was better off avoiding the cinegamma. Are you pretty sure I don't have a CCD problem? And, that I've just uncovered an anomaly that's easily worked around?

thx,

Skip Hunt

Jeff Donald · February 28th, 2003, 01:16 PM

Skip, the thing to do is see if it is repeatable. If you can consistently reproduce the effect by using certain modes or switching on or off various functions, then you might have a problem. But a one time occurrence of odd coloration in the sky does not make a manufactures defect. If it can be consistently repeated by other users then you have further evidence of a possible manufactures defect. Until further evidence I would say it is a one time glitch.

Skip Hunt · February 28th, 2003, 01:29 PM

Agreed, but this has not been a one time occurance. It has been infrequent and the discoloration is subtle, but it has occurred before. When I see others post sample images with simular coloration anamolies and find out that Panasonic ended up replacing these users cams, it makes me wonder if it is in fact a problem that Panasonic is trying to keep contained. I'll see if I can easily reproduce this problem this weekend and report back.\

thx,

Skip

Stephen van Vuuren · February 28th, 2003, 01:55 PM

Skip:

If you were any cinegamma mode, any clipping will cause issues (see www.adamwilt.com for more). The frame you posting has severe clipping in the white corner, so I suspect that's much of the reason.

Skip Hunt · February 28th, 2003, 02:47 PM

Ok... I read the adamwilt section on clipping (jeez, somebody should offer to build that guy a free search engine - took me 15mins to find it) And, although the clipping scenario also seems very plausible, Adam's article doesn't lead me to believe "white clipping" could have anything to do with it. However, his article continues with explanations for "...other codec problems..." and that explanation does seem likely. Basically, getting back to the same Quicktime Codec problem. Although his article codec references are out of date (QT4.1), I'm betting the same issue that causes certain supersaturated colors (reds inparticular) to get blocky, is also causing hue shifts in highly saturated colors. That day in particular, the sky was exceptionally blue and I was shooting about an hour before sunset. If you look at the area of my image that has the "purplish tinge", it is also kinda blocky looking and occurs in the most blue saturated portion of the image. Does anyone else think the QT DV codec itself is most likely the culprit?

thx for the reference Stephen,

Skip

Stephen van Vuuren · February 28th, 2003, 02:53 PM

Skip:

This is the specific section I was talking about:

http://www.adamwilt.com/24p/index.html#GammaSettings

Jeff Donald · February 28th, 2003, 02:57 PM

The codec is certainly a possibility also. It's easy to prove though. Give your original tape to someone with an Avid or any PC codec. If the coloration disappears, you have your culprit.

Skip Hunt · February 28th, 2003, 03:09 PM

Thanks guys!

I'll do more testing. As I stated to begin with, these occurances of strange hue shifts, etc. do seem to occur infrequently. The gamma settings, codec limitations, color temp shifts, etc. are all likely culprits. I've had the cam for 4 mos now and it hasn't presented itself as such a blaring problem to demand a replacement yet. I've been attributing occassional-to-rare hue anamolies as just that, anamolies and not flaws per se. It's just when other users report back that "Pana has acknowledged my color problem, and replaced my cam.. yada, yada" it kinda makes you wonder if Pana is just taking the easy way out with an insistant customer who believes they have a problem and won't rest until their cam is replaced... ? Or, is there really a hardware problem here? I'll reserve further judgement until I can do more tests and look for it.

thanks again,

Skip Hunt

February 27th, 2003, 08:57 AM	#1
Skip Hunt Regular Crew Join Date: May 2002 Location: Austin, Texas Posts: 37	advice on weird discoloration... Hey all, I posted this over at 2-pop and haven't been able to get a response. I thought maybe someone over here could offer some advice or speculation. I've been getting some wierd discoloration off and on especially when I'm shooting something with a clear blue sky as background. To me it looks as if where the image should just be deep blue, it shifts abrubtly to having a purplish red tinge. I'm not sure if contrasty shots like this are just pushing the limits of DV25 and 1/3in CCDS? Or, if I have a malfunctioning cam on my hands. There seems to be others posting images with this discoloration as well. I thought the blue was just too saturated, but I had my chroma level cranked down to about -5 and the chroma phase at -4. Could one of these settings be causing it? I believe this was shot with WB at 5600, but I did have a A WB preset I was using as well. Or, am I just in denial and need to go ahead and deal with the horror of trying to get Panasonic to replace my cam? I've uploaded a frame that shows what appears to be a purplish tinge in the top right-hand quadrant of the frame. I've seen this infrequently in other shots with smooth gradiation like a clear blue sky. Does anyone think there's anything wrong with this image? (other than it's not interesting ;-) I'm really wondering if I've been "working around" a cam characteristic with lowered chroma levels, etc. and that may actually be a camera flaw. Can anyone tell me if this kind of purplish tinge is completely normal for a DV shot like this? http://www.shardt.com/vagabond/Untitled-1.html thx, Skip Hunt

February 28th, 2003, 06:45 AM	#3
Jeff Donald Warden Join Date: Mar 2002 Location: Clearwater, FL Posts: 8,287	The color temperature of the sky does change. I would shoot a series of tests with a subject that has a consistent color temperature. If the tests show no change in color balance then I would chalk it up to WB and color temperature changing during the shot. During those times when color temperature changes at a rapid rate you may want to try to manual WB more frequently. __________________ Jeff Donald Carpe Diem Search DVinfo.net for quick answers \| Where to Buy? From the best in the business: DVinfo.net sponsors

February 28th, 2003, 10:33 AM	#5
Jeff Donald Warden Join Date: Mar 2002 Location: Clearwater, FL Posts: 8,287	The color temperature of the sky does change. I've a Minolta color meter and the temp of the sky can change several thousand degrees, depending on angle of the sun, cloud conditions, etc. I think you probably encountered a change in color temp and an anomaly in your Auto Tracing White Balance. It may have over corrected or just made an error in White Balance. __________________ Jeff Donald Carpe Diem Search DVinfo.net for quick answers \| Where to Buy? From the best in the business: DVinfo.net sponsors

February 28th, 2003, 01:16 PM	#7
Jeff Donald Warden Join Date: Mar 2002 Location: Clearwater, FL Posts: 8,287	Skip, the thing to do is see if it is repeatable. If you can consistently reproduce the effect by using certain modes or switching on or off various functions, then you might have a problem. But a one time occurrence of odd coloration in the sky does not make a manufactures defect. If it can be consistently repeated by other users then you have further evidence of a possible manufactures defect. Until further evidence I would say it is a one time glitch. __________________ Jeff Donald Carpe Diem Search DVinfo.net for quick answers \| Where to Buy? From the best in the business: DVinfo.net sponsors

February 28th, 2003, 01:55 PM	#9
Stephen van Vuuren Space Hipster Join Date: Oct 2001 Location: Greensboro, NC Posts: 1,508	Skip: If you were any cinegamma mode, any clipping will cause issues (see www.adamwilt.com for more). The frame you posting has severe clipping in the white corner, so I suspect that's much of the reason. __________________ stephen v2 www.insaturnsrings.com

February 28th, 2003, 03:47 AM	#2
Christopher Go Major Player Join Date: Nov 2002 Location: Honolulu, HI Posts: 429	Hey Skip, I certainly hope it isn't a problem especially what with the 24p audio sync "issue" a while back. I don't think I understand properly, did you white balance using a white object or card prior to shooting or did you just use an available preset?

February 28th, 2003, 08:34 AM	#4
Skip Hunt Regular Crew Join Date: May 2002 Location: Austin, Texas Posts: 37	Chris, I don't remember for sure. I didn't have anything white handy, so I did a manual WB from my sock for shots I was doing in the shade. For that particular shot, I'm not certain if I was using my A channel, manual WB or the 5600k preset. Jeff, I did have some shots of this same tower from another angle that didn't show any purple tinge. The color temp of the sky changing from one end to the other sounds plausible, but it still seems odd it would be represented so abruptly with the purplish color. Seems like it would be gradual. Skip

February 28th, 2003, 01:04 PM	#6
Skip Hunt Regular Crew Join Date: May 2002 Location: Austin, Texas Posts: 37	It's possible I suppose. However, my AWB tracking is off. I'll have to do some more testing this weekend if the sky is clear to see if I can reproduce it. I will add that it seems like its most frequently shots like the one I referenced where the intensity goes to extremes. The other shots from the shame shoot that did not exhibit this anomaly, had resonably consistant blue from corner to corner. I believe I was using both the cinematrix and cinegamma as well. I don't usually use both because I noticed strange color anomalies early on with this combo. I was just experimenting with them again because I've read they make an overall better combo. But, I'm beginning to think I was better off avoiding the cinegamma. Are you pretty sure I don't have a CCD problem? And, that I've just uncovered an anomaly that's easily worked around? thx, Skip Hunt

February 28th, 2003, 01:29 PM	#8
Skip Hunt Regular Crew Join Date: May 2002 Location: Austin, Texas Posts: 37	Agreed, but this has not been a one time occurance. It has been infrequent and the discoloration is subtle, but it has occurred before. When I see others post sample images with simular coloration anamolies and find out that Panasonic ended up replacing these users cams, it makes me wonder if it is in fact a problem that Panasonic is trying to keep contained. I'll see if I can easily reproduce this problem this weekend and report back.\ thx, Skip

February 28th, 2003, 02:47 PM	#10
Skip Hunt Regular Crew Join Date: May 2002 Location: Austin, Texas Posts: 37	Ok... I read the adamwilt section on clipping (jeez, somebody should offer to build that guy a free search engine - took me 15mins to find it) And, although the clipping scenario also seems very plausible, Adam's article doesn't lead me to believe "white clipping" could have anything to do with it. However, his article continues with explanations for "...other codec problems..." and that explanation does seem likely. Basically, getting back to the same Quicktime Codec problem. Although his article codec references are out of date (QT4.1), I'm betting the same issue that causes certain supersaturated colors (reds inparticular) to get blocky, is also causing hue shifts in highly saturated colors. That day in particular, the sky was exceptionally blue and I was shooting about an hour before sunset. If you look at the area of my image that has the "purplish tinge", it is also kinda blocky looking and occurs in the most blue saturated portion of the image. Does anyone else think the QT DV codec itself is most likely the culprit? thx for the reference Stephen, Skip

February 28th, 2003, 02:53 PM	#11
Stephen van Vuuren Space Hipster Join Date: Oct 2001 Location: Greensboro, NC Posts: 1,508	Skip: This is the specific section I was talking about: http://www.adamwilt.com/24p/index.html#GammaSettings __________________ stephen v2 www.insaturnsrings.com

February 28th, 2003, 02:57 PM	#12
Jeff Donald Warden Join Date: Mar 2002 Location: Clearwater, FL Posts: 8,287	The codec is certainly a possibility also. It's easy to prove though. Give your original tape to someone with an Avid or any PC codec. If the coloration disappears, you have your culprit. __________________ Jeff Donald Carpe Diem Search DVinfo.net for quick answers \| Where to Buy? From the best in the business: DVinfo.net sponsors

February 28th, 2003, 03:09 PM	#13
Skip Hunt Regular Crew Join Date: May 2002 Location: Austin, Texas Posts: 37	Thanks guys! I'll do more testing. As I stated to begin with, these occurances of strange hue shifts, etc. do seem to occur infrequently. The gamma settings, codec limitations, color temp shifts, etc. are all likely culprits. I've had the cam for 4 mos now and it hasn't presented itself as such a blaring problem to demand a replacement yet. I've been attributing occassional-to-rare hue anamolies as just that, anamolies and not flaws per se. It's just when other users report back that "Pana has acknowledged my color problem, and replaced my cam.. yada, yada" it kinda makes you wonder if Pana is just taking the easy way out with an insistant customer who believes they have a problem and won't rest until their cam is replaced... ? Or, is there really a hardware problem here? I'll reserve further judgement until I can do more tests and look for it. thanks again, Skip Hunt