|
|||||||||
|
Thread Tools | Search this Thread |
June 6th, 2003, 05:04 PM | #16 |
Regular Crew
Join Date: Mar 2003
Location: Amsterdam NL -Turnhout BE
Posts: 158
|
Niko, that started this discussion used a photo from my website as an example and as a proof of his statement.
To be honest I don't agree with his observations. As an answer and a contribution to the discussion I have added the unaltered and interlaced picture and the original avi to my site. So everybody who wants can study what is going on. I will be curious about the different opinions. The sun was very bright. I have years of experience with digital Photocams and I don't see anything extraordinairy in that very picture. Some parts of the flag are turned flat to the bright sun and there reflections are overexposed. But that is very normal and expected. Here is a photo taken with the newest Canon EOS D10 with some inevitable overexposed parts but not due to reflections of bright sunlight. And this one is taken with the Canon EOS D1 with clear reflections in the yellow showing Niko's "problem". On the above mentioned site videoframes are compared and dón't reveal the supposed problem. Till now I have seen not any picture from that Niko himself. I cannot imagine that he does not have any sample. That is contradictory with his complaints! So let us wait with what pictures he will come to base his observations on. |
June 6th, 2003, 05:33 PM | #17 |
New Boot
Join Date: Jun 2003
Location: Kranj, SVN
Posts: 6
|
Jan, actually I started the thread in the other forum and I just used the link to your page since I found there the picure of the flag that showed very evidently the effect I found on my TRV950. So the thread didn't start from your picture, but from my very detailed observations of the movies and pictures I made by myself. I just used yur picture as an example since I had no way to post my pictures on the web.
I made several hundreds of pictures with TRV950 that show very evidently the effect of white clipping of the yellow colour on TRV950. Some of the readers also successfully reproduced the effect. My observations were not incidental, I constructed several test pages in order to examine the effect and to define what is actually going on. After that I tested several video cameras: Sony TRV950, TRV940, TRV900, TRV75, VX2000, and Canon GL2. For comparison I also made the photos with photo camera Canon PowerShot G2. ONLY the TRV950 exhibited the effect of erroneous white clipping. Thanks for the link to your picture, taken by Canon (beautiful!). I quickly measured the parameters and Canon has performed absolutely perfect. Some overexposure (man in the middle and the crown in the left upper side) is visible, but it is shown correctly. Now I am carefully selecting and examining the test pictures and I'm going to post them on the web. I'll let you know about this and I am very interested in your opinion. I would really like somebody experienced to replicate the tests I made. Niko |
June 6th, 2003, 05:58 PM | #18 |
Regular Crew
Join Date: Mar 2003
Location: Amsterdam NL -Turnhout BE
Posts: 158
|
|
June 6th, 2003, 08:11 PM | #19 |
Wrangler
Join Date: Dec 2002
Location: Mays Landing, NJ
Posts: 11,794
|
Personally I've never seen the problem that Niko mentions, but I always shoot in full manual mode with the PDX-10...
|
June 8th, 2003, 07:57 AM | #20 |
Major Player
Join Date: May 2003
Location: St. Louis, Missouri
Posts: 936
|
Well thanks for all your input guys! I haven't checked this thread for a couple days and I'm surprised to see how it's grown...
Humor me for a minute as I "seem" to go off-topic... I love superbikes. In World Superbike competition the two dominant teams are Ducati and Honda (over the years). They have diametrically opposed strategies. Ducati does things by emotion and feel, using less science, analysis, and numbers. "How's the bike doing?" "Good, but sharpen up the front suspension..." Honda's approach is ALL numbers and datalogs and unbelievable computerized detail... The irony is that Ducati wins as often as Honda does... I'm a Ducati man even though I've mostly owned Hondas... I was hoping for some other "Ducati people" to give me some input on my original question... Great picture, good detail, low noise, no/low hiss on audio... that kind of thing. You guys are telling me the datatrace is off by 1/10th on the 3rd corner when all I really want to know is which bike is the most satisfying to ride for the money... Thanks for the input though. As I see it most of you are happy with the trv950/pdx10 though it can wash out some colors under certain conditions... Also I mostly hear good things about the vx2000 which I am considering... I'd consider a used trv900 to be an everyday toy type cam that is capable of professional results... It's just that at $1400 you are only $500 or so away from a pdx10... so to summarize it looks like I'll either go with a used trv900 or a vx2000/pd150 (even though those two cams are apples and oranges)... The thing about a pdx10 is if the audio was as good as my md system I'd get it no questions asked... the dvx100 is said to be that good... kind of nice to forget the double system and just record to the cam... Decisions... decisions. Thanks Boyd and Jan... I covered all your links... been to Beale many times before... pdx10 looking the best for now... wish it had more DOF possibilities though... like to stick with Sony... incomplete sentence assault over. |
June 9th, 2003, 12:58 PM | #21 |
New Boot
Join Date: Jun 2003
Location: Kranj, SVN
Posts: 6
|
I finally assembled a small page with some examples of the TRV950 pictures showing the white clipping effect. Jump to it:
http://www.procom.si/private/nikolaj/TRV950/index.htm Matt, I understand you, finally you just wanted to hear from somebody which camera is good for you and which is not, without a lot of data and discussion. I just used technical discussion to investigate what is going on with the TRV950 and how to remove the effect. You can freely jump over it and check my pictures to see if tha camera satisfies your needs or not. What I wanted to do with my answer to your question was to warn you before you buy. We are not talking about "the datatrace is off by 1/10th on the 3rd corner", we are talking about "the front wheel is rectangular", you know what I mean. Just look at the pictures and get your "Ducati" opinion :-) Jan, you picture is to me slightly overexposed (having two girls with dark skin in the foreground and green bushes in the background I would use a negative correction or manual mode), but nevertheless the picture is correctly processed. The white clipping on the girls' foreheads is gradual and reaches luminance value of 237 which is almost ideal 100 IRE white level. The white gradient on the girls arm is very gradual and smooth and still does not reach the white clipping value. I will try to make a similar picture with TRV950 and post it on a web. Jan, I would really like to hear your opinion about my pictures. In a few days I will post more pictures with colour gradients comparing TRV950, TRV950 and XM2 (GL2). I will also post the original pictures I constructed for the tests. Boyd, could you shoot some pictures with your camera(s) to see whether they show similar behaviour in automatic mode? (Guys, if you don't like technical discussions, just skip this post!) Thanks, Nikolaj |
June 9th, 2003, 05:44 PM | #22 |
Major Player
Join Date: May 2003
Location: St. Louis, Missouri
Posts: 936
|
Thanks for "biting" on my analogy... tee-hee :D That made me smile... I like how you expanded on it too!
I look forward to watching the progress of this thread... Thanks again guys... you keep me looking at this everyday... |
June 9th, 2003, 06:03 PM | #23 |
Wrangler
Join Date: Dec 2002
Location: Mays Landing, NJ
Posts: 11,794
|
Well you certainly have proved the adage that "a picture is worth a thousand words" ;-) The yellow flowers show the effect very clearly. I'll try to run some PDX-10 and VX-2000 tests when I get a chance.... a little busy here at the moment.
|
June 9th, 2003, 06:50 PM | #24 |
Major Player
Join Date: Sep 2002
Location: US & THEM
Posts: 827
|
i dont see anything wrong here in the sense that if you turned on the 100% zebra the washed out specular should show in the viewfinder and that would be your cue to switch to manual
try switching to manual and reduce aperture until the zebra just disappears else set custom preset to 4 cliks to left of center - thats why its there... for you to customise the cam to your needs auto can never be a catchall for every situation. bear in mind that the 950 operating system is almost a direct lift of the vx2k, which explains why it shoots 2 stops over the bar in auto mode.
__________________
John Jay Beware ***PLUGGER-BYTES*** |
June 11th, 2003, 05:32 PM | #25 |
New Boot
Join Date: Jun 2003
Location: Kranj, SVN
Posts: 6
|
John, zebra is a nice tool to come close to the proper exposure, but it still does not increase the dynamic range of the camera. In high contrast situations the white clipping in TRV950 simply cuts-off the colours above a certain lever simply too sharp, without a smooth transition. This makes the pictures to look unnatural in some areas.
I agree that auto mode can not catch up for all situatios, but it should handle properly a majority of them, which is here not the case. Custom presets are here to change the settings away from the standard settings according to a personal taste and not to aproach the standard settings. What attracted my interest was your comment that TRV950 heritated the operating system from VX2000 which shoots 2 f-stops over the bar. Can you please elaborate this. According to my comparisons the VX2000 produced much darker picture in high contrast circumstances and never had a problem with clipping. Maybe your explanation could lead to certain conclusions and solutions. BTW, I put full size original pictures to my web page and also added five MPEG movies to illustrate our discussion even more: http://www.procom.si/private/nikolaj/TRV950/index.htm Boyd, I would really appreciate your test pictures, just to verify, if the effect is present in all cameras, or just my sample is faulty. I just found a web page dealing with very similar behaviour of the JVC JY-VS200U. Very interesting, looks like TRV950 isn't the only cam with this kind of a problem: http://www.dvcentral.org/Reviews/jyvs200u.html#01_Summary Niko |
June 12th, 2003, 01:50 PM | #26 |
Major Player
Join Date: Sep 2002
Location: US & THEM
Posts: 827
|
Niko
Perhaps the meaning of my previous post has been lost in translation... In all your examples the use of the 100%zebra would have informed you that you are burning out the highlights You must take corrective action under such circumstances by reducing exposure the way you do this is a matter of personal preference but the options are 1 use the spot light button 2 add an exposure bias in custom preset 3 reduce aperture manual mode 4 increase shutter speed (not too high) OR a combination of all the above the alternative is to tolerate what SONY LAB TECHNICIANS have programmed your camcorder to do good luck
__________________
John Jay Beware ***PLUGGER-BYTES*** |
June 15th, 2003, 06:58 PM | #27 |
New Boot
Join Date: Jun 2003
Location: Kranj, SVN
Posts: 6
|
John, thank you for the step-by-step explanation. No, the meaning of your previous post has not been lost in the translation :-)
I tried all four methods that you described long ago. It helps to a certain extent, but anyway, why should I take any corrective actions in order to get proper exposure? Here are my comments to your suggestions: 1. Spot light measurement regularly produces slightly underexposed pictures. 2. I usually use -3 notches of the exposure bias (that is -3/4 f-stop). It works, results in slightly underexposed pictures and still does not remove the effect in the high contrast scenes. 3. TRV950 does not have a manual aperture mode as it had TRV900 :-( 4. Hmmm... shutter speed in TRV950? It has the most weird shutter speed setting I've ever seen. You can only adjust the mechanical shutter speed, electronic shutter speed (which is NOT shown in the screen, even not during the playback!!!) does whatever it wants, it still changes somehow even in the manual setting. ... and the combination of all measures simply doesn't remove incorrect behaviour of the camera (clipping of the colours). Please, visit my home page again, I have posted lots of test pictures, comparisons to TRV900 and XM2, and added some oscillograms to clarify my discussion. I would really like to hear your comments. And finaly, I am not willing to tolerate what the Sony lab technicians have programmed, finaly I didn't buy the second hand camera at the flea-market, I purchased a brand new device. Thanks again for your interest, Niko |
June 17th, 2003, 05:57 PM | #28 |
Major Player
Join Date: May 2003
Location: St. Louis, Missouri
Posts: 936
|
Thanks Niko... I just decided to get a trv900 afterall. I'll wait a while before pulling the trigger on something that costs more.
I appreciate the input... the links... and your observations. Thank you to everybody else too! |
| ||||||
|
Thread Tools | Search this Thread |
|