Second Impressions: Thoughts on Color & Sharpness

[Nick Hiltgen]

Well before everyone get's hung up bashing HDTV's I Have a sony 42" that is definitely not top of the line. But when I plaed backa DVD through component inputs of footage shot on a DVX100 (no A) it looked really good, and that was pretty lousy compression, I sincerely think that something might have been wrong with the test set up at circuti city (maybe you should try best buy... just kidding) I sincerely believe that until there is a portable platform for displaying HD your best bet is to go with the best Sd you can get your hands on. For most people that was the DVX-100 but I believe that is now the XL2.

What I have noticed, HDTV CRT's (tube), such as Barry G's WEGA, look much better and are much more forgiving with SD material. The DLP, plasmas, LCD, projection, etc., from what I have seen, all look like garbage with SD material. Maybe there are some that look good but I have yet to see one.

[Yi Fong Yu]

often at stores like CC they have to use special devices to split one signal and feed 10+ displays to demo TVs/HDTVs. are you sure you used a DVDplayer that is hooked up directly into ONE display or via that splitter? those splitters often damage the original signal. i would not judge HDTV via that method. best method is one DVD player, hooked into the HDTV and then recalibrated/color corrected using the AVIA or HT discs.

[Christopher Go]

I agree with Yi Fong Yu, those demo units at Circuit City are never calibrated correctly, and although they are meant to be used as examples of what you could buy, these monitors are just pulled out of their boxes and plugged in for show.

In fact, I think many people fail to calibrate their monitors, which defeats the whole purpose of buying based on specs, reviews, and tests. Too many "bigger so it's better" kind of mindsets. Who cares if your camera does this and that if you watch it on some crap monitor, right? To me, this is like placing an excellent shotgun mic right on the camera instead of on a boom pole for dramatic shots - I mean, why did you buy it then?

Sorry for the rant.

[Marty Hudzik]

Well what I do know is this was the TV that they have sitting in there little theater like room and it was playing jurassic park off of a single DVD player. That looked fine. Not Hi Def good but at least as good as standard def on a standard def tv. We directly popped out that disk and put in mine. And it looked bad. Something weird was going on because I have been authoiring DVD's now for 3+ years and know how to compress mpg2. So I know it wasn't the DVD. Lets just forget the whole HiDef thing for now. Sorry I brought it up. I went there expecting to be blown away since that screen might be able to resolve the XL2 better.

Peace All!

[A. J. deLange]

Yes, there is an adapter for Nikon 35 mm lenses from PS and I haven't put the 16x back on my XL1s since I got one. The results with my old Nikon lenses are amazing. You simply cannot expect a 16x or 20x zoom lens to compare in performance to a prime lens, especially one that is only being used close to its axis. The newer Nikkor (17-35mm, 28-70mm) zooms also give great results. I haven't tried the 20x (nor any other aspect of the XL2 yet - tomorrow!) but don't expect it to be much sharper than the 16. Nonetheless, it is a tremendous acheivement at less than $2K (IMO).

As for color saturation: I much prefer that the camera accurately reproduce color. If I want more saturation for a particular shot I'll put it in during post. But the public wants highly saturated color. Look at the color films being manufactured today. On the still photography boards the posters complain about how flat raw files converted with one piece of software compared to raw files converted with another. The preferred ones use Kodak's digital sciences which boost the saturation (just as their films and the films of other manufacturers do). No criticism of Kodak is meant here. They are just giving their customers what they want and they do make pro films with natural color rendition. I expect that the manufacturers must go round and round on this when trying to design a new camera (still or video). It's plain what they should do for the low end market - goose the saturation. But what about the prosumer (e.g. XL2) user? He should, presumably, understand enough color theory to appreciate the implications of accurate color reproduction but clearly many prosumers (and even pros) in the still camera market do not. I'm too new to the video arena to have an appreciation for the feelings here.

[Barry Goyette]

I've posted a few more clips taken with the settings I posted at the beginning of this thread. You will see a few clips in the middle where I've cycled the CP off and on to show the effect of the preset. What impresses me most so far about the color and gamma controls on the xl2, beyond the ability to oversaturate colors beyond any need...is that it looks pretty good doing it...very little posterization or artifacts from the adjustments, and second, the dynamic range with cinegamma and cinecolor is just amazing (I have a clip in 24p that I'll post later that blew me away in terms of highlight and shadow detail/color) but for now note the closeup of the chalk with the shadow (almost as good).

Realistically, these clips are oversaturated far beyond what was visible at the festival, but I'm posting them to show some of the color range of the camera.

http://homepage.mac.com/barrygoyette/FileSharing24.html

Barry

[Barry Goyette]

A.J.

I'm definitely in your camp regarding the need for a camera that produces colors accurately from the start. Although I will say that adding color in post is problematic with DV due to the relatively high compression used in the codec. The Xl2's adjustments give us a good opportunity to get it closer in camera with pre-compression 12bit processing.

Barry

[A. J. deLange]

Barry,

Yes, you have a point there. The problem I have is that you are more or less stuck with whatever decision you make at the time of recording. I think you do have some latitude with respect to increasing/decreasing saturation (and adjusting gamma as well) but the 8 bit width is indeed limiting. What I'd like, of course, is 12 bits of R, 12 bits of G and 12 bits of blue right off the sensor (which is what I get from my still camera or scanner) to fiddle with WRT color temp, saturation etc., but I doubt I'll see that at a price I can afford in my lifetime.

A.J.

[Evan Fisher]

Barry,

Do you find it easier to tweak DV image colors in post if you shoot slightly under saturated and slightly lower contrast?

I ask this question because coming from a film background, we often take lo con ftg into color timing. And when I shoot digital stills, I shoot Raw format which gives me a low contrast, low saturation image (which is my backup/no longer negative) and I color correct from there achieving beautiful results.

I am also asking this because a lot of people mention the "washed out" image from the XL2 (I'm very happy with the camera btw). Maybe it was Canon's thought to default to a lower contrast/saturation image that may be easier to manipulate in color correction (just a thought).

[Barry Goyette]

Evan

Part of my motivation to test the color matrix and gain settings on the xl2 was to answer some of the comments that had been made regarding the xl2's saturation. In practice, the xl2s default settings look great to me on my HD set at home, although they do look a little flat when viewed on my computer (to be expected given the difference in color space and profiles), or on my production monitor (which is set to have a flat look--at default, the xl2 looks like everything else I display on it).

Definitely it's better in the film world to start with a lower contrast original, as there is more data available to the colorist, however in the DV world working with a highly compressed source,and 8 bits of data, My feeling is that getting it in camera...or at least getting close....is probably a good idea. I've done some experimenting in the past with color correcting in camera versus the same effect in post. In-camera is gives a cleaner final product. To answer your question directly..Is it easier to tweak DV colors by starting from a low contrast, under saturated point...probably easier to adjust contrast in this situation...not necessarily color...posterization is the most common artifact besides noise that comes out of post color and tonal correction...and posterization will typically become more of a problem as you expand the range of color and contrast. Regardless, I think it is best to start as close as you can to your final image (if you know what you want when your are shooting).

At the very least, these first tests show me that the xl2 color processing is operating very cleanly...with little of the artifacts I've seen from my gl2, and dvx associated with gain adjustments.

Finally, these samples show that the camera is capable of going "overboard" on color (not possible with the gl2)...which means you have all the range you need...that I think is a good thing. I personally would never shoot with the camera turned up this high...unless it was for a particular purpose, but it's nice to know that the camera has more gas than you need.

Barry

[Jacques Mersereau]

<<<-- Originally posted by Marty Hudzik : Well what I do know is this was the TV that they have sitting in there little theater like room and it was playing jurassic park off of a single DVD player. That looked fine. Not Hi Def good but at least as good as standard def on a standard def tv.-->>>


Right, but their are reasons that Jur. Park looked far better than the DV
footage.
(A sorta repeat from my eariler post)

1) shot in film so there is much more resolution to work with. That matters.
DV's macro blocks don't help. In fact, a guru from Sorenson told me that
one should do a "save as" uncompressed before encoding to mpeg for maxium
results when using DV.
Others do that and then add a .1 of a pixel blur to help break those
macro blocks up even more.

2) Hardware noise reduction and the best MPEG2 encoders are used
for the Hollywood big boys. That get rid of unwanted pixel information
and with film's larger color space, unwanted banding.

So, even though the final output is NTSC at 720x486, when you start
with 2Kx2K or 4Kx4K pixels before encoding (as opposed to 720x480 at 5:1
compression), and larger bit depth, the final mpeg 2 render
will always look much better coming from film (or HD) than DV.

[Marty Hudzik]

So this is going nowhere fast.

I never said I thought my XL2 should look as good or better than Jurassic Park on DVD. OF course that is going to look better.

However if I pop in JP at home on a standard def TV it looks very clean. Of course the characteristics of film are there because it was initially shot on film.

IF I watch my encoded XL2 dvd at home it looks really clean too. OF course it is clean but with the look of the XL2....not the look of Jurassic Park.

On a DVD player hooked to the HDTV at CC JP looked clean. My XL2 footage looked artifacted. There is definitely an issue with whatever upsampling they do. I don't know what it is but it is there. Can it be worked around? Eventually. BUt I was going off of what Barry had said.....something like "go to your local circuit city and watch your XL2 footage on a hi def tv and be prepared to be amazed!"....those aren;t his exact words but something like that.

To restate my opinion. My standard def pro monitor can see a lot of detail and clarity from my XL2 that none of the televisions I have tested can't. I was hoping that a hi def TV would be able to see this too. In my limited test it not only didn't see it but seriously degraded the image beyond what it looks like on SD.

I guess it is wrong to think that the XL2 is ever going to look as good as it does on my field monitor on any retail television under $10,000.00. IS that the case?

I admit one tv is not a good sample to make conclusions.....but on a day where I needed some validation as to why I spent the extra money on the XL2 over the DVX100A the HD didn't work in favor of the XL2. Granted the DVX didn't look any better but it didn't look any worse. And my personal preference is that I don't really care for a lot of the little quirks of the XL2 body and controls. But I keep coming back to it and seeing the incredible clarity on my production monitor. That makes me want to keep it and work out the kinks with the design that I don't care for.

But when I think that I am willing to live with what I view as shortcomings of the design for greater image quality, and then on most of my clients tv's/monitors you won't be able to see it....then I wonder why I invested this much and then deal with quirks when the end result is the same as if I used DVX.

[Marty Hudzik]

I have complained of lack of color.....you all know that. I played with the settings and came up with one that gives me the look I want in most areas. Here are a few frame grabs.

www.iciclestudios.com/images/gandi/horse.jpg
www.iciclestudios.com/images/gandi/cup2.jpg
www.iciclestudios.com/images/gandi/parts.jpg
www.iciclestudios.com/images/gandi/samples.jpg
www.iciclestudios.com/images/gandi/vinyl.jpg

they are jpg compressed......sorry. I need to keep my bandwidth down. Hopefully these will help to get the idea across that the XL2 can produce really nice color images. Viewing these on a tv is essential as they look a little washed out on my LCD laptop screen.

Let me know what you think......I resized them to 960x480 to retain the proper aspect ratio.

[Barry Goyette]

Marty--I was going to stay out of this one...but you just had to drag me back in...

First off, I never said make a DVD of your footage and take it down to Circuit City...I said take the Camera to Best Buy....I finished it off with, I think you'll be impressed (not my actual words either).

second. you've made a habit of reporting lots of flaws in your images, and then blaming it on the camera. A good scientist forms a hypothesis (the xl2 sucks...for instance..) and then sets out to disprove it....which in your own way you are doing....sort of...but a good scientist doesn't publish his findings until he is reasonably secure his findings are correct. You started out thinking your camera was broken because the color was too dull. Then you tweaked the color and said the viewfinder was bad because it didn't match the image on your monitor (even though you knew that you could adjust it)...then you said the image was bad because you you played a DVD of it on one TV at Circuit City. A good scientist would probably have played with it on a few sets, just to make sure the set wasn't the cause of the problem....The XL2 is an excellent professional product that allows you to set it up to fit your personal needs...it might be best to work out the kinks, and then post on any problems that you find.

Finally, you keep saying that you expected to be "blown away" by the camera, as compared to the DVX, and thus you are disappointed....indicating along the way that you were duped by comments made here on the forum. Well I can't remember anyone ever saying that the camera blows away anything. Any comments I've made have had a very measured tone relative to specific aspects of the camera, and I stand behind every one of them...except that you probably shouldn't view DV on the 55" plasma monitor at CC. Several users have been impressed by the resolution in 16:9, and you yourself say that you have been impressed with this aspect. Regardless, this is a DV camera, and thus it is very much limited by the DV spec. Perhaps you might be happier if your expectations were a little lower to start with.

I hope you don't think I'm slamming you here, I don't mean to. I just think weight of your critiques would be stronger if you got your data better in line before reporting it.

Barry