The Fifth Element Color Test

[lnx64]

A lot of people in the Home Theater business, love to show this one shot from the movie The Fifth Element, to show off color, because getting all the colors shown correctly is hard. Flesh tones vs the building in the backgrounds concrete color, difficult to balance. And admittedly, NTSC can't quite grasp it well. But I just had to show off.

The screenshot usually in question on home theater forums: http://www.projectorreviews.com/imag...loo1_large.jpg

And viola, my old Samsung presenting the very shot: http://img.photobucket.com/albums/v3...on/tvcolor.jpg

So yea, if anyone has this shot to show off, we should show off our old TV's showing this "color test" that home theater forums like to show.

Because sometimes I feel our old TV's can do it quite well, despite being RF. And in certain circumstances, like TV's older than mine, can get the red hair even better than any thing modern.

[Pete Deksnis]

Brandon

Love a HUE challenge! As soon as I read your post I cranked up the CT-100 for a half-hour warm-up. A small Asus Eee netbook is the reference image provider.

My results are definitely different from the 'reference' in these two areas: [1] the lip gloss on the actress is virtually nonexistent on the reference and [2] the hair is perhaps best described as light rust orange on the 'reference' and a darker reddish-orange on the CRT.

Note: sorry the photo attached is a bit overexposed.

The TV image you see at the left of the photo is on RCA's f38310, its first hi-def set (circa 2000), fed with component video from a Blu-ray disc; the netbook screen is in the middle; and on the right, hopefully the best-calibrated RCA CT-100 (1954) in operation today fed with composite video from a full-screen DVD. [the CT-100 is modified to accept composite video; audio is delivered to the set's front-end via a DRAKE VM2310 video modulator, thereby utilizing virtually all the CT-100's circuitry]

In conclusion, why does lip gloss, apparent on my sets, not show on the 'reference?' Why do the first and last RCA color sets both show virtually identical hair color although the 'reference' is different? And finally, why does the photo show the same hair color on the Eee 'reference' when my eyeballs see it much different as described above?

Me-thinks it points out the difficulty in evaluating color video system performance.

Pete

BACK ON LINE! Catch my CT-100 site: http://www.earlytelevision.org/Deksnis/home_page.html

[Phil Nelson]

This made me curious to try my RCA CTC-11.



All right, so it's not the best-calibrated CTC-11 in operation today

The screen on the right is my tired Sony KV-1311CR TV/monitor from the 1980s. It is getting video direct from the laptop and the CTC-11 is fed through a cheap RF modulator.

My auto-everything camera makes screens look too blue under incandescent lights, and of course everything is overexposed. Hard to judge colors from photos, as Pete noted.

I'm too lazy to haul this stuff out to the workshop for a "Pepsi challenge" with my CTC-7, but I fear it wouldn't compare well to the CTC-11, since I have never bothered to do any of the usual setup on it.

Phil Nelson
Phil's Old Radios
http://antiqueradio.org/index.html

P.S. Pete, it's great to see your website back online!

[lnx64]

Interesting results. Mine seems to have a tad too much green it seems.

[old_tv_nut]

Fun, but pretty meaningless, I'm afraid. No one knows what an NTSC (DVD) version really is supposed to look like except the studio guys who did the original color timing.

Even worse, this file is a picture taken with a still camera with unknown color response from the screen of a particular projector, and as such will be FAR from the original source and probably not represent the actual color of the projected image too accurately either.

Now, if you got a BluRay disc of this and displayed it on a calibrated HD set, you would see color reasonably close to what the colorist who did the BluRay transfer intended.

And, by the way, that probably is different from what was done in theaters.

[Eric H]

I have a Bluray of the movie, I should see how good of a picture I can get from my 2005 Sony Rear Projector.

[lnx64]

I was playing the bluray version, the remastered one that didn't have a sucky picture.

[Pete Deksnis]

Originally Posted by old_tv_nut

Fun, but pretty meaningless, I'm afraid.

Agreed.

My own standard for setting hue and color saturation is the color bar signal from a Digital Video Essentials DVD using their supplied color filters. After that, I let the video source take over.

Trouble though is the filters. [Well no, see below -- Pete] According to DVE, they are balanced for a D65 gray scale (X = .313, Y = .329) and I do my best to adjust CT-100/15GP22 gray to the 1953 NTSC-standard Illuminant C (X = .310, Y = .316).

I suspect that's why the green test never looks correct on a 15GP22-based set, but much better on the f38310.

Pete

[lnx64]

My TV was calibrated as close as it can to D65, but it's strange, we found that it likes to deviate from that calibration, depending what is on the screen. If it's a pure greyscale pattern, it passes. But if there's other stuff on the screen, it deviates.

Must be a tired tube. Oh well.

[old_tv_nut]

Originally Posted by Pete Deksnis

Agreed.

My own standard for setting hue and color saturation is the color bar signal from a Digital Video Essentials DVD using their supplied color filters. After that, I let the video source take over.

Trouble though is the filters. According to DVE, they are balanced for a D65 gray scale (X = .313, Y = .329) and I do my best to adjust CT-100/15GP22 gray to the 1953 NTSC-standard Illuminant C (X = .310, Y = .316).

I suspect that's why the green test never looks correct on a 15GP22-based set, but much better on the f38310.

Pete

Hmmm - I have to check out what DVE says. Doesn't make sense to me that the filters should bear any relation to the particular white point - they are just supposed to separate R, G, and B phosphors. As I think you pointed out once, they do not do this perfectly with a true NTSC blue, which has a lot more cyan wavelengths in it than the modern blue.

[lnx64]

Both my PVM and this old Samsung is calibrated with the DVE's color filters for NTSC.

Along with white balancing my professional equipment, I find it funny that they both still don't even look close. But both are D65 calibrated.

http://img.photobucket.com/albums/v3...comparison.jpg

[Pete Deksnis]

Originally Posted by old_tv_nut

Hmmm - I have to check out what DVE says. Doesn't make sense to me that the filters should bear any relation to the particular white point - they are just supposed to separate R, G, and B phosphors. As I think you pointed out once, they do not do this perfectly with a true NTSC blue, which has a lot more cyan wavelengths in it than the modern blue.

Yes thank you, I've been re railed and am on track. As the attached image shows, the magenta and blue bars show unexpected brightness through a green filter, while the red bar between them is properly dark. This is a shot of color bars on the face of a 15GP22 through the green DVE filter.

Those same magenta and blue bars show far less brightness through that same green filter on a 'modern' CRT in the RCA f38310, which appears to demonstrate that the original blue phosphor of the 15GP22 has a much greater green component than its modern counterpart.

[lnx64]

I did a test with the Disney logo, and I got to say, this looks REALLY nice, and I love the fact that no NTSC dot crawl showed up.

http://img.photobucket.com/albums/v3...n/_IGP7136.jpg