[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
green light still shining
- To: "TIG list" <telecine at alegria.com>
- Subject: green light still shining
- From: "Mikael Reichel" <mikael.reichel at culinaire.se>
- Date: Tue, 24 Mar 1998 15:56:58 +0100
- Resent-Date: Tue, 24 Mar 1998 06:59:37 -0800
- Resent-From: telecine at alegria.com
- Resent-Message-ID: <"Ip9Kr.A.MEE.in8F1" at sun>
- Resent-Sender: telecine-request at alegria.com
- Resent-To: multiple recipients of <telecine at alegria.com>
I thought the referee had the Spirit hand up long time ago...
Replying to Paul Grace
>I think you are missing a point.....
I am sure I am missing many points here, not even the TIG in all its
capacity can relate it all. But letīs see if I can cast some light on some
points of interest after all.
>>In the Spirit case it is not a matter of "getting away" with
>>something more than actually understanding and making best use of how we
>>humans actually operate.
Correct, but it is obviously not all. Sure it is a compromise, then all
designs are. In this case I would guess quite a conscious one based on cost
against net performance gain at the output of the machine. What bothers me a
bit is how all the critics on the Spirit hit on this particular fact
strictly from a "number point of view". Are you not supposed to first and
foremost put specs aside and judge image quality?
>>The lower resolution red and blue are then
>>matrixed with a half resolution luminance signal to get the green.
I did not say that, wrong quote. Also it is not correct (unless scanner was
changed at some point).
The Spirit has three independent 960pix RGB photosensors in addition to the
1920pix luma sensor. Each of these RGB arrays have a dichroic filter in
front of it. Here comes one of the finer points with the luma sensor - it
has no chroma separation filter. In a traditional TK design, luma would have
to be derived from matrixing the RGB output of CCD/PMTīs and that is after
having passed through (non ideal) optical filters.
Do I need say more?
>In post production we strive for equal wide bandwidth to allow for
precision keying in composites.
>This is why high end visual FX houses use 4:4:4 in Ursa and Flame/Inferno."
1. Unfortunately, 4:4:4 is not a quality statement (nor is 4:2:2, 4:1:1 or
4:2:0) even though some may still believe so. It says nothing of what runs
through it. Equal and high bandwidth may at first glance sound like a major
requirement but as you know the wider the bandwidth the more noise and if
blue already had a problem....
Everone who is into HD scanning know what care and time it takes to generate
clean transfers. Noise, dirt and scratches are all there but in som much
greater detail using so much more data...
2. Are you saying that the Spirit is no good for high end visual FX work
whereas the URSA is?
3. 4:4:4? There are other alternatives to 4:4:4 that do not offer its
restrictions when transferring into computer domain. I would prefer to use a
pure computer format to download an interlace-free progressively scanned
image. It kind of makes computer FX work more fun unless of course it is
stills we are discussing here.
>Film work is even more demanding in quality and any
datacine/telecine/scanner must offer a wide band response to >faithfully
capture all the data we need for VFX along with other increased technical
Paul this sounds like sale lit talk (I happen to have written some in a
previous life). These demands are to begin with, interlace, square pixels
and higher frame rates if I can wish. But then I am only a theoretician.
Think I said this recently - resolution and noise are at opposing ends of
the stick. Clever engineering can make the stick shorter. Tradition
(adherence to old technology for the sake of comfort) will make it longer.
(new technology is not neccessarily better...)
>Equating this to the eye's perception is misleading as that has nothing to
do with the image
>manipulation we need to do to the images prior to any transmission.
Hmm must have got this wrong then, I thought FX WAS all about the eyes (and
In my mind "perception" is one of several very important considerations to
take into account these days. Perception also made up the very clever TV
system that has lasted for some 40 years now. MPEG could not exist without
perceptual visual weighting. Obviously noise free, detail rich transfers are
important for the technical manipulation as well for the eye but the fact
remains that we have so many degrading factors in our television system(s)
today that we have been led to believe the more "horsepower" is the only way
out. However, even MPEG with all its tricks, will look better if the source
material is "sampled" higher with matching detail content. So there is no
argument with me on that one.
If you want to see that less is more you should look at the progressive
480/720 demos at NAB. Thats HD for your eyes but maybe not for your vanity.
>Paul "anarchy in the UK " Grace
Mike "interlace sucks" Reichel
"I do not distinguish by the eye but by the mind which is the proper judge
"The future of television is the Internet"
Phone/fax/voicemail +46-(0)8-541 31500 Mobile +46(0)708-541792
Adress; Ytterbystrand 6, S-185 94 Vaxholm, SWEDEN
Thanks to Peter Stansfield for supporting the TIG in 1998..
No product marketing allowed on the main TIG. Contact rob at alegria.com
944 subscribers in 36 countries on Tue Mar 24 06:57:30 PST 1998
complete information on the TIG website http://www.alegria.com/tig3/