Fwd: [Tig] LUTs for Video File Codecs - carrots

Tim Montgomery tim.montgomery
Wed May 3 18:24:25 BST 2006


--
Thanks to oktobor for supporting the TIG
--
It seems that I'm incorrect in referring to this as a LUT.  But a "LUT type structure"   see below.


-----Original Message-----
From: Stan Chayka
Sent: Wednesday, May 03, 2006 10:53 AM
To: 'tig-bounces at tig.colorist.org'
Subject: RE: [Tig] LUTs for Video File Codecs - carrots


It is interesting to observe how the cycle of life (technology wise) repeats
itself.

The original thought that has brought this particular issue to debate is
whether a better correction job could be done if the colorist could see his
work with the compression degradation simulated in a LUT type structure.

The choice of utilizing LUTs is not the issue and should not become the
focus.

The issue as I understand it is 'Should images be adjusted to a particular
target viewing technology'. This choice has been with us for many years and
most of us are old enough or so to remember the transition from Subcarrier
Encoded images to Digital Component Images was very painful and required
redundant monitoring to be sure that both formats were 'LEGAL' and still
maximized.

Since the Composite signal no longer rules, this concern has gone away, but
the idea of creating a particular 'look' to satisfy the delivery is not new.

Now to comment on that particular idea of Optimizing the image to the
delivery;

I agree that this is a desirable way to produce your best results when the
total process is within your control.

A few comments about where the pitfalls are;

 This means that the job you do is only best viewed in the context of the
delivery system it was targeted for.

Dynamics such as noise reduction, image sharpness, dirt concealment, bit
depth on colorspace and even the very content of the image may be optimized
for one and appear very in-appropriate for other viewing systems.

 The job would  need to be labeled as 'For Compression ONLY' or for 'DLP
presentation ONLY' or for NTSC only. I'm sure this would not be universally
acceptable in practice yet it is something of a requirement to follow thru
on 'target specific' material.

I acknowledge that this does actually occur in certain situations such as
correcting in a Digital Intermediate context for release back on film and
I'm sure there are plenty more examples.

If your work is targeted for diminishment by a CODEC, and you have some
certainly that the intermediate products will never be used in any other
context, by all means setup a dual monitoring system with a real codec
signal path display to be sure of your efforts.

If however, your end product is going to be judged by others in various
viewing venues, be precise and live to the standards of the best viewing
system available. Everything else IS a step down from there and is not your
responsibility to make them selectively better than they actually are.

Would you expect sound to be mixed to the performance of some piezio
transducer in a cell phone? or is the correct approach to mix to the best
sound system available?

Unless the sound was only used in such cases as ring tones and such.

These are my comments and only represent what I have observed and been part
of for the last 30 or so years.

Stan Chayka




-----Original Message-----
From: tig-bounces at tig.colorist.org
[mailto:tig-bounces at tig.colorist.org]On Behalf Of Tim Montgomery
Sent: Wednesday, May 03, 2006 10:10 AM
To: tig at tig.colorist.org; woz
Subject: Re: [Tig] LUTs for Video File Codecs - carrots


--
Thanks to oktobor for supporting the TIG
--
I guess the principals of encoding are the debate now.  I'm more concerned
with achieving quality no matter what the deliverable is (the deliverable is
a client request and not something we can change).  If LUTs are being used
to ensure a consistent grade between projected and emitted light, why can't
there be a simulation for encoded signals.  I guess I am less concerned with
the encoding operation and more concerned with it's effects on the image and
how to compensate for it (it's seems to be agreed upon that it does have an
effect).

On Wednesday, May 3, 2006 10:55 AM, woz <warrenl at gmail.com> wrote:
>Grading for codec degradation is rather like putting the cart before the
>carrot (I made that last part up because I couldnt quickly think of a
>suitable metaphor).
>The point is that while you should be aware of the effects of  noise and
>detail etc on a particular codec, if your accepted material becomes
>unacceptable because your choice of codec then you should look elsewhwere
>for a suitable codec.
>I will agree that relatively trivial changes such as noise reduction might
>help a little but overall you grade for the look, the codec is merely a
cart
>in which to safely carry the carrot (or onion if you prefer).
>However as a general rule, (if you want one) with the type of
>temporal/spatial codecs now in use, the less detail and change of detail
you
>have in your pictures, and the lower the resolution, the better they will
>code within a given bitrate.
>
>woz
>itv/3sixtymedia
>----- Original Message -----
>From: "Tim Montgomery" <tim.montgomery at finishedit.com>
>To: <tig at colorist.org>; <tig-bounces at tig.colorist.org>
>Sent: Tuesday, May 02, 2006 9:39 PM
>Subject: [Tig] LUTs for Video File Codecs
>
>
>--
>Thanks to oktobor for supporting the TIG
>--
>Is there such a thing as a LUT that simulates Codecs like Quicktime and
>Windows Media?  Is there a general rule when grading for codec
>degredation?
>
>
>
>_______________________________________________
>Tig list - http://tig.oktobor.com/mailman/listinfo/tig
>TIG wiki: http://tig.colorist.org/wiki
>
>



_______________________________________________
Tig list - http://tig.oktobor.com/mailman/listinfo/tig
TIG wiki: http://tig.colorist.org/wiki






_______________________________________________
Tig list - http://tig.oktobor.com/mailman/listinfo/tig
TIG wiki: http://tig.colorist.org/wiki




More information about the Tig mailing list