#Gamma
21 messages in this thread
Hi Greg,
Just got your book, and I must say, it's very impressive (good pic too! <g>).
Required reading for any animator, IMHO — I've got a question about gamma. If
I set my display gamma where the two grey squares match, and adjust everything
to give the best possible output there (on the monitor), then set the output
file to 2.2 (for the PAR) am I going to get the right output? I don't have a
good NTSC monitor, only a small TV/monitor to check the PAR output so I can't
really trust what I'm seeing.
John:
Glad you like the book (and the picture!)
You are using the same settings that we are, so I would venture to say that
your output is moving in the right direction! Pick up an SVHS monitor at your
local electronics superstore to hook up to the PAR to really see what you have
been creating!
Greg Pyros
Greg-
Sounds good but I'm still confused about one point — When I had the (computer)
monitor set for 2.2 gamma, the picture looked correct, but when I switched it
back to 1.25 (where the grey boxes say the gamma should be) the picture was too
dark (on the computer monitor). Output file gamma is still set to 2.2. My
question is, should I keep the computer monitor set at 1.25 gamma and use it as
the control monitor and trust the output file gamma to be correct (to the PAR)?
Am I going to get (on the PAR) what I am looking at on the (computer) monitor?
One more question…when I tried rendering with fields to the PAR, I had no
luck (as a matter of fact, when it didn't work, I switched back to rendering
frames, but the PAR distorted every frame until I rebooted). Is there something
special I need to know about this operation?
–John
You have a composite, NTSC monitor (Television set), and NTSC has built in
gamma correction, by nature. Therefore, you cannot trust the grey boxes to
give you correct gamma on your TV (unless you're using your computer's RGB
monitor, in which case the TV will be even MORE different). If you are using
your computer monitor, then calibrate the monitor gamma there to output images
as close as possible to what you see on a GOOD monitor or TV set.
You should at least rent a decent monitor for a day and get your system
calibrated. There's just no way to do that with a television set (is your TV
calibrated properly…) Mainly, what you want from monitor gamma, is to match
final output as closely as possible with what you see on your monitor.
Keep in mind that the bandwidth on computer monitors (and that they are
non-interlaced, for the most part) is so much higher than TV and video, that
you'll never get quite the same output.
John Tissavary (LUNA cie)
Hey John! Good to hear from you —
I'm not sure I'm getting my question worded properly — It is (has been) my
understanding that if I calibrate my computer monitor with the grey boxes
(which ends up at 1.25, by the way), this will give me a gamma-corrected
reference with which I should make all of my adjustments (lighting, etc.). If
I then set my output file gamma (when outputting to the PAR) to 2.2 (NTSC
level?) will I be getting the same picture?
If I am understanding you correctly, you're saying that I should be doing this
the other way around — by renting a monitor, calibrating it, (how?) then
adjusting the computer monitor to match —
Please excuse my density — I still don't have a complete grasp on this thing
(obviously! <bg>) I appreciate your help!
–John
The rule of thumb is to set your computer monitor gamma to where the grey boxes
match, and then do some tests to the PAR with the same scene (I've used a
rendered grey-scale) with various file output gamma between 1.8 and 2.2 (dark
to light). Some folks like the darker look 1.8 gives them, even though
technically, 2.2 is optimal NTSC gamma.
– G
Gary, Since you're obviously using the PAR and infamous for being picky about
video quality I was wondering if I could get your thoughts on its output. I was
told by someone here in Milwaukee the other day that the PAR's output is "soft"
and not nearly as good as going single frame to BetaSP. Since the PAR can do
component I thought it should be very close if not just as good. Any thoughts?
Dave Mennenoh
Obviously using the PAR? Me? I'm the last of the video snobs, David. When I
need to go to video, Jamie renders it up in the lab and we convert the frames
to Abekas format using the ASDG software and then send it off to PVR, downtown.
They either put it on D1 or (now) Digital Betacam.
– G
You just mentioned the PAR in a message to someone and I assumed that you were
using one. Never assume I guess… Have you seen the output from one?
Dave Mennenoh
Actually, no. Jamie's got one up in the lab, but it didn't come with a hard
disk, and things have been so busy that he hasn't had the chance to get one and
install it. I sense that he's on the verge, however… (the Siggraph tape is
coming up fast, and it'll be great to use for tests)
– G
Well Gary congrats for being a "video snob"! Actually I am interested in a
comparison of the PAR with component and other component output. Here we use a
LVR 3000 for output and then edit on M II tape, tryng to keep it all in
component form. Most of the stuff I see on the forum seems geared to the VHS
crowd using the PAR. One other question now, exactally how would I go about
finding if the LVR does gamma correction? I've been using a setting of 1.8 and
getting reasonable resultsusing a Targa +64 framebuffer.
Carl
Even though the PAR "does component", David, keep in mind that this is a
function of the frame buffer and scan converter, not the amount of JPEG
compression used, which is also important in determining the all-around quality
of a PAR-output animation. Your person in Milwaukee could very well be
reffering to artifacts of JPEG compression, bandwidth of the CAV video signal,
or perhaps both when he called it "soft". As the amount of compression used can
vary between using different hard disks, and the actual characteristics of the
animation's frames, blanket statements about the level of quality you can
expect to get from the PAR are a little unrealistic, I think. The best way to
really compare would be to do the *same animation* frame-by-frame to BetaSP,
and then to the PAR, CAV to BetaSP. Do some complicated, and some simple
frames. Make sure the PAR is operating at it's highest Block Limit setting at
all times. Then sit down and A/B the two versions. It may be readily apparent
just how much quality difference there is. If it isn't so apparent, you know
the PAR is doing a pretty good job! I haven't gotten a chance to do this,
David, so I just don't know how much of a difference you will find. I plan to
make this sort of comparison in the near future.
-Alan
Thanks Alan, but your reply didn't really give me anything to go on. I will
never (for a long time anyway) be able to do a comparison between BetaSP and
the PAR. I was hoping that someone using one could give me their impressions of
it vs. single frame. When you get the chance to do this, I'd really like to
hear your thoughts on it. Oh, on the game companies paying for well done
.fli's. I believe you're right about that and I'm sure MartinF could attest to
that also, but it was just a thought about the student version. I really think
there shouldn't be a student version. Schools should just buy the full blown
version and keep good security measures. As MichelleB stated about the school
that has all the dongles inside the cases. That makes perfect sense to me – the
cases can be locked and the CPU's can be locked to the tables they're on.
Dave Mennenoh
David, if I get the chance to make this comparison, I promise to share my
findings (or opinions, as it were) with yourself and the forum in general.
There is a chance that I will be able to do this soon. We'll see.
As 80-90% of all my work in 1993 was game-related (AEGIS, Deadzone, etc.) I can
tell you from personal experience that .FLx files are VERY popular with the
game companies.
As to the idea of stricter security standards at schools, it is a nice idea,
but impossible due to the fact that Adesk sells educational (full-blown) copies
to LOTS of individuals (students, teachers, etc.) as well as the institutions
themselves. So the best school-security in the world could never keep the
educational software out of "public" hands.
-Alan
Thanks for your help — That makes sense.
When you said you were calibrating your monitor you didn't specify your
computer monitor. For instance, I never render to my computer monitor, it
doesn't do me any good since my stuff goes out to videotape and the two just
look so different.
But if you're calibrating your computer monitor, then yes… the gray boxes
should match (as much as possible).
Your file output gamma should be what looks good. This is where you'll really
benefit from a video reference monitor. If you use 2.2 as gamma, you'll get a
fairly "hot" (bright, lower contrast) look, 1.8 will be less hot, more neutral,
1.5 will be darkish and contrasty. Think of gamma as a curve of electricity –
the higher the value the more juice you're cranking into the picture (mostly
the mid-range). I use output gamma 1.8 most of the time, but others have
different preferences, and I'll crank that value up if I want a "hotter"
picture.
John Tissavary (LUNA cie)
Thanks, John — That clears up a lot!
Joseph, PMJI, but I haven't noticed anyone say anything about this yet. But,
you need to put the editor into a 24bit graphics mode to adjust the gamma
correctly. Leaving it at the 256 color mode will not give you the proper gamma.
I had the exact same thing happen to me, but when I put it to 24bit the gamma
looked about perfect at 1.8.
Dave Mennenoh
Thanks David (or is it Dave?) — I hadn't considered that, and it makes a lot
of sense
–John
Glad to help Joseph (or is it Joe?) <g>
Dave Mennenoh
John, actually — <bg>
–John Wilson