HOW-TO: Calibrating Display to Match HTPC Output - Page 4 - AVS Forum
Forum Jump: 
Reply
 
Thread Tools
post #91 of 486 Old 03-31-2005, 04:11 PM - Thread Starter
AVS Special Member
 
cyberbri's Avatar
 
Join Date: May 2004
Location: San Jose, CA
Posts: 8,095
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 52
Yes, it's pretty weird. Like I put in the most recent section, if I pushed contrast up so the whitest white my display produces comes down to 235, the upper grey range turns reddish-pink and orange at different values, similar to what the highlights look like on the pic I just attached.


I'll have to go back and look at some different things. On my monitor here at work, just for kicks I opened up the user settings. It turns out my monitor is default at 9300K temp. When I change it to 6500K temp ("Image Management" setting), the overall brightness (amount of light) drops and there is a slight yellow tint to the screen. My display at home probably "seems" like it's somewhere in between the 6500K and 9300K I'm seeing right now on this monitor, with it currently closer to the 6500K temp's overall screen brightness (amount of light) and getting up towards how the 9300K looks if I crank up the contrast in the service menu.

Just when I thought I was done trying to calibrate for a while... (Even now, I still can't get my r/g/b balance perfect, so everything looks great until I pop in Finding Nemo and see slight banding...) I wonder if the only way I'd really get it "perfect" would be to hire a pro... (can HTPCs and DLP/digital displays even show Finding Nemo without banding?)
LL
cyberbri is offline  
Sponsored Links
Advertisement
 
post #92 of 486 Old 03-31-2005, 05:51 PM
 
ChrisWiggles's Avatar
 
Join Date: Nov 2002
Location: Seattle
Posts: 20,730
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 14
It is not abnormal for many digital displays to colorshift as you move the white level control to maximum. You will likely see either hard clipping of whites electronically, or colorshifting. This is why I mentioned in my guide that if your display colorshifts instead of clips, it's probably best to shoot for the max white on your display to line up near 254 to minimize any visible colorshifting on highlight details. While clipping can be objectionable on data that may be above nominal reference white, I find this less objectionable than colorshifting that may occur.
ChrisWiggles is offline  
post #93 of 486 Old 03-31-2005, 06:24 PM
3no
Senior Member
 
3no's Avatar
 
Join Date: Aug 2004
Location: Atlanta
Posts: 375
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:


Originally posted by cyberbri
Have you found with your 720p projector that you could calibrate contrast to 235 for peak white?

Ummm, I clearly need to be careful in this thread on how I respond to that. If you mean by “peak white” max brightness from the projector, then the short answer is “yes”, I could set the contrast control on my projector to project it's max white level at 235 reference white (and every higher white), but my observation is that really bright scenes look better with white differentiation somewhat higher than 235. I've played with this quite a bit trying to find the best balance between increasing perceived contrast and not crushing either blacks or whites.

There seems to be an optimum contrast setting on my projector (Hitachi PJ-TX100 LCD) where I get the 255 background to it's brightest, have the most contrast between high whites (230-235) and the 255 background, and extend white differentiation somewhat above 235 (to 243). That seems to be the best compromise for PQ as well. Black level calibration was fairly straightforward, setting brightness to just barely distinguish the 17 bar from the 0 background to the point where it might be imagination. As you would expect, I had to cycle between black and white as the contrast and brightness controls do have some effect on each other.

I have used both DVE and the Philips Pattern Generator previously, but I created these patterns so I could have better control over finding my optimum settings by using alternating bands/background and expanding range at the ends of the gray spectrum.

Quote:


On my DLP TV, if I try to get the brightest white the display is "capable" of at 235, it messes with other values below that (high grey turns pink/orange)

The PJ-TX100 has a factory mode setup function that allows gray "color" to be set separately for 31 different intensity bands to correct for that effect.
3no is offline  
post #94 of 486 Old 03-31-2005, 06:27 PM
 
ChrisWiggles's Avatar
 
Join Date: Nov 2002
Location: Seattle
Posts: 20,730
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 14
3no, thanks for the patterns, I just want to point out that one does need to understand well what the video overlay's are doing with the levels. I hesitate to recommend using still images to calibrate unless you are very clear and understand specifically what the video overlay is doing, namely by observing video patterns that are run through those overlays and processing rather than images. If you do know well and understand what your video is doing, then by all means use still images to assist you for specific tweaking of calibration.
ChrisWiggles is offline  
post #95 of 486 Old 03-31-2005, 07:19 PM
AVS Special Member
 
maxleung's Avatar
 
Join Date: Mar 2003
Location: Calgary, Alberta, Canada
Posts: 3,495
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 13
3no, thanks too for the patterns - it is really cool seeing the white patterns on my CRT monitor - the difference between 235 and 255 is very very pronounced (gamma on my monitor is roughly 2.6). I can even see a clear difference between 253 and 255! The Contrast on my monitor is already at the maximum too.

This will be very useful with my DLP projector...

I would love to have an mpeg2 version of this image, but I'm not certain how the RGB image would convert to YCrCb (or is it YPrPb?)
maxleung is offline  
post #96 of 486 Old 04-01-2005, 03:51 AM
3no
Senior Member
 
3no's Avatar
 
Join Date: Aug 2004
Location: Atlanta
Posts: 375
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:


Originally posted by ChrisWiggles
3no, thanks for the patterns, I just want to point out that one does need to understand well what the video overlay's are doing with the levels. I hesitate to recommend using still images to calibrate unless you are very clear and understand specifically what the video overlay is doing, namely by observing video patterns that are run through those overlays and processing rather than images. If you do know well and understand what your video is doing, then by all means use still images to assist you for specific tweaking of calibration.

I certainly make no claim to knowing the internals well, the reason I went after a still image in the first place was to allow calibration with only the projector and video card/drivers. I've kept the video card settings at neutral and adjusted only the projector at this point. To your point, and maxleung's request, I'll make an mpeg2 version this evening. However... now I have three adjustment points (TheaterTek, video driver, projector). Instinct says to leave the latter 2 alone, as adjusted with the still image, and adjust TT as it is the new element in the chain. Advice?
3no is offline  
post #97 of 486 Old 04-01-2005, 08:48 AM
AVS Special Member
 
maxleung's Avatar
 
Join Date: Mar 2003
Location: Calgary, Alberta, Canada
Posts: 3,495
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 13
I calibrate TT by taking a screenshot and compare the RGB values to the original frame as captured by a reference mpeg2 decoder like DGIndex. Adjusting the video controls in TheaterTek (which uses the VMR9 procamp controls) actually changes the RGB values of the image - it is extremely coarse as the controls work in the 8 bit domain.
maxleung is offline  
post #98 of 486 Old 04-01-2005, 10:48 PM
3no
Senior Member
 
3no's Avatar
 
Join Date: Aug 2004
Location: Atlanta
Posts: 375
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:


Originally posted by ChrisWiggles
I hesitate to recommend using still images to calibrate unless you are very clear and understand specifically what the video overlay is doing, namely by observing video patterns that are run through those overlays and processing rather than images.

Quote:


Originally posted by maxleung
I would love to have an mpeg2 version of this image, but I'm not certain how the RGB image would convert to YCrCb (or is it YPrPb?)

Per ChrisWiggles' advice and maxleung's request, attached is an mpeg2 version of the black and white reference calibration screens. To keep the file size down the video is actually on a one second loop, so you may see a slight flicker at that rate.

(note: file removed temporarily for issue resolution)
3no is offline  
post #99 of 486 Old 04-01-2005, 10:54 PM
3no
Senior Member
 
3no's Avatar
 
Join Date: Aug 2004
Location: Atlanta
Posts: 375
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
The more I play with grayscale calibration, the more I'm finding that while the setup I describe several posts back (where the projector essentially expands the input range 16 --> 243 to it’s blackest black --> whitest white) looks great on material of "average" brightness, material that is very dark with low contrast (e.g. 24 Season 3) or very bright with high contrast (e.g. Kill Bill vol 2 B&W wedding chapel massacre) looks terrible with the same setup.

[edit: the following problem was resolved as a DGIndex YUV -> RGB setup issue, discussed in a later post]
I took maxleung's approach and snapped bmp’s of several frames from the DVDs (via DGIndex from ripped vob), and inspected them in Photoshop. Imagine my surprise. Tons of content below 16,16,16 in the dark 24 scenes, and at both extremes in KB2 (as low as 0,0,0 and as high as 255,255,255). So what am I doing wrong? Does this material really contain such extreme excursions below reference black and above reference white? And aren’t 0 and 255 totally illegal for video material? Here’s a snap from KB2 and one from 24.
3no is offline  
post #100 of 486 Old 04-02-2005, 01:25 AM
 
ChrisWiggles's Avatar
 
Join Date: Nov 2002
Location: Seattle
Posts: 20,730
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 14
What are you using for video rendering in your PC?
ChrisWiggles is offline  
post #101 of 486 Old 04-02-2005, 09:05 AM
3no
Senior Member
 
3no's Avatar
 
Join Date: Aug 2004
Location: Atlanta
Posts: 375
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:


Originally posted by ChrisWiggles
What are you using for video rendering in your PC?

Premier 6.5/Matrox RT.X100

I had/have at least two problems Chris.

1) Checking video levels: I am using DGIndex to capture a frame as a bmp, then check levels with Photoshop. Turns out I had the YUV -> RGB option in DGIndex set to PC Scale. Changed it to TV Scale and the levels in those dark and bright/contrasty scenes now all read within 16-235, with a very few spots slightly higher than 235. Verified accurate with a frame grab from DVE Title 12 Chapter 14 using the same process. Solved.

2) Rendering video levels: Now that I can accurately check levels in the mpeg2 files, I checked the Premier/Matrox output for my level calibrator video and sure enough, the renderer is doing a colorspace conversion to squeeze my levels into the NTSC values. I'll work on fixing that when I have some time later today, but if you or anyone knows Premier/Matrox I'd appreciate a hint (or even an educated guess).

I'll repost the mpeg2 video after this is solved.
3no is offline  
post #102 of 486 Old 04-02-2005, 03:33 PM
AVS Special Member
 
maxleung's Avatar
 
Join Date: Mar 2003
Location: Calgary, Alberta, Canada
Posts: 3,495
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 13
Thanks for your efforts 3no!

BTW, it annoys me to no end that DGIndex defaults to PC levels. Well, at least it is switchable to TV (aka Video) levels.

I'm thinking we at the AVSForum should make our own calibration disc - I really like your calibration screens 3no. I wonder if we can use AVISynth and a (free?) mpeg2 encoder, and master our own super-duper AVS calibration discs. But, I don't know if there are any paint programs that will allow us to create images in the YUV2 colorspace...sure, we can make RGB screens and then convert them, but there may be color resolution problems that we haven't forseen yet.

Oh well, that's just my brainstorm of the day...
maxleung is offline  
post #103 of 486 Old 04-04-2005, 07:20 PM
3no
Senior Member
 
3no's Avatar
 
Join Date: Aug 2004
Location: Atlanta
Posts: 375
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
maxleung, chriswiggles-
Thanks for your advice and comments.

Here's a video version of the black/white reference calibration screens. I switched renderers to TMPGEnc which can output video levels outside the 16-235 range. I'll work later on improving edge definition of the bands, and maybe other features (moving bands?).

Comments/suggestions welcome from everyone.

 

bw_cal.zip 280.435546875k . file
3no is offline  
post #104 of 486 Old 04-12-2005, 08:34 PM
Senior Member
 
yipchunyu's Avatar
 
Join Date: Jan 2002
Posts: 244
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
learn from this page, VMR9 is better than overlay and we should recalibrate the display for better image. However, I don't really understand the procedure (sorry for my poor understanding of English).
Does it means that we use some pattern e.g. 3no's listed in the previous post and re-calibrate the display and make it only display 16~235 range and make it 0~15 same as 16 and 236~255 same as 235?
yipchunyu is offline  
post #105 of 486 Old 04-12-2005, 08:35 PM
Senior Member
 
yipchunyu's Avatar
 
Join Date: Jan 2002
Posts: 244
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
learn from this page, VMR9 is better than overlay and we should recalibrate the display for better image. However, I don't really understand the procedure (sorry for my poor understanding of English).
Does it means that we use some pattern e.g. 3no's listed in the previous post and re-calibrate the display and make it only display 16~235 range and make it 0~15 same as 16 and 236~255 same as 235?
Besides, i found that VMR9 is not as sharp as overlay, is it true? how to compensate this as I use AE700 for display and the soft screen tech already make the image not as sharp as others.
p.s. sorry for my poor english
yipchunyu is offline  
post #106 of 486 Old 04-13-2005, 02:45 PM
AVS Addicted Member
 
stanger89's Avatar
 
Join Date: Nov 2002
Location: Marion, IA
Posts: 17,310
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 59 Post(s)
Liked: 115
Basically you get a test/calibration disk (like Digital Video Essentails, or the THX optimizer will probably work, it's on a number of "THX" disks). It will have a pattern that will have a video black background (ie 16) and then a below black area, what you do is adjust the brightness so the below black area matches the background.

Contrast is a little more tricky, most of the procedures were defined for CRTs which have no hard maximum white. Basically you can calibrate so that 235 produces the maximum white your PJ will do or, 255. You'll have to play and see what you like best yourself. What I would probably do is calibrate so that a little bit beyond reference white (235) is visible.

See what an anamorphoscopic lens can do, see movies the way they were meant to be seen
stanger89 is offline  
post #107 of 486 Old 04-13-2005, 02:53 PM
VBB
Advanced Member
 
VBB's Avatar
 
Join Date: May 2003
Location: Long Beach, CA
Posts: 651
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 21
Yipchunyu,

I'm surprised cyberbri hasn't been all over your post yet . What you need to do is very simple. Play 3no's video in your preferred player, making sure that the player's color controls are set to neutral. Then, using the black reference part of the video, adjust your projector's brightness so that 17 is the last shade of grey you can distinguish from the black background. That way, you will make sure that studio level reference black (16) is the darkest black your Panny can project. As for white, you won't be able to calibrate it so that 235 is the brightest white your projector is able to produce, but you don't really need to either. I recommend lowering your contrast down to the point where you don't notice any colorshifting, which seems to be an issue with most LCD projectors. Hope that helps. Everybody else, feel free to correct me if necessary .

Henry
Sharp LC-70LE857U
Onkyo HT-RC360
Mede8er MED600X3D

DarbeeVision Darblet DVP-5000

Synology DS1813+

X-Rite i1Display Pro

VBB is offline  
post #108 of 486 Old 04-13-2005, 03:06 PM - Thread Starter
AVS Special Member
 
cyberbri's Avatar
 
Join Date: May 2004
Location: San Jose, CA
Posts: 8,095
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 52
yipchunyu,

The calibration part of your question has already been answered.

For VMR9 softness... What do you use to play DVDs, what codecs do you use, etc.? Are you using ffdshow for resizing, etc.?

Different video codecs produce different levels of picture quality, so changing to a better video codec may improve your picture.

FFDSHOW can be used to resize and sharpen (and it has a lot of other filters, mostly unnecessary for simple DVD watching). If you use Zoom Player, ffdshow can be added as an Extra Filter in the DVD Options. Inside ffdshow, you can select different resize values (2x DVD resolution, 1920x1080, etc.), and can use Luma Sharpen at the same time as the resize, or use a separate Sharpen filter (Unsharp Mask, XSharpen, MSharpen, etc.)


Hope that helps...
cyberbri is offline  
post #109 of 486 Old 04-13-2005, 06:20 PM
Senior Member
 
yipchunyu's Avatar
 
Join Date: Jan 2002
Posts: 244
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
VBB and stanger89,
Regarding the questions I posted. I used 3no's files to recalibrate the pj as u suggested. However, haven't time to check some video clips with it. Hope it works well as I'm afraid the dynamic iris feature of this pj may affect the result.

Cyberbri,
yes i use ffdshow for resizing, i also apply some other filters as well. I use dscaler as the video renderer. When I said VMR9 is softness is compared to overlay. (with all other settings are the same). I just want to know is it the basic nature of VMR9?

Anyway, thx you all
yipchunyu is offline  
post #110 of 486 Old 04-14-2005, 03:00 PM
VBB
Advanced Member
 
VBB's Avatar
 
Join Date: May 2003
Location: Long Beach, CA
Posts: 651
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 21
Make sure to set the dynamic iris to auto when you calibrate. From my experience with the Sony HS51, the Auto Iris function needs to be on to properly set brightness and contrast.

Regarding VMR9, yes, that is the nature of it. Just compensate by sharpening slightly more in ffdshow. Like cyberbri, I recommend using Lanczos 2 with a Luma sharpen of .30 and no Chroma sharpening.

Henry
Sharp LC-70LE857U
Onkyo HT-RC360
Mede8er MED600X3D

DarbeeVision Darblet DVP-5000

Synology DS1813+

X-Rite i1Display Pro

VBB is offline  
post #111 of 486 Old 04-14-2005, 03:32 PM
AVS Addicted Member
 
stanger89's Avatar
 
Join Date: Nov 2002
Location: Marion, IA
Posts: 17,310
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 59 Post(s)
Liked: 115
Quote:


Originally posted by VBB
Regarding VMR9, yes, that is the nature of it. Just compensate by sharpening slightly more in ffdshow. Like cyberbri, I recommend using Lanczos 2 with a Luma sharpen of .30 and no Chroma sharpening.

FWIW, (semantics here) VMR9 is actaully more faithful in a way, Overlay tended to be oversharpened in most drivers so the move to overlay results in a (comparatively) softer image.

See what an anamorphoscopic lens can do, see movies the way they were meant to be seen
stanger89 is offline  
post #112 of 486 Old 04-14-2005, 03:53 PM - Thread Starter
AVS Special Member
 
cyberbri's Avatar
 
Join Date: May 2004
Location: San Jose, CA
Posts: 8,095
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 52
Quote:


Originally posted by VBB

Regarding VMR9, yes, that is the nature of it. Just compensate by sharpening slightly more in ffdshow. Like cyberbri, I recommend using Lanczos 2 with a Luma sharpen of .30 and no Chroma sharpening.

I think the amount of sharpening needed to get a certain result may depend on the resize value (may need more or less sharpening with higher/lower values). For me, at 1440x1200 (2x horiz + 2.5x vert resize), I don't notice a whole lot of difference between .30 and .50 or so, although I can notice a difference between 0 and .30 (this is adjusting on the fly, not a/b flipping screenshots). At that resize, once you get up above .60 I see more of the effect, but I choose to keep the effect slighter so I don't accentuate ringing, etc. too much, and I think a lot of stuff is sharp enough as it is without needing too much extra help to get the desired effect.

I have heard that Theater Tek's VMR9 Renderless is "sharper" / "less soft" than ZP's, so it may also be partly a function, to a certain extent, of the player's implementation of the VMR9 renderer.

FWIW
cyberbri is offline  
post #113 of 486 Old 04-15-2005, 02:42 PM
AVS Special Member
 
maxleung's Avatar
 
Join Date: Mar 2003
Location: Calgary, Alberta, Canada
Posts: 3,495
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 13
Sharpening can be done by programming the pixel shaders when using VMR9 - I am pretty sure this is why TheaterTek looks sharper than most VMR9 players. Unfortunately, I don't feel like running a D3D sniffer tool to verify this - I'm sure Andy would be pissed if somebody did this.

That reminds me, I really need to play 3no's file...
maxleung is offline  
post #114 of 486 Old 04-15-2005, 03:54 PM
AVS Special Member
 
takisot's Avatar
 
Join Date: Jan 2002
Location: Athens, Greece
Posts: 1,101
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Cyberbi, Chris,Max and all other fellow Avsforum members thank you for the knowledge you provide gratis in this forum!
After having read this thread and other threads linked to this one (like Chris Wiggles guide) I have to post a question that is bugging me:
My HTPC is connected to my BenQ 8700 Dlp via dvi and I run a Zoom/dscaler/ffdshow VMR9 windowless configuration.
I used DVE to set Brightness/contrast following the standard procedure of lowering the pj brightness until the BTB bar disappears. If I understood correct , by doing that I have set my black level at digital 7 ( where the bar is encoded on DVE) and not at digital 16 in order to be able to see the BTB info I SHOULD see according to the DVD producers that I would have seen on a CRT due to the float phenomenon. Am I right so far?
Assuming that I am right, I then lowered my brightness a bit more until the dithering of DLP micro-mirrors stopped, as It has been suggested somewhere else in this forum.
So my question is: Should I do this? If yes, after stopping the DLP micro-mirrors dithering, is my black level BELOW digital 7 so now I am seeing more BTB info than the one I was supposed to see? Or am I getting confused here?
Thanks in advance

A Never Ending Quest Τowards Visual Nirvana..
takisot is offline  
post #115 of 486 Old 04-15-2005, 08:26 PM
 
ChrisWiggles's Avatar
 
Join Date: Nov 2002
Location: Seattle
Posts: 20,730
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 14
Im confused as to where the digital 7 came from.

For DLP, I would shoot for having a black frame(digital 16) be dither-free because the dither can be irritating. This is no hard and fast and some will recommend bumping the brightness up one or two clicks to preserve just a touch below 16 for some detail that might ostensibly fall below 16 but be visible on a CRT due to a CRT's different behavior. The exact setting is a tad subjective here. The easy advice I could give would be to lower the black level until black(digital 16) elements in a pattern no longer produce any dithering. You may settle a couple clicks above this subjectively by choice.

What I want to make clear though, is that you *shouldnt* clearly see bars that are below black. They should be largely not visible. On a dlp this is something akin to clipping them off at the display. (others, don't argue about BTB and such here, it's not synonomous with clipping because the information can be used/processed and affect the display for elements that are above black and clearly visible).

I hope this helps, I guess I'm still confused where the digital 7 came from that's out of nowhere?
ChrisWiggles is offline  
post #116 of 486 Old 04-16-2005, 02:47 AM
AVS Special Member
 
takisot's Avatar
 
Join Date: Jan 2002
Location: Athens, Greece
Posts: 1,101
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Hi Chris.
Until today I was calibrating brightness just as you advise.
But then I read the following post by Bob Pariseau which mentions Digital 7 as the point were the BTB bar is encoded on DVD, that got me thinking:
For your convenience I quote the full Bob Pariseau's post:


QUOTE:
Bob Pariseau
AVS Special Member

Registered: Jul 2004
Location:
Posts: 1141
psujohnny,
Setting blacks (and whites) properly can be a little complicated. But the first rule of thumb is that whatever looks better TO YOU is likely the proper setting (once your eye is used to just how good a calibrated image can look). You can use Avia or DVE to test how far your preferred setting is from a given "reference" setting and to return to your preferred setting in a precise manner.

"Black" is encoded on the DVD as digital 16. The "Blacker than Black" bar is encoded on DVE as digital 7. The other two bars on DVE are digital 21 (the inner bar) and 25 (the middle bar).

On Avia the dimmer moving bar is digital 19 and the brighter bar is digital 21.

Blacker than Black detail represents light levels that were recorded below what the cinematographer and DVD editor chose to arbitrarily identify as "Black" for reference and quality control purposes.

Technically, Black and Above detail should always be visible and Blacker than Black detail should *NOT* be visible -- although it needs to be present so that the signal processing that happens before the pixels light up can take advantage of it. So technically you should adjust things so that digital 17 and above (only) are visible against a black, digital 16, background. If you raise Brightness beyond that point, Blacker than Black details (below 16) become visible as kind of darker "holes" in a now slightly grayish "Black" background.

If you adjust with DVE so that the Blacker than Black bar JUST becomes invisible, then you probably are still showing some levels of Blacker than Black data between 8 and 16 "since you have made 7 your point of invisibility . If you adjust with Avia so that the dimmer moving bar is JUST invisible (i.e., as the Avia narration suggests) then you are possibly concealing data in the range of 17 to 19 since you have made 19 your point of invisibility. Meanwhile Guy Kuo of Ovation recommends raising Brightness a bit further for digital displays so that the dimmer moving black bar in Avia is JUST BARELY visible -- and so do I. Which is better?

THAT is a question of what the guys who produced the DVD INTENDED you to see!

You see the guys who are setting the balances when the DVD transfer is made are not so much trying to hit some mathematically perfect setting as they are trying to predict what the result will look like on your TV. That's because they are trying to hit a compromise that will work across the entire dynamic range of the film.

Now it is most typically the case that quality control for DVD transfers is done by watching the results on CRT based monitors, partly because that's what they have to work with and partly because they assume that's what most of the buying audience will be using. CRT based monitors have "floating" blacks -- black levels that rise and fall according the average brightness of what's on screen at the moment. So they assume that when the resulting DVD is played at home certain "Blacker than Black" details will float into view in scenes with medium to high brightness and will be concealed in darker scenes, and so they keep tweaking the black level on the transfer to try to get the visible level of detail "just right" -- based on the "float" that's happening in THEIR monitors.

Which means that no matter what level you set on your TV at home, you won't see precisely what they intended you to see unless your TV happens to have exactly the same degree of "float" as their studio monitors!

Digital displays such as LCDs and Plasmas typically have no "float" -- although there are some very few digital displays out there which are designed to emulate the "float" of a CRT. Now "float" is not really a good thing. Ideally a display would have no "float". But since the DVDs are made assuming your display has this flaw, you may need a compromise setting in your "better engineered" display in order to see what the DVD producers intended you to see.

The Avia moving bar chart is made with a half screen of gray to trigger a certain amount of "float" in TVs that do that. By making the leftmost moving bar JUST invisible, you have achieved a compromise setting that works across a fair range of CRT based systems that have a common degree of float. That is, you have set a black level that's deliberately a little too low knowing that it will float up. Avia includes variations of the moving bar pattern that have different amounts of stuff on the other half of the screen so that you can check just how much "float" seems to be affecting your black settings. One such pattern is totally black except for the moving bars. Another has a bright white half screen and a third has gray steps on the right side. Check all 4 patterns to see how the two moving black bars increase or decrease in visibility at any given set of level settings. You will need to get up close to your screen so that the sensitivity of YOUR EYE is not altered too much by the brightness on the other half of the screen.

For a digital display without "float" that Avia setting -- as described in the narration -- is too low. You will lose some black details you were supposed to see. So instead target a bit higher Brightness where the leftmost moving bar is JUST BARELY visible above the black background. Again you can use the other charts to double check whether your TV has "float" or not.

So is the DVE narration just wrong? No.

You see there's another school of thought which says that for digital displays without "float" what you really want to do is to make a compromise setting that's deliberately set a bit TOO HIGH so as to force some Blacker than Black detail to become visible! If you make the Blacker than Black bar in DVE JUST invisible, then you have made the top half of the Blacker than Black range -- from digital 8 to 15 -- just barely visible. The idea is supposed to be that bringing some of this Blacker than Black detail up into visibility will better match what the DVD producers intended you to see IN BRIGHTER SCENES where that stuff WOULD have floated into view if your TV only had some "float". But on the other hand, setting it that way means you will see TOO MUCH Blacker than Black detail in darker scenes -- which could yield noisier dark scenes since the quality control on Blacker than Black data is lower all along the reproduction chain.

Others say these folks are simply "thinking too much" and should just relax and set the display to the "proper" level where 17 is visible over 16 and 15 is not visible under 16 -- and then go buy good DVDs where the editors weren't trying to second guess how your TV works so much.

So how do you choose which is really best for YOUR TV?

Well of course the big test is what looks best to you while watching movies. But there are some other tests. First, digital displays (and to some degree CRT systems) aren't equally good at divvying up the gray scale range for all pairs of black and white end points. If you have several candidate setting pairs of Brightness and Contrast, one way to choose between them is to look at the Black to White gray scale ramps (or the higher detail, partial range, "reverse" gray ramps on DVE) to see which candidate setting produces what looks like the smoothest gray ramp.

A professional would do the same thing when measuring gray steps with a light sensor to try to achieve the best linearity across the range of the gray ramp.

Also, some TV's will have inherent limitations on just how far you can raise Brightness. You know of course that if you raise white levels with Contrast too high you may run into limits of your display -- points you don't want to cross -- points which you need to remember as upper limits for how far you might choose to raise Contrast. For example, you might see blooming on a CRT system or your might see thresholding (loss of white detail) on a digital system. Well some displays have limits on how high you can set Brightness. On those displays if you try to raise the base black level too far the display gets into a kind of rounding error as it tries to raise the over-all light output of "Black" a tiny bit. Digital displays have only a fixed number of steps of light output for each pixel, and so if the display wants to raise the light output just a little bit it has to barely turn on only a scattering of pixels. So if you get up close to your display and see widely scattered pixels just barely turned on in a nearly regular pattern (usually green or purple) in what should be "Black" then you know you have Brightness set too high. And thus that will become an upper limit on how high you can raise Brightness for your setup. Keep in mind that Brightness and Contrast interact so you need to check this for each pair of settings.

Finally there's the problem that not all source content is created equal. This is particularly so when watching broadcast TV. For broadcast TV, and for some cases of poorly produced DVDs, you will need to decide whether you want to (1) calibrate for perfect content and let the imperfect stuff look not quite as good as it might, or (2) pick a compromise set of settings that is not as good as it could be for perfect content but is less likely to annoy you when watching less than perfect content, or (3) save a few different sets of settings so that you have a "perfect" set and also settings that seem to work better with crappy source content.

Personally, I recommend calibrating for high quality, "perfect", source content and letting the crappy stuff look as bad as it wants to.
--Bob

A Never Ending Quest Τowards Visual Nirvana..
takisot is offline  
post #117 of 486 Old 04-16-2005, 08:13 AM - Thread Starter
AVS Special Member
 
cyberbri's Avatar
 
Join Date: May 2004
Location: San Jose, CA
Posts: 8,095
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 52
Takisot,

That still does not say to set your brightness at Digital 7. 7 is where the BTB bars are encoded. Black should be at 16, with above 16 visible and 16 and below as black. He talks about how CRTs have a floating black level, to varying degrees depending on the display, and how if you have a digital display you may choose to slightly vary your brightness setting to include any slight detail that may fall in that range. Whether or not they float all the way down to 7 is another matter. The values may be at 15 or 14 or 12, but maybe not all the way down to 8 or 9. So setting your black level at 7 means that you will probably never see any "blacks", and what is supposed to be blacki/16 will be an elevated dark grey. You can set brightness a few notches below 16 to show data that might come in just below "black", but that's something you have to decide on your own. Ie., if this makes the letterbox bars no longer the darkest black your display can produce, you may not want noise/light there.



I know that on my DLP, if I raised my brightness/black level to include any possible 16-below information, I would never have "blacks" that are black - the only time I would have actual blacks would be if there was 16-below information. And the "black" on my display isn't a deep black anyway like a good RP-CRT, so I can't afford to have no "blacks". So I have chosen to set my brightness/black level at 16, with 17 barely visible, and a good gradation through the low end and all the way through the spectrum. See my original post at the top of this thread to see how I reached my settings.
cyberbri is offline  
post #118 of 486 Old 04-16-2005, 09:08 AM
AVS Special Member
 
takisot's Avatar
 
Join Date: Jan 2002
Location: Athens, Greece
Posts: 1,101
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:


Originally posted by cyberbri


That still does not say to set your brightness at Digital 7. 7 is where the BTB bars are encoded. Black should be at 16, with above 16 visible and 16 and below as black.

Yes, but if I lower my brightness in order to have the Below Black Bar disappear into the Digital 16 Black background of DVE title 12 chapter 13, doesn't that mean that I have automatically set my brightness at Digital 7 (where the BTB bar is encoded)?

A Never Ending Quest Τowards Visual Nirvana..
takisot is offline  
post #119 of 486 Old 04-16-2005, 09:53 AM - Thread Starter
AVS Special Member
 
cyberbri's Avatar
 
Join Date: May 2004
Location: San Jose, CA
Posts: 8,095
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 52
No, that means that Digital 7 has been blended into the black background because you are setting your display to 16 (with 17 visible and 16 as the darkest black your display will go). If you can distinguish the Digital 7 BTB bar from the background, then your "black" floor/level is higher than 7, so any blacks in movies won't appear black anymore.

For example, if you look at DVE on a PC display that is calibrated to 0-255, 0 is the blackest the display shows, and 255 is the brightest. So digital 7 will be up from the blackest black, and 16, where "black" on DVDs is, is a dark grey. That's why Overlay, used with PC levels (0-255), expands the 16-235 range to 0-255 so you don't have to re-calibrate your display or watch grey blacks. So VMR9, which does NOT expand to PC levels, passes or maintains the whole 0-255 range, of which mainly 16-235 is used. This is why you need to lower your brightness so that black is at or around 16 (below that blends in with black).
cyberbri is offline  
post #120 of 486 Old 04-16-2005, 10:11 AM
AVS Special Member
 
takisot's Avatar
 
Join Date: Jan 2002
Location: Athens, Greece
Posts: 1,101
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 0 Post(s)
Liked: 10
Quote:


Originally posted by cyberbri
No, that means that Digital 7 has been blended into the black background because you are setting your display to 16 (with 17 visible and 16 as the darkest black your display will go). If you can distinguish the Digital 7 BTB bar from the background, then your "black" floor/level is higher than 7, so any blacks in movies won't appear black anymore.


Sorry, I still dont get it. How do I get digital 16, t when the digital 7 BTB bar blends into black background? If , by lowering gradually the brightness my black level matches the black of the BTB bar, then logically my black level should be at at 7 not 16! If the BTB bar was at digital 15, then it would make sense to me the way you are exlaining it!
Am I making some kind of fundamental mistake here?

A Never Ending Quest Τowards Visual Nirvana..
takisot is offline  
Reply Home Theater Computers

User Tag List

Thread Tools
Show Printable Version Show Printable Version
Email this Page Email this Page


Forum Jump: 

Posting Rules  
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off