What is BT.1886 gamma? - Page 2 - AVS Forum
Forum Jump: 
 10Likes
Reply
 
Thread Tools
post #31 of 96 Old 07-10-2014, 07:09 AM
Member
 
randal_r's Avatar
 
Join Date: Mar 2010
Posts: 174
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 54 Post(s)
Liked: 67
Quote:
Originally Posted by Wouter73 View Post
Hi,<br>
I read a lot of discussions on gamma, and while I do believe I have a rudimentary understanding of what gamma is and what it means to have a certain number of gamma (2.22,2.3,2.4 and so on), I also read about BT.1886 gamma. But I can't find a clear explanation about what it is? I understand it is some kind of standard, but what does it do exactly?<br><br>
The reason I ask, is because in autocal I have the option to choose bt1886.<br><br>
Thanks!<br><br>
Wouter.
Isn't it amazing how a simple question can take on life on it's own and mutate into a completely different subject if given enough time. Since everyone is giving their two cents worth I give you mine. Download the PDF file that Sotti linked to you in post #4 of this thread (well worth reading). Then watch this episode of Twit - home theater geeks #208 http://twit.tv/show/home-theater-geeks/208

Be careful of the advise you follow but be patient with those who supply it. There are some very knowledgeable individuals on this forum. All in all, it is up to you to sift through and separate the brass from the garbage.
randal_r is offline  
Sponsored Links
Advertisement
 
post #32 of 96 Old 07-10-2014, 09:28 AM
AVS Special Member
 
HDTVChallenged's Avatar
 
Join Date: Jan 2002
Posts: 8,576
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 146
Quote:
Originally Posted by gwgill View Post
BT.1886 uses an input black point offset strategy, so it is more accurate to say "an EOTF using input black offset and a power of 2.2", but "BT.1886-like with a power of 2.2" is fine by me too.
Still doesn't make sense. BT1886 gamma is 2.4 ... period.

Quote:
Originally Posted by gwgill View Post
Except that in practice, there really isn't such a think as a "straight" power law, since in general no display has a zero black level..
True ... but in my experience the display gamma presets usually act as if the black level is zero. In fact, my LG actually calculates the 11pt luminance PL targets for you (based on black level =0)

My point is that if your projected BT1886 luminance curve is "hugging" a power-law luminance curve w/ gamma =2.2, you probably should start your multipoint BT1886 calibration from the display's PL w/ gamma=2.2 preset ... not 2.4, 2.0 or 1.8.

Last edited by HDTVChallenged; 07-10-2014 at 09:34 AM.
HDTVChallenged is offline  
post #33 of 96 Old 07-10-2014, 11:10 AM
Advanced Member
 
spacediver's Avatar
 
Join Date: May 2013
Location: Toronto
Posts: 943
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 120 Post(s)
Liked: 75
Quote:
Originally Posted by HDTVChallenged View Post
Still doesn't make sense. BT1886 gamma is 2.4 ... period.
Makes perfect sense. The BT1886 is characterized by two properties:

1) the input offset
2) the particular exponent that is used 2.4

There are reasons why you might want to change this to 2.2 - for example brighter lit environments, or differently encoded material. In such a case, it's sometimes useful to be able to preserve the input offset property while changing the exponent. The phrase "BT.1886-like function with an exponent of 2.2" makes perfect sense with the assumption that others understand what is meant by it.
spacediver is online now  
post #34 of 96 Old 07-10-2014, 01:24 PM
AVS Addicted Member
 
Join Date: Sep 2002
Location: Wetumpka, AL
Posts: 15,493
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 181 Post(s)
Liked: 199
Quote:
Originally Posted by randal_r View Post
...snip...

Then watch this episode of Twit - home theater geeks #208 http://twit.tv/show/home-theater-geeks/208

....snip...
Thanks for the link.

It sure provided clarity about why certain gammas worked/ didn't work for me.

Samsung 64F8500, Panasonic 65VT50, Oppo 95, Tivo Roamio for OTA, Dish VIP722, Marantz AV8801 preamp, Rotel Amps, Atlantic Tech 8200 speakers, Seaton Submersive HP, Calman 5, Chromapure, Accupel DVG-5000, i1Display3pro, i1pro2, eecolor colorbox.
JimP is offline  
post #35 of 96 Old 07-10-2014, 04:22 PM
AVS Special Member
 
LarryInRI's Avatar
 
Join Date: Jun 2007
Location: in RI
Posts: 5,273
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 5 Post(s)
Liked: 151
Quote:
Originally Posted by spacediver View Post
Makes perfect sense. The BT1886 is characterized by two properties:

1) the input offset
2) the particular exponent that is used 2.4

There are reasons why you might want to change this to 2.2 - for example brighter lit environments, or differently encoded material. In such a case, it's sometimes useful to be able to preserve the input offset property while changing the exponent. The phrase "BT.1886-like function with an exponent of 2.2" makes perfect sense with the assumption that others understand what is meant by it.
Absolutely (pun intended). I've been using a BT.1886-like function with an exponent of 2.27 for many older transfers and the results are great.

Larry
LarryInRI is offline  
post #36 of 96 Old 07-10-2014, 05:27 PM
Advanced Member
 
gwgill's Avatar
 
Join Date: Jan 2013
Location: Melbourne, Australia
Posts: 677
Mentioned: 7 Post(s)
Tagged: 0 Thread(s)
Quoted: 100 Post(s)
Liked: 88
Quote:
Originally Posted by HDTVChallenged View Post
Still doesn't make sense. BT1886 gamma is 2.4 ... period.
By definition, "BT.1886-like" is not BT.1886.
Quote:
Originally Posted by HDTVChallenged
True ... but in my experience the display gamma presets usually act as if the black level is zero. In fact, my LG actually calculates the 11pt luminance PL targets for you (based on black level =0)
If this is in fact what it's doing, then there will be a discontinuity when it transitions between the black level and the power curve, although it may be a very small discontinuity if the black level is low. But part of my suspicion is that most of the time these systems are fooling you - they are omitting the black level when describing the curve, and it gets added in afterwards without telling you.
gwgill is offline  
post #37 of 96 Old 07-11-2014, 06:37 AM
Member
 
randal_r's Avatar
 
Join Date: Mar 2010
Posts: 174
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 54 Post(s)
Liked: 67
Quote:
Originally Posted by JimP View Post
Thanks for the link.

It sure provided clarity about why certain gammas worked/ didn't work for me.
I regret that you came away feeling that you got nothing out of it.
randal_r is offline  
post #38 of 96 Old 07-11-2014, 10:14 AM
AVS Special Member
 
WiFi-Spy's Avatar
 
Join Date: Feb 2004
Location: Seattle, WA
Posts: 2,310
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 26 Post(s)
Liked: 33
Quote:
Originally Posted by LarryInRI View Post
Absolutely (pun intended). I've been using a BT.1886-like function with an exponent of 2.27 for many older transfers and the results are great.

Larry
In CalMAN there is a "Sliding Power" gamma that works like BT.1886 except you pick the curve.

Tyler Pruitt - Technical Liaison at SpectraCal
WiFi-Spy is offline  
post #39 of 96 Old 07-11-2014, 11:02 AM
AVS Addicted Member
 
Join Date: Sep 2002
Location: Wetumpka, AL
Posts: 15,493
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 181 Post(s)
Liked: 199
Quote:
Originally Posted by randal_r View Post
I regret that you came away feeling that you got nothing out of it.
How did you get that from what I posted?

I got a lot out of it.

I think the big mistake is the assumption that BT1886 works best for everyone without mention of viewing conditions. Whether its bat cave or sun room conditions....it's not mentioned.

Samsung 64F8500, Panasonic 65VT50, Oppo 95, Tivo Roamio for OTA, Dish VIP722, Marantz AV8801 preamp, Rotel Amps, Atlantic Tech 8200 speakers, Seaton Submersive HP, Calman 5, Chromapure, Accupel DVG-5000, i1Display3pro, i1pro2, eecolor colorbox.
JimP is offline  
post #40 of 96 Old 07-11-2014, 11:05 AM
AVS Special Member
 
TVOD's Avatar
 
Join Date: Jun 2003
Posts: 4,881
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 74 Post(s)
Liked: 70
I suspect video enthusiasts set their black level to the lowest possible without clipping. My experience with mastering several years ago was that displays were set slightly above to ensure information at black was still rendered, often by use of the negative pludge on color bars. This would slightly lower the equivalent gamma on the display.

709 defines a gamma curve with a linear breakpoint around 8% (on gamma corrected video). If a pure exponential curve is used without a breakpoint, then in theory a lower value needs to be used. The flip side is that in CRT displays I don't remember there being gamma compensation so it's sort of a moot point. It really comes down to what the monitor is displaying.

Last edited by TVOD; 07-11-2014 at 02:23 PM.
TVOD is offline  
post #41 of 96 Old 07-11-2014, 11:27 AM
Advanced Member
 
spacediver's Avatar
 
Join Date: May 2013
Location: Toronto
Posts: 943
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 120 Post(s)
Liked: 75
Quote:
Originally Posted by TVOD View Post
709 defines a gamma curve with a linear breakpoint around 8% (on gamma corrected video).
encoding gamma you mean, right?
spacediver is online now  
post #42 of 96 Old 07-11-2014, 01:09 PM
Member
 
randal_r's Avatar
 
Join Date: Mar 2010
Posts: 174
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 54 Post(s)
Liked: 67
Quote:
Originally Posted by JimP View Post
How did you get that from what I posted?

I got a lot out of it.

I think the big mistake is the assumption that BT1886 works best for everyone without mention of viewing conditions. Whether its bat cave or sun room conditions....it's not mentioned.
My apologies, by your statement; "It sure provided clarity about why certain gammas worked/ didn't work for me", last part "didn't work for me" lead me to an assumption stated in my former post.

Through the various pieces of literature that I have come across on the subject, it seems to only be that BT-1886 is an alternative to Gamma 2.4. Where the need for a higher Gamma the BT-1886 does not apply. There is no way what Gamma to set one's display to for a particular movie. I challenge anyone to find where a Gamma setting is on a DVD cover. Gamma settings are not written in stone. When is a display image unwatchable? Is it Gamma 1.8, 2.0, 2.2, 2.4 or 2.6? I agree with your statement: "I think the big mistake is the assumption that BT1886 works best for everyone without mention of viewing conditions". In the Home Theater Geeks #208 when Joel Silvers described the viewing conditions in his sunroom in Florida, he stated that his display was around 1.8 .

Again, I apologizes for misconstruing your statement.
randal_r is offline  
post #43 of 96 Old 07-11-2014, 01:59 PM
AVS Special Member
 
TVOD's Avatar
 
Join Date: Jun 2003
Posts: 4,881
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 74 Post(s)
Liked: 70
Quote:
Originally Posted by spacediver View Post
encoding gamma you mean, right?
Yes, the YUV gamma corrected video. I've used 'encoded video' before, which is a holdover from the composite days (commonly used around the time the earth was thawing out from the last ice age), but it's somewhat inaccurate referring to component video.
TVOD is offline  
post #44 of 96 Old 07-11-2014, 02:17 PM
AVS Special Member
 
TVOD's Avatar
 
Join Date: Jun 2003
Posts: 4,881
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 74 Post(s)
Liked: 70
Quote:
Originally Posted by randal_r View Post
Is it Gamma 1.8, 2.0, 2.2, 2.4 or 2.6? I agree with your statement: "I think the big mistake is the assumption that BT1886 works best for everyone without mention of viewing conditions". In the Home Theater Geeks #208 when Joel Silvers described the viewing conditions in his sunroom in Florida, he stated that his display was around 1.8.
If one raises the brightness on the display, say to match the dark areas of the screen to the ambient reflections, and the brightness offset is applied before the display's exponential correction (along with gain normalization) the effective gamma value decreases.

As for the gamma of what's on specific material, it can really gets down to whatever a colorist thought looked good, and how the monitor was set. Even with objective measurements and standards, the final look is often a subjective decision. Standards such as BT1886 only attempt to make it look consistent on various displays. Even BT1886 has a variance which depends on a slight change in the brightness setting where a breakpoint is used between gamma values.
TVOD is offline  
post #45 of 96 Old 07-12-2014, 11:03 AM
AVS Special Member
 
HDTVChallenged's Avatar
 
Join Date: Jan 2002
Posts: 8,576
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 146
Quote:
Originally Posted by spacediver View Post
Makes perfect sense. The BT1886 is characterized by two properties:

1) the input offset
2) the particular exponent that is used 2.4

There are reasons why you might want to change this to 2.2 - for example brighter lit environments, or differently encoded material. In such a case, it's sometimes useful to be able to preserve the input offset property while changing the exponent. The phrase "BT.1886-like function with an exponent of 2.2" makes perfect sense with the assumption that others understand what is meant by it.
The exponent (aka gamma) is 2.4 ... period ... by definition. What you are trying to do is to create yet another "sub-standard," which is the exact opposite of what intended with BT1886. If that's your position, then why bother with BT1886 at all? Just go ahead and randomly adjust your luminance curve until it looks good.

Last edited by HDTVChallenged; 07-12-2014 at 11:07 AM.
HDTVChallenged is offline  
post #46 of 96 Old 07-12-2014, 03:04 PM
Advanced Member
 
spacediver's Avatar
 
Join Date: May 2013
Location: Toronto
Posts: 943
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 120 Post(s)
Liked: 75
Quote:
Originally Posted by HDTVChallenged View Post
The exponent (aka gamma) is 2.4 ... period ... by definition. What you are trying to do is to create yet another "sub-standard," which is the exact opposite of what intended with BT1886. If that's your position, then why bother with BT1886 at all? Just go ahead and randomly adjust your luminance curve until it looks good.
I'm not trying to create anything, I'm trying to articulate to you how we can use language to communicate ideas. In this case, using an input offset with a 2.2 exponent can be efficiently communicated using the phase "BT.1886 like curve with an exponent of 2.2". Most people who understand BT.1886 fully grasp what is meant by this phrase.

I also sense the idea that you're not comfortable with people using customized exponents - you should understand that not all people are able to watch material in ideal lighting conditions, and some degree of flexibility is required to maintain the intended perceptual result. And you should recognize that not all material was mastered with a 2.4 exponent.
spacediver is online now  
post #47 of 96 Old 07-12-2014, 03:49 PM
AVS Special Member
 
TVOD's Avatar
 
Join Date: Jun 2003
Posts: 4,881
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 74 Post(s)
Liked: 70
Quote:
Originally Posted by HDTVChallenged View Post
The exponent (aka gamma) is 2.4 ... period ... by definition.
Not exactly. 2.4 is the reference curve, but Appendix 1 of the ITU document says:

"The EOTF specified in Annex 1 is considered to be a satisfactory, but not exact, match to the characteristic of an actual CRT. When it is desired to match a CRT, the Lw and LB parameters of the EOTF can be set to the corresponding values of the CRT that are being matched. For moderate black level settings, e.g. 0.1 cd/m2, setting the LB of the EOTF to 0.1 will give a satisfactory match to the CRT. In the event the CRT is operated at a lower black level, e.g. 0.01 cd/m2, the EOTF will provide a better match with LB set to a lower value such as 0.0 cd/m2. When it is necessary to more precisely match a flat panel display characteristic to a CRT, the alternative EOTF formulation specified below may provide a solution."

The following parameters in the document shows a breakpoint in the exponential correction. Beyond this issue is the rec 709 linear section in the lowlights. Should that also be part of the final curve? Using a pure exponential curve has more gain reduction in the dark areas.
TVOD is offline  
post #48 of 96 Old 07-12-2014, 04:01 PM
AVS Special Member
 
TVOD's Avatar
 
Join Date: Jun 2003
Posts: 4,881
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 74 Post(s)
Liked: 70
Quote:
Originally Posted by spacediver View Post
And you should recognize that not all material was mastered with a 2.4 exponent.
When CRTs were in common use, the Sony monitors were a de facto standard with their gamma. Now that has largely been replaced with displays using LUTs. Hopefully they should closely adhere to a EOTF standard, but I have my doubts. Some 'broadcast' LCDs I've seen are simply horrible on black levels.
TVOD is offline  
post #49 of 96 Old 07-13-2014, 09:50 AM
AVS Special Member
 
HDTVChallenged's Avatar
 
Join Date: Jan 2002
Posts: 8,576
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 146
Quote:
Originally Posted by TVOD View Post
Not exactly. 2.4 is the reference curve, but Appendix 1 of the ITU document says:

"The EOTF specified in Annex 1 is considered to be a satisfactory, but not exact, match to the characteristic of an actual CRT. When it is desired to match a CRT, the Lw and LB parameters of the EOTF can be set to the corresponding values of the CRT that are being matched. For moderate black level settings, e.g. 0.1 cd/m2, setting the LB of the EOTF to 0.1 will give a satisfactory match to the CRT. In the event the CRT is operated at a lower black level, e.g. 0.01 cd/m2, the EOTF will provide a better match with LB set to a lower value such as 0.0 cd/m2. When it is necessary to more precisely match a flat panel display characteristic to a CRT, the alternative EOTF formulation specified below may provide a solution."

The following parameters in the document shows a breakpoint in the exponential correction. Beyond this issue is the rec 709 linear section in the lowlights. Should that also be part of the final curve? Using a pure exponential curve has more gain reduction in the dark areas.
Don't confuse the two sections. The "reference" (top) section is the actual "recommendation/standard" the lower (alternate) section only applies if you are trying to match two specific monitors ... because you want to use them side-by-side ... If I'm reading the technicalese correctly.

IOW, if your IPS based LCD display can only get down to 0.12Nits, you're never going to be able to "match" it to a CRT with a black level of 0.01Nits ... but you can get close to matching it to a CRT with its black-level cranked up to 0.12Nits ... YMMV.
HDTVChallenged is offline  
post #50 of 96 Old 07-13-2014, 10:02 AM
AVS Special Member
 
HDTVChallenged's Avatar
 
Join Date: Jan 2002
Posts: 8,576
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 146
Quote:
Originally Posted by spacediver View Post
I also sense the idea that you're not comfortable with people using customized exponents - you should understand that not all people are able to watch material in ideal lighting conditions, and some degree of flexibility is required to maintain the intended perceptual result. And you should recognize that not all material was mastered with a 2.4 exponent.
True ... but what you are doing doesn't really solve the problem. At some point, if you want to have an actual standard you simply *have to stop* doing things the "wrong way" and start doing them the "correct" way. Period.

At some point, if you want accurate color, you have to stop setting displays to 9500K + whitepoints. At some point, if you want digital/HDTV to take off, you have to turn of all the analog transmitters.

Seriously, why stop with just "custom gamma" ... go ahead and mess with the whitepoint ... whatever makes it "look good."
HDTVChallenged is offline  
post #51 of 96 Old 07-13-2014, 10:08 AM
AVS Special Member
 
HDTVChallenged's Avatar
 
Join Date: Jan 2002
Posts: 8,576
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 146
PS: To be fair ... I suspect that the problem will not be solved within HDTV lifetime. Perhaps there's hope if UHD/4K actually gets us back to linear encode/decode along with other radical departures from the past ... but that appears to still be half a decade or so away.
HDTVChallenged is offline  
post #52 of 96 Old 07-13-2014, 11:57 AM
Advanced Member
 
spacediver's Avatar
 
Join Date: May 2013
Location: Toronto
Posts: 943
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 120 Post(s)
Liked: 75
Quote:
Originally Posted by HDTVChallenged View Post
True ... but what you are doing doesn't really solve the problem.
what problem does this not solve, exactly?

Quote:
Originally Posted by HDTVChallenged View Post
At some point, if you want to have an actual standard you simply *have to stop* doing things the "wrong way" and start doing them the "correct" way. Period.
You really should watch that "all about gamma" episode (if you haven't already). Not everyone has the luxury of watching content in studio conditions, 100 percent of the time.

Here, let me give an example - I'm curious to see how you respond.

1) It is a sunday afternoon and the wimbledon's men final is on. Your kids are home and you want to spend time with them while watching the game. You have two options:

1) Use BT.1886 and turn down all the lights to studio conditions.
2) Use a function that uses the same input offset as BT.1886, but with an exponent of 2.0, and turn lights up enough that your kids can play with their toys when they get periodically bored of watching the game.

2) allows you to spend time in the same room as your kids, and allows you to ameliorate the impact of using non ideal ambient lighting conditions.

Do you not agree that having the option to adjust your luminance function for the second scenario is valuable?
spacediver is online now  
post #53 of 96 Old 07-13-2014, 03:59 PM
Advanced Member
 
CalWldLif's Avatar
 
Join Date: Feb 2011
Location: South Bay'Los Angeles California.
Posts: 518
Mentioned: 1 Post(s)
Tagged: 0 Thread(s)
Quoted: 70 Post(s)
Liked: 29
seems the lower the desired gamma the less curve that would be generated.
might as well just go with a flat gamma setting
CalWldLif is offline  
post #54 of 96 Old 07-13-2014, 09:16 PM
Advanced Member
 
gwgill's Avatar
 
Join Date: Jan 2013
Location: Melbourne, Australia
Posts: 677
Mentioned: 7 Post(s)
Tagged: 0 Thread(s)
Quoted: 100 Post(s)
Liked: 88
Quote:
Originally Posted by HDTVChallenged View Post
PS: To be fair ... I suspect that the problem will not be solved within HDTV lifetime. Perhaps there's hope if UHD/4K actually gets us back to linear encode/decode along with other radical departures from the past ... but that appears to still be half a decade or so away.
The way an image is encoded, can't solve the problem of how to adapt the image for the display limitations and viewing conditions.

Modern HDTV video has a specified encoding - that's what Rec709 is all about. In reality they are only nominally encoded that way, the actual encoding is "whatever looks good to the colorist & director on their reference monitor". To see exactly what they see, you need a display that has the same capabilities as they used, setup the same way, in the same viewing conditions. If at least that was standard, you would have enough information to know how to adapt the video to give the same appearance on your display in your viewing conditions.
gwgill is offline  
post #55 of 96 Old 07-13-2014, 11:47 PM
AVS Special Member
 
HDTVChallenged's Avatar
 
Join Date: Jan 2002
Posts: 8,576
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 146
Quote:
Originally Posted by spacediver View Post
what problem does this not solve, exactly?
The lack of an actual "gamma" setup standard. In fact all you're doing is making the issue worse by trying to tinker with the BT1886 exponent.

Quote:
Originally Posted by spacediver View Post
You really should watch that "all about gamma" episode (if you haven't already).
If you are refering to the "Home theater geeks" video, I have ... at least twice.

Quote:
Originally Posted by spacediver View Post
Here, let me give an example - I'm curious to see how you respond.

1) It is a sunday afternoon and the wimbledon's men final is on. Your kids are home and you want to spend time with them while watching the game. You have two options:

1) Use BT.1886 and turn down all the lights to studio conditions.
2) Use a function that uses the same input offset as BT.1886, but with an exponent of 2.0, and turn lights up enough that your kids can play with their toys when they get periodically bored of watching the game.

2) allows you to spend time in the same room as your kids, and allows you to ameliorate the impact of using non ideal ambient lighting conditions.

Do you not agree that having the option to adjust your luminance function for the second scenario is valuable?
Firstly, I wouldn't have the TV installed in a room where "afternoon lighting" would have such an impact that I would have to resort to "de-tuning" the TV.

Secondly, I don't see the point of screwing with BT1886 ... if you *need* a different gamma for "afternoon," then you can assume that "studio conditions" will not be met no matter what you do. So, in this case, I would just go ahead and set up a power-law curve ... and probably with just a quickie 2pt greyscale at that ... save time and avoid wasting effort that not really going to do much for image quality.

Thirdly, I'm assuming that the "input offset" parameters in BT1886 were derived, in part, on empirical measurements of CRTs set to a nominal gamma of 2.4 ... so changing *just* the exponent for a "BT1886 like" EOTF is questionable ... at the least.

Fourthly, as it turns out *my* BT1886 curve is actually pretty good at handling "daylight" conditions. So if you still insist on this direction, I would recommend that instead of screwing with the exponent, one would do better to raise the black-level and setup a new BT1886 curve according to the established definition.

Finally, as I've said elsewhere, if you're not going to bother with controlling your ambient light levels, then worrying about the minutiae of gamma or calibration in general is pointless. You might as well just use 15000K "torch mode." I could go on, but I think I'm starting to repeat myself.
HDTVChallenged is offline  
post #56 of 96 Old 07-14-2014, 12:02 AM
AVS Special Member
 
HDTVChallenged's Avatar
 
Join Date: Jan 2002
Posts: 8,576
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 3 Post(s)
Liked: 146
Quote:
Originally Posted by gwgill View Post
. To see exactly what they see, you need a display that has the same capabilities as they used, setup the same way, in the same viewing conditions. If at least that was standard, you would have enough information to know how to adapt the video to give the same appearance on your display in your viewing conditions.
Yes ... my point is that advocating for anything other than BT1886 as currently defined only makes this issue worse. To be blunt, I don't see how having multiple "EOTF" setups on the end user's display helps anything. Have one setup for BT1886 and perhaps one set to some variation of PL@2.2 ... one or the other should be "close enough" to cover the majority of video content, and the proper selection should be pretty obvious visually.

Dealing with ambient lighting conditions is better handled by actually dealing with the ambient light ... not the EOTF. And you don't have to make the room "theater dark" either.
HDTVChallenged is offline  
post #57 of 96 Old 07-14-2014, 12:15 AM
Advanced Member
 
gwgill's Avatar
 
Join Date: Jan 2013
Location: Melbourne, Australia
Posts: 677
Mentioned: 7 Post(s)
Tagged: 0 Thread(s)
Quoted: 100 Post(s)
Liked: 88
Quote:
Originally Posted by HDTVChallenged View Post
In fact all you're doing is making the issue worse by trying to tinker with the BT1886 exponent.
I can't agree with that. Viewing conditions vary, so some compensation for that has to be made. In the past, TV's had to put up with some pretty crude controls for that purpose - "Contrast" and "Brightness", because that's what the technology (CRT) made available cheaply. Back in the studio they could use more sophistication, and they did, incorporating gamma encoding in the cameras that made a compensation for the viewing condition difference between the bright studio and the typical not so bright TV in the dim living room. With modern digital processing we don't have to put up with such limitations, we can also apply a gamma curve on each display now. BT.1886 doesn't attempt to address a variable viewing condition adjustment, but extending it to allow for a change in gamma is a very reasonable approach, much better than "Contrast" and "Brightness" controls. This is supported by the type of adjustment curve you get if you run two different ambient light levels into and out of CIECAM02 - you get something that resembles a power curve.
Quote:
Originally Posted by HDTVChallenged
Secondly, I don't see the point of screwing with BT1886 ... if you *need* a different gamma for "afternoon," then you can assume that "studio conditions" will not be met no matter what you do. So, in this case, I would just go ahead and set up a power-law curve ... and probably with just a quickie 2pt greyscale at that ... save time and avoid wasting effort that not really going to do much for image quality.
Switching to an output offset black then leaves you with a less than optimal display if you are using an LCD. The use of input vs. output black offset is not connected with the power chosen. Choosing input offset black (ie, just like BT.1886) and a power value that suits your viewing conditions is probably a better choice.
Quote:
Originally Posted by HDTVChallenged
Thirdly, I'm assuming that the "input offset" parameters in BT1886 were derived, in part, on empirical measurements of CRTs set to a nominal gamma of 2.4 ... so changing *just* the exponent for a "BT1886 like" EOTF is questionable ... at the least.
Wrong assumption. Input offset is a choice made on the basis of how we respond to light - like most of our senses, it is basically ratiometric. So if the dark shadow steps are to be visible starting at a heightened black level, then you need to offset into the curve, so that the steps from black maintain a similar ratio to black. BT.1886 choice of 2.4 power represents an approximation to the CRT characteristic response. All together it is a simplified CRT mode (pure 2.4 power) + input offset to maintain shadow detail in the face of an elevated black point. See Appendix 1 for an EOTF that is meant to more closely mimic a CRT.
Quote:
Originally Posted by HDTVChallenged
Fourthly, as it turns out *my* BT1886 curve is actually pretty good at handling "daylight" conditions. So if you still insist on this direction, I would recommend that instead of screwing with the exponent, one would do better to raise the black-level and setup a new BT1886 curve according to the established definition.
That's only going to help if you are facing a lot of glare. You can be in brighter viewing conditions without this being the issue, and as a result just need to reduce the gamma slightly. (In theory you could measure the display with glare included using a tele-instrument, but in practice this may raise the black too much, because the video encoding assumes some (unspecified) level of glare anyway.)
Quote:
Originally Posted by HDTVChallenged
Finally, as I've said elsewhere, if you're not going to bother with controlling your ambient light levels, then worrying about the minutiae of gamma or calibration in general is pointless. You might as well just use 15000K "torch mode." I could go on, but I think I'm starting to repeat myself.
Certainly if the viewing conditions are dynamically variable, any hope of a quality visual result is small. But a static difference to reference conditions is easily compensated for.
sillysally likes this.

Last edited by gwgill; 07-14-2014 at 12:25 AM.
gwgill is offline  
post #58 of 96 Old 07-14-2014, 12:24 AM
Advanced Member
 
gwgill's Avatar
 
Join Date: Jan 2013
Location: Melbourne, Australia
Posts: 677
Mentioned: 7 Post(s)
Tagged: 0 Thread(s)
Quoted: 100 Post(s)
Liked: 88
Quote:
Originally Posted by HDTVChallenged View Post
Yes ... my point is that advocating for anything other than BT1886 as currently defined only makes this issue worse. To be blunt, I don't see how having multiple "EOTF" setups on the end user's display helps anything. Have one setup for BT1886 and perhaps one set to some variation of PL@2.2 ... one or the other should be "close enough" to cover the majority of video content, and the proper selection should be pretty obvious visually.
You seem to be contradicting yourself. You're saying "there should be one EOTF" then you're saying that "perhaps there should be more than one". The practical reality is that pure BT.1886 doesn't suit everyone situation, although it seems a good place to start. There is no reason why input offset black should only be used with a 2.4 power - these two aspects are completely unconnected. BT.1886 simply adopted input offset black because it addresses certain issues well. Input offset black EOTF's existed before BT.1886 - ArgyllCMS dispcal is testament to that.
gwgill is offline  
post #59 of 96 Old 07-14-2014, 01:12 AM
Advanced Member
 
spacediver's Avatar
 
Join Date: May 2013
Location: Toronto
Posts: 943
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 120 Post(s)
Liked: 75
Quote:
Originally Posted by HDTVChallenged View Post
Firstly, I wouldn't have the TV installed in a room where "afternoon lighting" would have such an impact that I would have to resort to "de-tuning" the TV.
Quote:
Originally Posted by HDTVChallenged View Post
Finally, as I've said elsewhere, if you're not going to bother with controlling your ambient light levels, then worrying about the minutiae of gamma or calibration in general is pointless. You might as well just use 15000K "torch mode." I could go on, but I think I'm starting to repeat myself.
We're not discussing whether one should control ambient lighting conditions - obviously one should strive to whenever possible - but rather how to effectively deal with situations where one cannot control ambient lighting conditions. I'm not sure why one should just use 15000k torch mode just because they can't get good ambient lighting conditions. Isn't it more rational to make the best out of a non ideal situation instead of throwing a fit with your tv?
spacediver is online now  
post #60 of 96 Old 07-14-2014, 02:01 AM
Advanced Member
 
spacediver's Avatar
 
Join Date: May 2013
Location: Toronto
Posts: 943
Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 120 Post(s)
Liked: 75
Quote:
Originally Posted by HDTVChallenged View Post

Fourthly, as it turns out *my* BT1886 curve is actually pretty good at handling "daylight" conditions. So if you still insist on this direction, I would recommend that instead of screwing with the exponent, one would do better to raise the black-level and setup a new BT1886 curve according to the established definition.
Read this paper by Mark Fairchild, and in particular pay attention to figure 5 - it's described on page 10, and keep in mind these are lightness exponents (kind of like OETF's of the human visual system) which are the inverse of a perceptually uniform EOTF. Note how the required gamma exponent for perceptual uniformity drops from from 1/0.33 (~3), to 1/0.44 (~2.25) when the adapting background increases in luminance. What this shows is that you need to change the exponent of your EOTF to maintain consistent perceptual results across different viewing conditions. Note that this is quite a separate issue of increasing the gain within an EOTF as black level rises (this is what input offset does).

Quote:
Originally Posted by HDTVChallenged View Post
Thirdly, I'm assuming that the "input offset" parameters in BT1886 were derived, in part, on empirical measurements of CRTs set to a nominal gamma of 2.4 ... so changing *just* the exponent for a "BT1886 like" EOTF is questionable ... at the least.
I don't think this is the case. See this post, and this one, where a member of this forum derived the BT.1886 function independently by using an input offset raised to the power of the reciprocal of the exponent (2.4). I believe Graeme implemented the same idea in his own software before BT.1886 was created. The BT.1886 function, as I understand it, is simply a function where the exponent is applied to the (video level + offset) as a whole (hence the term input offset). All the other stuff in there is just to ensure the function is ranged between the black level and 1.

Quote:
Originally Posted by HDTVChallenged View Post
Yes ... my point is that advocating for anything other than BT1886 as currently defined only makes this issue worse. To be blunt, I don't see how having multiple "EOTF" setups on the end user's display helps anything. Have one setup for BT1886 and perhaps one set to some variation of PL@2.2 ... one or the other should be "close enough" to cover the majority of video content, and the proper selection should be pretty obvious visually.
BT.1886 and PL@2.2 (output offset) are, by definition, multiple EOTF's. Furthermore, why would you choose 2.2 with an output offset over 2.2 with an input offset??
spacediver is online now  
Reply Display Calibration

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