Rich H
Second Unit
- Joined
- May 22, 2001
- Messages
- 283
Hi folks. We sort of hashed some of this stuff out in another forum, but I'd like to put these questions and observations to the many knowledgeable and interested folks here:
I find myself fascinated by the notions of "picture quality," especially as it relates to subjective Vs objective evaluations. By subjective I mean "does it fool our eyes into believing?" Vs Objective "How does the display measure in relation to adopted standards? And are those adopted standards (e.g. NTSC) putting a hammerlock on what is considered "good" picture quality?
"Good" picture quality is often associated with NTSC standards and how closely a display can track those standards. But I also believe good picture quality can be "released" from those constraints, and evaluated on general notions of how natural and realistic the image looks - how convincing it is over-all and how easily it allows us to suspend our belief.
Most people think an ISF'd display automatically equates to a more realistic image. However, the calibration rendered by an ISF technician does not have the goal per se of “creating a real-looking picture.” It is instead goal number one for an ISF technician to bring the performance of your display in line with NTSC imaging standards. A smoother, more realistic image is often the happy by-product of a display calibrated to NTSC standards, especially compared to most artificial-looking factory settings. But the essential idea of such calibration is that a movie image should look the same on my display as it looks on your display, as it looked on the original broadcast monitors used for transferring the image.
Many HT enthusiasts are thrilled by an image that imitates projected film. I'm one of them. But I also have a foot in the other camp that is thrilled by an image that has amazing verisimilitude; where the image on a display looks amazingly close to how it looks in real life. I think most of us are when we encounter such a display.
I had my plasma professionally (ISF) calibrated several months ago. It now runs a nice D6500K etc. Given the received wisdom that "a display will only perform to it's picture quality potential when properly calibrated to professional standards..." It was interesting to find it wasn't the case (to my eyes). Previous to the ISF, I had calibrated the plasma based on it's strengths and weaknesses, to provide the most natural, believable, and realistic image I could attain. I actually used "reality" as my comparison point, observing how people and objects appear in real life (looking at real skin etc.) and calibrating for this look. It was amazingly successful: my plasma had a looking-through-a-window-at-real-events effect on good source material. Whereas, with the ISF calibration I traded that realism for flatter, over-all more subdued, romantic, film-like look. Yes, the ISF settings do mimic projected film more, and no doubt my display now looks more like the professional monitor on which the DVD material was mastered. But do the images look more convincing now? No. Less so. Do I see more detail in dark areas, color areas, or high-light areas with the ISF settings compared to my settings: No. Quite a few other people have had the same experience as I have. (And of the twelve people I've demoed my settings vs the ISF settings to, all agreed mine created the more realistic, convincing image).
Personal observations: Based on just about every ISF'd image I've seen, I still feel I'm seeing a Hagan-Daz version of reality - warm, creamy 'n smooth, and over-rich in color (a generalization). I've found dialing down the over-all color level, even on ISF'd settings, decreases the sense of "people and objects as neon signs," - that over-lush glow to everything - to a flatter, more texturally real image. (To analogize: A mat picture vs a glossy photo). Also, I find at least in the case of my plasma, carefully dialing the contrast and brightness beyond ISF settings (but still way below torch mode)also increase the realism of the image. I don't get any obvious loss of detail in high-lights - only a more realistically dynamic image. Etc.
During the calibration I sometimes had the sense the plasma was being made to jump through hoops made for another show animal (CRT), and that it's performance was evaluated solely on how well it mimicked that animal's abilities, rather than how well it displayed it's own. As a "broadcast monitor" my plasma might ultimately fail, due to it being unable to eek the last crumbs of detail in the bottom of the gray scale. BUT, as a "virtual reality machine" it excels. To lift a quote from a Director Of Photography who also notes the compelling images from plasmas: "CRTs are naturally going to score better in tests designed around the performance envelope of the CRT." But "I think we're finally seeing that ISF test performance numbers aren't capable of articulating all of the qualitative properties we're using when we perceive a picture."
So, there are two issues here:
1. How NTSC standards and professional calibration relate to the latest technologies. For consistency sake, the current standards make some logical sense..if strict accuracy to the original transfer is the goal. But when we calibrate every display type we encounter to look more "the same," more like standard broadcast monitors, what might we be loosing? Do we ignore some of the specific strengths of new technologies when we calibrate them around an envelope designed for CRT?
2. How NTSC standards themselves seem to be held as the only measure of how images should look. I assert that more elbow room should be made in the "hall of respectability" for varying notions of picture quality. For many, adherence to NTSC standards may be the end of the story in evaluating a display's worth. But to others adherence to NTSC standards is not, in of itself, good enough to guarantee the most convincing image possible.
3. I'm aware that, as in any trade or endeavor, in the ISF and broadcast technology world there is debate about these subjects. I'd enjoy hearing from the pros just what type of debates are ongoing these days.
Whew. Sorry for the long post. I look forward to being educated or corrected wherever I've gone wrong. And this isn't by any means meant to impugn the idea of ISF calibration, nor the great job many calibrators do in making so many people happier with their display. (I also realize that a full ISF calibration is likely to reap more rewards on CRT-based technology, given there is more to tweak there vs fixed pixel devices). I just want to peak out of the ISF / NTSC box to see what's there
Cheers,
I find myself fascinated by the notions of "picture quality," especially as it relates to subjective Vs objective evaluations. By subjective I mean "does it fool our eyes into believing?" Vs Objective "How does the display measure in relation to adopted standards? And are those adopted standards (e.g. NTSC) putting a hammerlock on what is considered "good" picture quality?
"Good" picture quality is often associated with NTSC standards and how closely a display can track those standards. But I also believe good picture quality can be "released" from those constraints, and evaluated on general notions of how natural and realistic the image looks - how convincing it is over-all and how easily it allows us to suspend our belief.
Most people think an ISF'd display automatically equates to a more realistic image. However, the calibration rendered by an ISF technician does not have the goal per se of “creating a real-looking picture.” It is instead goal number one for an ISF technician to bring the performance of your display in line with NTSC imaging standards. A smoother, more realistic image is often the happy by-product of a display calibrated to NTSC standards, especially compared to most artificial-looking factory settings. But the essential idea of such calibration is that a movie image should look the same on my display as it looks on your display, as it looked on the original broadcast monitors used for transferring the image.
Many HT enthusiasts are thrilled by an image that imitates projected film. I'm one of them. But I also have a foot in the other camp that is thrilled by an image that has amazing verisimilitude; where the image on a display looks amazingly close to how it looks in real life. I think most of us are when we encounter such a display.
I had my plasma professionally (ISF) calibrated several months ago. It now runs a nice D6500K etc. Given the received wisdom that "a display will only perform to it's picture quality potential when properly calibrated to professional standards..." It was interesting to find it wasn't the case (to my eyes). Previous to the ISF, I had calibrated the plasma based on it's strengths and weaknesses, to provide the most natural, believable, and realistic image I could attain. I actually used "reality" as my comparison point, observing how people and objects appear in real life (looking at real skin etc.) and calibrating for this look. It was amazingly successful: my plasma had a looking-through-a-window-at-real-events effect on good source material. Whereas, with the ISF calibration I traded that realism for flatter, over-all more subdued, romantic, film-like look. Yes, the ISF settings do mimic projected film more, and no doubt my display now looks more like the professional monitor on which the DVD material was mastered. But do the images look more convincing now? No. Less so. Do I see more detail in dark areas, color areas, or high-light areas with the ISF settings compared to my settings: No. Quite a few other people have had the same experience as I have. (And of the twelve people I've demoed my settings vs the ISF settings to, all agreed mine created the more realistic, convincing image).
Personal observations: Based on just about every ISF'd image I've seen, I still feel I'm seeing a Hagan-Daz version of reality - warm, creamy 'n smooth, and over-rich in color (a generalization). I've found dialing down the over-all color level, even on ISF'd settings, decreases the sense of "people and objects as neon signs," - that over-lush glow to everything - to a flatter, more texturally real image. (To analogize: A mat picture vs a glossy photo). Also, I find at least in the case of my plasma, carefully dialing the contrast and brightness beyond ISF settings (but still way below torch mode)also increase the realism of the image. I don't get any obvious loss of detail in high-lights - only a more realistically dynamic image. Etc.
During the calibration I sometimes had the sense the plasma was being made to jump through hoops made for another show animal (CRT), and that it's performance was evaluated solely on how well it mimicked that animal's abilities, rather than how well it displayed it's own. As a "broadcast monitor" my plasma might ultimately fail, due to it being unable to eek the last crumbs of detail in the bottom of the gray scale. BUT, as a "virtual reality machine" it excels. To lift a quote from a Director Of Photography who also notes the compelling images from plasmas: "CRTs are naturally going to score better in tests designed around the performance envelope of the CRT." But "I think we're finally seeing that ISF test performance numbers aren't capable of articulating all of the qualitative properties we're using when we perceive a picture."
So, there are two issues here:
1. How NTSC standards and professional calibration relate to the latest technologies. For consistency sake, the current standards make some logical sense..if strict accuracy to the original transfer is the goal. But when we calibrate every display type we encounter to look more "the same," more like standard broadcast monitors, what might we be loosing? Do we ignore some of the specific strengths of new technologies when we calibrate them around an envelope designed for CRT?
2. How NTSC standards themselves seem to be held as the only measure of how images should look. I assert that more elbow room should be made in the "hall of respectability" for varying notions of picture quality. For many, adherence to NTSC standards may be the end of the story in evaluating a display's worth. But to others adherence to NTSC standards is not, in of itself, good enough to guarantee the most convincing image possible.
3. I'm aware that, as in any trade or endeavor, in the ISF and broadcast technology world there is debate about these subjects. I'd enjoy hearing from the pros just what type of debates are ongoing these days.
Whew. Sorry for the long post. I look forward to being educated or corrected wherever I've gone wrong. And this isn't by any means meant to impugn the idea of ISF calibration, nor the great job many calibrators do in making so many people happier with their display. (I also realize that a full ISF calibration is likely to reap more rewards on CRT-based technology, given there is more to tweak there vs fixed pixel devices). I just want to peak out of the ISF / NTSC box to see what's there
Cheers,