According to this article, 4K certainly does have something to offer for plenty of HTs, and you don't have to sit inside of 1.0 SW's for this to be the case:
Here's the chart he developed for someone with 20/20 vision, which means you can resolve 1/60th of a degree:
Bear in mind that this assumes that each pixel is perfectly resolved, which might not always be the case depending on your projector's optics and convergence. Both of those (roughly equivalent to MTF) can make a significant difference in a projector's ability to render single pixel-level detail that's detectable at normal viewing differences. As an example, I did a comparison between a Sim2 HT3000E and a BenQ W20000 (both single-chip DLP projectors) using Kung Fu Panda, and although the BenQ is certainly a nice projector, the Sim2's picture had a level of clarity and depth to small detail that the BenQ lacked; these kinds of factors can have an effect on our perception of resolution differences.
All that being said, my experience with 2K and 4K Digital Cinema projectors has illustrated that 4K can most definitely offer improvements for our projector-based HTs at viewing distances that many people utilize. I'm not sure that the same can be said of flat panel TVs at their current average size. _________________ There are 10 types of people in the world, those who understand binary and those who don't.
Link Posted: Mon Oct 31, 2011 8:12 pm Post subject:
Person99 wrote:
macgyver655 wrote:
Person99 wrote:
Hell, a Greek HT group did a single blind 720p/1080p shoot out between two identical Marantz PJs--only difference was the resoloution--both PJs showing the top half of the frame from the same source and the images stacked one over the other on the same screen at the same time. At viewing distances of 1.5x screen width (16:9) and greater the difference between them was pretty much undetectable (there was no correlation between what people "guessed" was the 1080p PJ and which one it is). People were not accurately able to identify the 1080p PJ until they moved up to between 1.1 and 1.3x (depending upon content and person). And this is with an HT group looking for differences--being told one was 720p and one was 1080p! Imagine how a "normal" casual observers would fair!
Now, imagine the seating distance you would need to tell the diference between 2K and 4K! .6 or .7 screen width? I don't know anyone who sits at that distance. Few sit under 1.0x the width of a scope image in a CIH HT. And the 4K will not really help as we don't perceive resolution changes in the horizontal as well as the vertical.
.
I dont know how this could be a valid conclusion of anything since the projectors have a native resolution and no matter what resolution you send to it, that resolution would be scaled to the native panel. So the only thing you could possibly identify is a difference of the fill(or compressed) pixels of the non native resolution. It would also depend on what the actual resolution of the source disc was and how the player scaled it. But the actual display resolution would always be the same. And the difference between 720p and 1080p pixel fill(or compressed) would be practically unidentifiable.
Its more likely a visual of artifacts and not a resolution difference since there is no resolution difference on screen.
I do not really follow this. They did 1:1 pixel mapping so scaling in the projector was not done. Scaling was done with the HTPC. Secondly downscaling produces far less artifacts that upscaling--in fact, downscaling can be done almost perfectly.
.
Scaling is scaling no matter where it is done. And in the digital domain, any scaling is artifacts. Plus, any difference between the source material and the native display has to be scaled, somewhere. It is what it is, no ifs, ans or buts about it.
And if the source resolution is different then the displays native resolution then each entire frame is rewritten and none if the original data maintained. Digital scaling is completely different then analog scaling. _________________ macgyver566@gmail.com
Link Posted: Mon Oct 31, 2011 8:17 pm Post subject:
Person99 wrote:
Since they used 1080 content as the HD source, the 720p projector was at a disadvantage technically since it had the scaled content. Given this disadvantage, it is even more surprising that from typical viewing distances it was tough to tell them apart. It was only when they got within distances where differences in fill and resolution could be noticed that they could tell the difference.
.
This depends on what the native resolution of the Marantz's were. You didn't give that info. Plus was the source device and source material both 1080P? _________________ macgyver566@gmail.com
Interesting.....so neither projector received a source resolution as to there native resolution, which means both projectors received entire frame scaling. Not surprising what the outcome was. No original data on either one. Each had a completely rewritten image. _________________ macgyver566@gmail.com
Link Posted: Mon Oct 31, 2011 11:07 pm Post subject:
Quote:
Interesting.....so neither projector received a source resolution as to there native resolution, which means both projectors received entire frame scaling. Not surprising what the outcome was. No original data on either one. Each had a completely rewritten image
.
Good point.
There is also another issue with this Greek test. Especially since it was done 5 years ago, is it not correct that most all digital projectors were lucky to fully resolve 600 lines on motion video? If this was true, then, yes, they would look pretty much identical unless they were fed static test patterns.
Quote:
Here's the chart he developed for someone with 20/20 vision, which means you can resolve 1/60th of a degree:
Bear in mind that this assumes that each pixel is perfectly resolved, which might not always be the case depending on your projector's optics and convergence. Both of those (roughly equivalent to MTF) can make a significant difference in a projector's ability to render single pixel-level detail that's detectable at normal viewing differences. As an example, I did a comparison between a Sim2 HT3000E and a BenQ W20000 (both single-chip DLP projectors) using Kung Fu Panda, and although the BenQ is certainly a nice projector, the Sim2's picture had a level of clarity and depth to small detail that the BenQ lacked; these kinds of factors can have an effect on our perception of resolution differences.
All that being said, my experience with 2K and 4K Digital Cinema projectors has illustrated that 4K can most definitely offer improvements for our projector-based HTs at viewing distances that many people utilize. I'm not sure that the same can be said of flat panel TVs at their current average size.
This chart (sorry - I hit the wrong selection) makes perfect sense as long as the signal source in each case IS the native of the display. In other words, the 4K projector is from a 4K source material and not one that is just scaled from a 1080 source. While there would be an improvement with scaled up, it could not, however, match true 4K source material which I think this diagram was based on.
If this was not the case, then why do we bother with 1080 source material - We could just stay with NTSC and scale everything to 4K.
Joined: 22 Mar 2006 Posts: 3910 Location: Comedy Central
Link Posted: Tue Nov 01, 2011 2:07 am Post subject:
Boilermaker wrote:
Quote:
Interesting.....so neither projector received a source resolution as to there native resolution, which means both projectors received entire frame scaling. Not surprising what the outcome was. No original data on either one. Each had a completely rewritten image
.
Good point.
There is also another issue with this Greek test. Especially since it was done 5 years ago, is it not correct that most all digital projectors were lucky to fully resolve 600 lines on motion video? If this was true, then, yes, they would look pretty much identical unless they were fed static test patterns.
Bob
You bring up a good point. That may be why the JVC 4k looked so good with the 4k demo material. Most of the scenes were slow moving or relatively static images.
Link Posted: Tue Nov 01, 2011 3:33 am Post subject:
Boilermaker wrote:
This chart (sorry - I hit the wrong selection) makes perfect sense as long as the signal source in each case IS the native of the display. In other words, the 4K projector is from a 4K source material and not one that is just scaled from a 1080 source. While there would be an improvement with scaled up, it could not, however, match true 4K source material which I think this diagram was based on.
If this was not the case, then why do we bother with 1080 source material - We could just stay with NTSC and scale everything to 4K.
Bob
Sure, to take full advantage of a 4K display, one would need 4K material. Does this mean that 4K projectors displaying upscaled 1080p material would offer no benefit over a 1080p projector? No, although it depends on the individual setup and projector. Just as upscaling from 480i to 720p and 1080p offered improved PQ in terms of pixel structure visibility, sharpness, and apparent detail, so will upscaling from 1080p to 4K. The applications just won't be nearly as widespread - at least not at current flat screen TV sizes. _________________ There are 10 types of people in the world, those who understand binary and those who don't.
Joined: 09 Mar 2006 Posts: 4901 Location: Flower Mound, TX
Link Posted: Tue Nov 01, 2011 7:43 pm Post subject:
HogPilot wrote:
According to this article, 4K certainly does have something to offer for plenty of HTs, and you don't have to sit inside of 1.0 SW's for this to be the case:
Here's the chart he developed for someone with 20/20 vision, which means you can resolve 1/60th of a degree:
This is a theoretical chart and in my experience is close but not exact. As you mentioned, there are additional factors than just resolution. I would also include color accuracy and a number of others.
In my experience and viewing with others and others tests, the hard lines of "benefit clear" in that chart is too aggressive and don't match reality as noted by the Greek test with two effectively identical PJ just with different resolutions.
Also, in the "starts to become appearant" zone, other factors influence the overall perception of picture quality more than the resolution. As an example, I had a good 720p DLP in my theater (fantastic ANSI CR, good resolution, great color accuracy, etc). When the 2nd and 3rd gen of 1080p LCDs were getting popular (the Panny, the Epson, the Mitsu), we did some head to heads. Even at 1.4x screen width, the 720p was almost always considered the overall better picture. _________________ Dave
A train station is where the train stops. A bus station is where the bus stops. On my desk, I have a work station....
Joined: 09 Mar 2006 Posts: 4901 Location: Flower Mound, TX
Link Posted: Tue Nov 01, 2011 8:12 pm Post subject:
macgyver655 wrote:
Scaling is scaling no matter where it is done. And in the digital domain, any scaling is artifacts. Plus, any difference between the source material and the native display has to be scaled, somewhere. It is what it is, no ifs, ans or buts about it.
Your objection was in the context of the Greek results. I still do not understand the objection.
Let me try to put the original stuff simply:
1) 2 PJs one 1080 one 720.
2) Identical source material at 1080
3) Identical HTPC scaling to the native resolution of the PJ where required.
So, on 1080 content, by any measure, the 720p PJ is at a DISADVANTAGE because its content is being scaled. Despite the disadvantage, it was difficult to tell the PJs apart from distances over 1.5x 16:9 screen width.
From this, we can draw 2 conclusions:
1) Downscaling the 1080 to 720 did not introduce any obvious artifacts which detracted from the overall image.
2) The difference is resolution are not significant at that viewing distance.
macgyver655 wrote:
And if the source resolution is different then the displays native resolution then each entire frame is rewritten and none if the original data maintained. Digital scaling is completely different then analog scaling.
I still don't understand your knee jerk reaction to scaling--it makes me believe you have not been around this much. Upscaling and downscaling are two significantly different operations. Upscaling is very very difficult to do extremely well because you can possible create and algorithm that "creates detail" without knowing the type of detail. This is what video upscaling has to do.
Now, downscaling is a much much much easier thing to do. You are analyzing the existing pixels and reducing the resolution by throwing out information. When you throw out info, you understand what is being thrown out and how to represent it as effectively possible and the new resolution. This is very well understood and can be done quite well. It does not generate any visible artifacts in 99.99999% of cases, it only potentially looses detail.
You can try the above for yourself in a semi-unscientific way. Grab a photo you've taken with your phone. Now do a "save as" with the resolution set lower than the original say 60-70% of the original on both axis. Now look at them side by side. YOu will not see "artifacts" in the lower resolution image, at best, you will see a small loss of detail. Should you want to do this experiment more scientifically, make sure you view the images 1:1 mapped on your monitor.
I can sit you in front of a 720p image from a 1080p source scaled by a Lumagen and there is no way you are going to be able to point out an artifact.
Further, downscaling does not ALWAYS result in a loss of detail. Lets take an extreme example the refutes your "any scaling is artifacts" statement: An all white image containing a black horizontal bar in the accross center whose height it 10% of the overall image height.
If that image is in 1080p, the black bar is 108 pixels high and taking up EXACTLY 10% of the height of the image. If that is scaled to 720p, the black bar is 72 pixels high and taking up EXACTLY 10% of the image. Absolutely no artifacts OR loss of detail was introduced with the scaling. In fact, observers could not tell the difference between the two unless they could see pixel spacing.
In fact, this pretty much holds true with minor deviations. For instance, lets say the 1080p black bar was 109 pixels tall the 720p would have to be 72. You would not be able to tell that information was lost (i.e. the bar went from 10.1% the height to 10% of the height).
You want to see a horrible artifact? Watch 24 Hz content at 60Hz. That introduces WAAAAAAAY more artifacts than scaling 1080p/24 to 720p/72. And I bet quite a few on here are watch at 60 Hz because 48 looks terrible on a CRT and 72 softens up too much at HD resolutions where you are pushing the PJ too close to its bandwidth limits. _________________ Dave
A train station is where the train stops. A bus station is where the bus stops. On my desk, I have a work station....
Link Posted: Tue Nov 01, 2011 8:16 pm Post subject:
Person99 wrote:
as noted by the Greek test with two effectively identical PJ just with different resolutions.
.
Well I wasn't going to comment on this but I just have to. The only thing identical between those 2 projectors was the name plate. All the internals would be different. Maybe the lamp is the same but I didn't look it up. _________________ macgyver566@gmail.com
Interesting.....so neither projector received a source resolution as to there native resolution, which means both projectors received entire frame scaling. Not surprising what the outcome was. No original data on either one. Each had a completely rewritten image.
Thanks for finding it Kal, you refreshed my memory as to some details.
You are making a huge assumption here and IIRC an incorrect assumption. You are assuming that 1080i film content was not deinterlaced correctly and scaled but scaled from the 540 field.
1080i film content can be deinterlaced quite easily to its 1080p source--which means the 1080p PJ did have the original source and the 720p has a downscaled image from the original source.
I have to laugh at your "completely rewritten" comments. If you did not "completely rewrite" an image when downscaling, it would look much worse then doing it the way we (us who write image manipulation software) do it. _________________ Dave
A train station is where the train stops. A bus station is where the bus stops. On my desk, I have a work station....
Joined: 09 Mar 2006 Posts: 4901 Location: Flower Mound, TX
Link Posted: Tue Nov 01, 2011 8:21 pm Post subject:
macgyver655 wrote:
Person99 wrote:
as noted by the Greek test with two effectively identical PJ just with different resolutions.
.
Well I wasn't going to comment on this but I just have to. The only thing identical between those 2 projectors was the name plate. All the internals would be different. Maybe the lamp is the same but I didn't look it up.
Technically there are some small differences (i.e. improvements in the chip to the 1080p model). But, for the most part, Marantz swapped the chip out and had a 1080p PJ. Again, any differences between the two PJs where at the disadvantage of the 720p PJ as the 1080p technically included improvements in CR not just resolution.
So again, given the disadvantage (technically speaking) of the 720p PJ, their results are even more surprising. _________________ Dave
A train station is where the train stops. A bus station is where the bus stops. On my desk, I have a work station....
Link Posted: Tue Nov 01, 2011 8:29 pm Post subject:
Person99 wrote:
macgyver655 wrote:
Scaling is scaling no matter where it is done. And in the digital domain, any scaling is artifacts. Plus, any difference between the source material and the native display has to be scaled, somewhere. It is what it is, no ifs, ans or buts about it.
Your objection was in the context of the Greek results. I still do not understand the objection.
Let me try to put the original stuff simply:
1) 2 PJs one 1080 one 720.
2) Identical source material at 1080
3) Identical HTPC scaling to the native resolution of the PJ where required.
So, on 1080 content, by any measure, the 720p PJ is at a DISADVANTAGE because its content is being scaled. Despite the disadvantage, it was difficult to tell the PJs apart from distances over 1.5x 16:9 screen width.
From this, we can draw 2 conclusions:
1) Downscaling the 1080 to 720 did not introduce any obvious artifacts which detracted from the overall image.
2) The difference is resolution are not significant at that viewing distance.
macgyver655 wrote:
And if the source resolution is different then the displays native resolution then each entire frame is rewritten and none if the original data maintained. Digital scaling is completely different then analog scaling.
I still don't understand your knee jerk reaction to scaling--it makes me believe you have not been around this much. Upscaling and downscaling are two significantly different operations. Upscaling is very very difficult to do extremely well because you can possible create and algorithm that "creates detail" without knowing the type of detail. This is what video upscaling has to do.
Now, downscaling is a much much much easier thing to do. You are analyzing the existing pixels and reducing the resolution by throwing out information. When you throw out info, you understand what is being thrown out and how to represent it as effectively possible and the new resolution. This is very well understood and can be done quite well. It does not generate any visible artifacts in 99.99999% of cases, it only potentially looses detail.
You can try the above for yourself in a semi-unscientific way. Grab a photo you've taken with your phone. Now do a "save as" with the resolution set lower than the original say 60-70% of the original on both axis. Now look at them side by side. YOu will not see "artifacts" in the lower resolution image, at best, you will see a small loss of detail. Should you want to do this experiment more scientifically, make sure you view the images 1:1 mapped on your monitor.
I can sit you in front of a 720p image from a 1080p source scaled by a Lumagen and there is no way you are going to be able to point out an artifact.
Further, downscaling does not ALWAYS result in a loss of detail. Lets take an extreme example the refutes your "any scaling is artifacts" statement: An all white image containing a black horizontal bar in the accross center whose height it 10% of the overall image height.
If that image is in 1080p, the black bar is 108 pixels high and taking up EXACTLY 10% of the height of the image. If that is scaled to 720p, the black bar is 72 pixels high and taking up EXACTLY 10% of the image. Absolutely no artifacts OR loss of detail was introduced with the scaling. In fact, observers could not tell the difference between the two unless they could see pixel spacing.
In fact, this pretty much holds true with minor deviations. For instance, lets say the 1080p black bar was 109 pixels tall the 720p would have to be 72. You would not be able to tell that information was lost (i.e. the bar went from 10.1% the height to 10% of the height).
You want to see a horrible artifact? Watch 24 Hz content at 60Hz. That introduces WAAAAAAAY more artifacts than scaling 1080p/24 to 720p/72. And I bet quite a few on here are watch at 60 Hz because 48 looks terrible on a CRT and 72 softens up too much at HD resolutions where you are pushing the PJ too close to its bandwidth limits.
Dave....... I'm going to say this only once. If the source video is anything other then the displays native resolution, none of the displays pixels are of the original content, no matter if it's upscaled or down scaled. The original pixels content is examined..... averaged..... and rewritten to the displays native resolution.
720 source to 720 native display= good pic. 1080 source to 720 display=fake picture
1080 source to 1080 display=good pic. 720 source to 1080 display= fake pic.
Greek test was 1080i source to 720p and 1080p native displays=fake pic on both projectors....... So the 720p seemed to have a better looking fake pic then the fake pic on the 1080p projector.
Oh, and be the way....I understand digital technology better then you may think I do........ _________________ macgyver566@gmail.com
Link Posted: Tue Nov 01, 2011 8:31 pm Post subject:
Person99 wrote:
macgyver655 wrote:
Person99 wrote:
as noted by the Greek test with two effectively identical PJ just with different resolutions.
.
Well I wasn't going to comment on this but I just have to. The only thing identical between those 2 projectors was the name plate. All the internals would be different. Maybe the lamp is the same but I didn't look it up.
Technically there are some small differences (i.e. improvements in the chip to the 1080p model). But, for the most part, Marantz swapped the chip out and had a 1080p PJ. Again, any differences between the two PJs where at the disadvantage of the 720p PJ as the 1080p technically included improvements in CR not just resolution.
So again, given the disadvantage (technically speaking) of the 720p PJ, their results are even more surprising.
Dave... you can't just swap the DLP chip from 1080p to 720p. ALL the electronics are different. Come on man....post some true facts...will ya..... _________________ macgyver566@gmail.com
Link Posted: Tue Nov 01, 2011 9:40 pm Post subject:
Person99 wrote:
This is a theoretical chart and in my experience is close but not exact.
Hence the caveats made by its creator and myself. I completely agree that there's much more that comes into play with respect to perceived PQ than just the display resolution. _________________ There are 10 types of people in the world, those who understand binary and those who don't.
Link Posted: Tue Nov 01, 2011 9:59 pm Post subject:
HogPilot wrote:
I completely agree that there's much more that comes into play with respect to perceived PQ than just the display resolution.
I'm curious as to what these things would be. And I'm being serious here. My bashing the Greek thing was only because it is flawed. But I'd like to know what the perceptions are. _________________ macgyver566@gmail.com
Link Posted: Tue Nov 01, 2011 10:19 pm Post subject:
macgyver655 wrote:
HogPilot wrote:
I completely agree that there's much more that comes into play with respect to perceived PQ than just the display resolution.
I'm curious as to what these things would be. And I'm being serious here. My bashing the Greek thing was only because it is flawed. But I'd like to know what the perceptions are.
Some of the things I discussed earlier - quality of the optics, quality of the display chip, as well as ANSI and on/off contrast, image brightness, image uniformity, screen quality/uniformity...there's probably more I'm forgetting. Not to mention the type of material you're showing - i.e., high-contrast material with abrupt transitions between light and dark (white text on a dark screen, a ship in space, etc) will show pixel structure (which is related to display resolution) more readily than high IRE material.
That being said, I could make out pixel structure on my RS50 on a 84" wide screen sitting at 7' (depending on the material), so there is certainly room for improvement in display resolution depending upon the setup parameters, quality of projector, and visual acuity of the person watching. Plus 4K displays have other benefits (which don't require 4K material to be realized) for even flat panel TV owners (full-resolution 1080p PIP, passive full resolution 3D without higher display refresh rates, etc). _________________ There are 10 types of people in the world, those who understand binary and those who don't.
Link Posted: Tue Nov 01, 2011 10:26 pm Post subject:
HogPilot wrote:
macgyver655 wrote:
HogPilot wrote:
I completely agree that there's much more that comes into play with respect to perceived PQ than just the display resolution.
I'm curious as to what these things would be. And I'm being serious here. My bashing the Greek thing was only because it is flawed. But I'd like to know what the perceptions are.
Some of the things I discussed earlier - quality of the optics, quality of the display chip, as well as ANSI and on/off contrast, image brightness, image uniformity, screen quality/uniformity...there's probably more I'm forgetting. Not to mention the type of material you're showing - i.e., high-contrast material with abrupt transitions between light and dark (white text on a dark screen, a ship in space, etc) will show pixel structure (which is related to display resolution) more readily than high IRE material.
That being said, I could make out pixel structure on my RS50 on a 84" wide screen sitting at 7' (depending on the material), so there is certainly room for improvement in display resolution depending upon the setup parameters, quality of projector, and visual acuity of the person watching. Plus 4K displays have other benefits (which don't require 4K material to be realized) for even flat panel TV owners (full-resolution 1080p PIP, passive full resolution 3D without higher display refresh rates, etc).
So pretty much it's a dollar and cents thing. For the higher quality options you have to spend more money. So your comment on the perceived PQ was based on a parts quality thing more so then a resolution thing. However on a dollar for dollar issue the resolution is would play more into it. Correct? _________________ macgyver566@gmail.com
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum You cannot attach files in this forum You can download files in this forum
Forum FAQ
Search
Memberlist
Photo Albums
Register
Profile
Private Messages
Log in
Joined: 21 Jan 2010
Link
