Don't Call it HDMI 2.0 with Michael Heiss - Page 2 - AVS Forum
Forum Jump: 
Reply
 
Thread Tools
post #31 of 51 Old 10-28-2013, 01:29 PM - Thread Starter
AVS Special Member
 
Scott Wilkinson's Avatar
 
Join Date: Sep 2001
Location: Burbank, CA
Posts: 1,350
Mentioned: 7 Post(s)
Tagged: 0 Thread(s)
Quoted: 358 Post(s)
Liked: 1198
Quote:
Originally Posted by Manni01 View Post

Scott, not only did I think that your interview did a great job at explaining these essential points to a non technical audience, but I believe your reaction to the minor inaccuracies in the content is admirable. If only all journalists or analysts could react the way you do when subjected to friendly and constructive criticism! Thanks and well done.


Many thanks for your kind words! I try to remain open to learning new things, even within my sphere of expertise.


Scott Wilkinson
AVS Editor
Scott Wilkinson is offline  
Sponsored Links
Advertisement
 
post #32 of 51 Old 10-28-2013, 04:23 PM
AVS Special Member
 
madshi's Avatar
 
Join Date: May 2005
Posts: 5,482
Mentioned: 9 Post(s)
Tagged: 0 Thread(s)
Quoted: 70 Post(s)
Liked: 140
Quote:
Originally Posted by Scott Wilkinson View Post


According to Bob Kisor, VP of technology at Paramount, increasing the color gamut from 709 to 2020 increases the amount of data by 125%.

I'm not sure what he means with a data increase. Of course increasing the color gamut means that the colors can be more saturated. Maybe he means that. But what I can say with 100% certainty is that the amount of bits transfered over HDMI does not increase if you increase the color gamut. The gamut is just a definition of how to interpret the data that is transferred. One 4K video frame with 10bit 4:4:4 BT.709 consumes exactly the same number of bits compared to a 4K video frame with 10bit 4:4:4 BT.2020. I know this for a fact.

Quote:
Originally Posted by Scott Wilkinson View Post

Just got off the phone with Joe Kane, who maintains that bit depth is, in fact, directly related to dynamic range, though it has nothing to do with the capabilities of a given display, and it doesn't push the gamut outward or the peak luminance upward.

I guess it's a question of how you define "dynamic range". Some people say that getting finer steps (even with the same peak brightness!) means increased dynamic range. If you define dynamic range like that then ok. Personally, when hearing "increased dynamic range" I think more of recording higher brightness levels than before.
madshi is offline  
post #33 of 51 Old 10-28-2013, 04:48 PM
AVS Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 2,583
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 267 Post(s)
Liked: 173
Quote:
Originally Posted by madshi View Post

I'm not sure what he means with a data increase. Of course increasing the color gamut means that the colors can be more saturated. Maybe he means that. But what I can say with 100% certainty is that the amount of bits transfered over HDMI does not increase if you increase the color gamut. The gamut is just a definition of how to interpret the data that is transferred. One 4K video frame with 10bit 4:4:4 BT.709 consumes exactly the same number of bits compared to a 4K video frame with 10bit 4:4:4 BT.2020. I know this for a fact.
I guess it's a question of how you define "dynamic range". Some people say that getting finer steps (even with the same peak brightness!) means increased dynamic range. If you define dynamic range like that then ok. Personally, when hearing "increased dynamic range" I think more of recording higher brightness levels than before.

My understanding from the podcast with Joe Kane (which was very interesting by the way) is that Joe's position is that to do a good job at reproducing a picture mastered for a DCI gamut, you need at least 12 bits. So while you *could* use only 8 bits, you do need more bits to do a good job if you increase the gamut. So in effect, if you increase the gamut beyond rec709 (especially if you go all the way up to rec2020 which is significantly larger than DCI) you need to increase the bit depth, hence it does have a knock on effect on bandwidth. He even advocates 16 bits for rec-2020!

His position is also that because you create less artefacts with 444 / 12 bits, you might actually need less bandwidth than if you use 420 8bits, as the artefact-less picture can be compressed more efficiently. I was VERY surprised by that one.

He also reckons that P3 (DCI) might be the most likely extended gamut for UHDTV. This means the VW1000/1100 would be one of the few consumer displays able to support it, although only at 420 8 bits as this is the 60p limitations on the Sony displays in 4K.

Looks like we're very far from seeing a UHDTV standard, that's for sure... And most of the 4K displays available today will clearly be obsolete when that happens due to their many limitations, the first one being HDMI 2.0 (which can't handle 120hz at any resolution), the second one being the limited gamut on most displays.
Manni01 is online now  
post #34 of 51 Old 10-28-2013, 06:04 PM
AVS Addicted Member
 
mark haflich's Avatar
 
Join Date: Dec 2000
Location: brookeville, maryland, usa
Posts: 20,246
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 627 Post(s)
Liked: 588
I just listened to the replay of today's JK podcast.missed the pod cast. A great interview by you Scott asking all the right questions.

Assuming everything is held constant, an expanded gamut will require more bits just to maintain the level of banding we presently see. You just can't pull the 2020 gamut off at 8 bits without severe banding and if you add bits the data rate increases.

And by increasing bit length there is room for coding for brightness peaks above our present standard. You need longer bit lengths to increase the dynamic range and you need longer bit lengths to provide more steps within a given range.

You did make one statement about decreasing the chroma subsampling rate and calling that a form of compression. Is decreasing the number of samples taken from a source say in 4:2:2 a form of compression or is it just a complete dropping of the information?

Mark Haflich
markhaflich@yahoo.com
call me at: 240 876 2536
mark haflich is offline  
post #35 of 51 Old 10-28-2013, 06:07 PM
AVS Addicted Member
 
mark haflich's Avatar
 
Join Date: Dec 2000
Location: brookeville, maryland, usa
Posts: 20,246
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 627 Post(s)
Liked: 588
Quote:
Originally Posted by Manni01 View Post

My understanding from the podcast with Joe Kane (which was very interesting by the way) is that Joe's position is that to do a good job at reproducing a picture mastered for a DCI gamut, you need at least 12 bits. So while you *could* use only 8 bits, you do need more bits to do a good job if you increase the gamut. So in effect, if you increase the gamut beyond rec709 (especially if you go all the way up to rec2020 which is significantly larger than DCI) you need to increase the bit depth, hence it does have a knock on effect on bandwidth. He even advocates 16 bits for rec-2020!

His position is also that because you create less artefacts with 444 / 12 bits, you might actually need less bandwidth than if you use 420 8bits, as the artefact-less picture can be compressed more efficiently. I was VERY surprised by that one.

He also reckons that P3 (DCI) might be the most likely extended gamut for UHDTV. This means the VW1000/1100 would be one of the few consumer displays able to support it, although only at 420 8 bits as this is the 60p limitations on the Sony displays in 4K.

Looks like we're very far from seeing a UHDTV standard, that's for sure... And most of the 4K displays available today will clearly be obsolete when that happens due to their many limitations, the first one being HDMI 2.0 (which can't handle 120hz at any resolution), the second one being the limited gamut on most displays.

Joe did not say it would most likely be P3 (DCI). He said he would like it to be P3. Also, I guess Madshi and I were wrong, J.K feels a higher chroma subsampling rate is more important than a longer bit length.. Finally, I learned xyYcc is not a defined color space, it is a way of coding a larger color space which could be defined as a manufacturer such as Sony chooses for the expanded gamut its sets can produce, called by Sony Triluminous.

Mark Haflich
markhaflich@yahoo.com
call me at: 240 876 2536
mark haflich is offline  
post #36 of 51 Old 10-29-2013, 02:23 AM
AVS Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 2,583
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 267 Post(s)
Liked: 173
Quote:
Originally Posted by mark haflich View Post

Joe did not say it would most likely P3 (DCI). He said he would like it to be P3. Also, I guess Madshi and I were wrong, J.K feels a higher chroma subsampling rate is more important than a longer bit length.. Finally, I learned xyYcc is not a defined color space, it is a way of coding a larger color space which could be defined as a manufacturer such as Sony chooses for the expanded gamut its sets can produce, called by Sony Triluminous.

Xvycc is a larger colorspace coded in relation to rec709, so it has a point of reference but it doesn't define a fixed larger colorspace like the others. We found that out when we had this long discussion with madshi and found out that you don't calibrate to xvycc as the extended gamut data is relative to rec709.
This is the reason why I think that's likely going to be what we're going to end up with, at least in the near future, unless they make the (unlikely) decision to choose a gamut that only a minority of displays can support, or decide to wait until we change the light source of our displays.
Xvcolor is certainly not the best option, but it looks like the only realistic option. It's not a manufacturer specific option, as it's been added to the hdmi feature list (on Sony's proposal) under that name and not under xvycc and it's in theory supported by other manufacturers. I have hopes for it to be used as a transport mecanism for a fixed wider gamut as suggested by Madshi during our long exchange on the subject, as this would give flexibility as well as a wider compatibility with a majority of displays.
Joe kane seems to be trying to push for the best and most radical change, which is great, but he didn't come across as being very practical regarding what is likely to happen.
I fear many incremental changes dictated by what the manufacturers can deliver at a realistic cost rather than the ideal implementation that we would all dream for.
Manni01 is online now  
post #37 of 51 Old 10-29-2013, 03:06 AM
AVS Special Member
 
madshi's Avatar
 
Join Date: May 2005
Posts: 5,482
Mentioned: 9 Post(s)
Tagged: 0 Thread(s)
Quoted: 70 Post(s)
Liked: 140
Quote:
Originally Posted by Manni01 View Post

My understanding from the podcast with Joe Kane (which was very interesting by the way) is that Joe's position is that to do a good job at reproducing a picture mastered for a DCI gamut, you need at least 12 bits. So while you *could* use only 8 bits, you do need more bits to do a good job if you increase the gamut. So in effect, if you increase the gamut beyond rec709 (especially if you go all the way up to rec2020 which is significantly larger than DCI) you need to increase the bit depth, hence it does have a knock on effect on bandwidth.

Yes, but the bandwidth increase comes solely from increasing the bitdepth, not from increasing the color gamut.

Quote:
Originally Posted by Manni01 View Post

He even advocates 16 bits for rec-2020!
Quote:
Originally Posted by mark haflich View Post

Assuming everything is held constant, an expanded gamut will require more bits just to maintain the level of banding we presently see. You just can't pull the 2020 gamut off at 8 bits without severe banding and if you add bits the data rate increases.

Joe is missing an important point here, though: Our eyes are most sensitive to brightness, and much less sensitive to color. That's why 4:2:0 is used in the first place. It's mostly banding in the *luma* (brightness) channel which bothers our eyes/brain. Banding in the chroma channels can also be visible, but it's usually much less a problem then luma banding. And for the luma channel it doesn't matter at all whether we use BT.709 or BT.2020 (!!). Some time ago I read that studies found out that the average human eye can see up to about 11bit steps in the luma channel, IIRC (I hope I remember that correctly). So my personal opinion is that 10bit will already be a major improvement, and might already be good enough - regardless of whether we're using BT.709 or BT.2020. Of course 12bit would be great and I'd welcome it. 12bit would be all we would ever need, IMHO. Furthermore studios can use dithering (or error diffusion) before encoding the movies to avoid any remaining banding problems. Some studios are already doing that for the Blu-Ray encodings, some don't, unfortunately. 16bit is overkill. Of course I have no problem at all to use 16bit, it doesn't harm. But IMHO it's not needed, simply because it's the luma channel we need to be worried about most, and the luma channel doesn't care at all about how big the gamut is.

If we do increase the encodable peak brightness a lot then maybe that would be an argument for going up to 16bit, though.

Quote:
Originally Posted by Manni01 View Post

His position is also that because you create less artefacts with 444 / 12 bits, you might actually need less bandwidth than if you use 420 8bits, as the artefact-less picture can be compressed more efficiently. I was VERY surprised by that one.

It's true that encoding with e.g. x264 in 10bit is more efficient than encoding in 8bit. The same might be true for 12bit. The reason for the efficiency increase is that there are less rounding errors in the whole encoding pipeline and smoother gradients which are easier to encode. However, this does *NOT* apply for 4:4:4. Encoding in 4:4:4 requires more bits than encoding in 4:2:0.

Think about it yourself: Luma and chroma channels are encoded separately. Doing 4:2:0 means that the chroma channels are downsampled/downscaled. So basically the Blu-Ray luma channel is encoded with 1920x1080 pixels while the chroma channels are encoded with 960x540 pixels. Now ask yourself: Does encoding a downscaled 960x540 frame need more or less bandwidth for artifact free results? The correct answer is: Less bandwidth. Otherwise DVD would require more bandwidth than Blu-Ray and Blu-Ray would require more bandwidth than 4K, which doesn't make any sense. Anytime you scale a video down, you save bandwidth (and lose quality, obviously). So again: 4:2:0 requires less bandwidth than 4:4:4.

Quote:
Originally Posted by mark haflich View Post

Also, I guess Madshi and I were wrong, J.K feels a higher chroma subsampling rate is more important than a longer bit length.

Does he really say that? Where? Do you have a link + timecode? I kinda doubt he really says that. If he does, most experts will disagree with him.
madshi is offline  
post #38 of 51 Old 10-29-2013, 04:23 AM
AVS Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 2,583
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 267 Post(s)
Liked: 173
Quote:
Originally Posted by madshi View Post

Yes, but the bandwidth increase comes solely from increasing the bitdepth, not from increasing the color gamut.

Joe is missing an important point here, though: Our eyes are most sensitive to brightness, and much less sensitive to color. That's why 4:2:0 is used in the first place. It's mostly banding in the *luma* (brightness) channel which bothers our eyes/brain. Banding in the chroma channels can also be visible, but it's usually much less a problem then luma banding. And for the luma channel it doesn't matter at all whether we use BT.709 or BT.2020 (!!). Some time ago I read that studies found out that the average human eye can see up to about 11bit steps in the luma channel, IIRC (I hope I remember that correctly). So my personal opinion is that 10bit will already be a major improvement, and might already be good enough - regardless of whether we're using BT.709 or BT.2020. Of course 12bit would be great and I'd welcome it. 12bit would be all we would ever need, IMHO. Furthermore studios can use dithering (or error diffusion) before encoding the movies to avoid any remaining banding problems. Some studios are already doing that for the Blu-Ray encodings, some don't, unfortunately. 16bit is overkill. Of course I have no problem at all to use 16bit, it doesn't harm. But IMHO it's not needed, simply because it's the luma channel we need to be worried about most, and the luma channel doesn't care at all about how big the gamut is.

If we do increase the encodable peak brightness a lot then maybe that would be an argument for going up to 16bit, though.
It's true that encoding with e.g. x264 in 10bit is more efficient than encoding in 8bit. The same might be true for 12bit. The reason for the efficiency increase is that there are less rounding errors in the whole encoding pipeline and smoother gradients which are easier to encode. However, this does *NOT* apply for 4:4:4. Encoding in 4:4:4 requires more bits than encoding in 4:2:0.

Think about it yourself: Luma and chroma channels are encoded separately. Doing 4:2:0 means that the chroma channels are downsampled/downscaled. So basically the Blu-Ray luma channel is encoded with 1920x1080 pixels while the chroma channels are encoded with 960x540 pixels. Now ask yourself: Does encoding a downscaled 960x540 frame need more or less bandwidth for artifact free results? The correct answer is: Less bandwidth. Otherwise DVD would require more bandwidth than Blu-Ray and Blu-Ray would require more bandwidth than 4K, which doesn't make any sense. Anytime you scale a video down, you save bandwidth (and lose quality, obviously). So again: 4:2:0 requires less bandwidth than 4:4:4.
Does he really say that? Where? Do you have a link + timecode? I kinda doubt he really says that. If he does, most experts will disagree with him.

Thanks Masdhi, all interesting comments. We don't disagree re the wider colorspace not requiring more bandwidth per se if we are already using 10 or 12 bits. But as you can't really go from 8 bits rec709 to 8 bits rec2020, I guess Scott and Joe are correct to say that if you go for a larger gamut, you NEED more bandwidth. Both statements are accurate:).

Re your last question, I don't have the timecode but Joe does say (towards the end) that he would by far go with 444 rather than getting a higher bit depth if he had to chose. He gives an example of the damage chroma subsampling can do with a still of a gray train with a red stripe, and shows that there is a thin black line between the grey and the red, both horizontally and vertically when 420 is applied. I was very surprised too, because to me this kind of compromise for video content is nothing compared to the ugly banding I see with 8 bit 420, which is due to the bit depth limitation, not the chroma subsampling.

So like you and Mark, I would have expected Joe to say that he would by far chose more bit depth. For video content, that's also what I would choose (more bit depth and 420).

You should listen to the whole podcast, it's interesting and that way you'd be able to comment the content itself instead of our interpretation of it.
Manni01 is online now  
post #39 of 51 Old 10-29-2013, 04:39 AM
AVS Special Member
 
madshi's Avatar
 
Join Date: May 2005
Posts: 5,482
Mentioned: 9 Post(s)
Tagged: 0 Thread(s)
Quoted: 70 Post(s)
Liked: 140
Quote:
Originally Posted by Manni01 View Post

Re your last question, I don't have the timecode but he does say (towards the end) that he would by far go with 444 rather than getting a higher bit depth if he had to chose. He gives an example of the damage chroma subsampling can do with a still of a gray train with a red stripe, and shows that there is a black line between the grey and the red, both horizontally and vertically when 420 is applied. I was very surprised too, because to me this kind of compromise for video content is nothing compared to the ugly banding I see with 8 bit 420, which is due to the bit depth limitation, not the chroma subsampling.

So like you and Mark, I would have expected Joe to say that he would by far chose more bit depth. For video content, that's also what I would choose (more bit depth and 420).

Well, discussing chroma subsampling is kind of "difficult" because there are different ways to subsample chroma (nearest neighbor being the simplest and most stupid way) and there are also different ways to upsample chroma (again nearest neighbor being possible but bad). So whatever artifacts Joe saw in that test image might have also been produced by bad chroma down- oder upsampling methods. But it's hard to say without having access to the specific test pattern. In real life content I find it very hard to see differences between 4:2:0 and 4:4:4. I find banding to be a much more commonly visible and distracting problem.
madshi is offline  
post #40 of 51 Old 10-29-2013, 04:45 AM
AVS Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 2,583
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 267 Post(s)
Liked: 173
Quote:
Originally Posted by madshi View Post

Well, discussing chroma subsampling is kind of "difficult" because there are different ways to subsample chroma (nearest neighbor being the simplest and most stupid way) and there are also different ways to upsample chroma (again nearest neighbor being possible but bad). So whatever artifacts Joe saw in that test image might have also been produced by bad chroma down- oder upsampling methods. But it's hard to say without having access to the specific test pattern. In real life content I find it very hard to see differences between 4:2:0 and 4:4:4. I find banding to be a much more commonly visible and distracting problem.

Yes, I totally agree. Even the "damage" shown by Joe is to me not very significant because there is no way to know when we watch the content whether the train is like that, or if it's a chroma subsampling issue. Also it's probably even less noticeable on a moving picture than on a still. Therefore even if the "damage" shown is due to the chroma subsampling in his example, I don't really care because seeing the downsampled picture would never take me out of the movie unless I could compare it with the non downsampled one or if I knew how the train looks like in reality (in that case without the think black line between the grey and the red) and cared about the discrepancy (if the movie is any good, I probably wouldn't as hopefully my attention would be directed to something else). However, like you, every time I see banding on bluray in low light areas, textures like fog or smoke or on skin tones, it is much more noticeable than that, it completely takes me out of the movie (even if it's a great one) and that's why I hate far more the 8 bit limitation of bluray and would get rid of it first if I had to choose.
Manni01 is online now  
post #41 of 51 Old 10-29-2013, 06:42 AM
Member
 
fumoffu's Avatar
 
Join Date: Aug 2007
Posts: 119
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 9 Post(s)
Liked: 11
Quote:
Originally Posted by madshi View Post

In real life content I find it very hard to see differences between 4:2:0 and 4:4:4. I find banding to be a much more commonly visible and distracting problem.

Some time I hate learning these types of things because I begin seeing them. I would agree that it is unlikely to know you are seeing these artifacts in "natural" image captured by cameras.

I was watching Disney's "The Many Adventures of Winne the Pooh" with my kids. This animation is characterizes with thick black lines and solid colors. I was noticing some odd color ringing around some of the lines in some situation. I need to look into this a bit further but I am suspicions of the chroma sub-sampling.

I am not wondering if traditional animation is a common case where the 4:2:0 artifacts can easily be seen?
fumoffu is offline  
post #42 of 51 Old 10-29-2013, 06:48 AM
AVS Special Member
 
madshi's Avatar
 
Join Date: May 2005
Posts: 5,482
Mentioned: 9 Post(s)
Tagged: 0 Thread(s)
Quoted: 70 Post(s)
Liked: 140
Quote:
Originally Posted by fumoffu View Post

I am not wondering if traditional animation is a common case where the 4:2:0 artifacts can easily be seen?

Yes, with animation 4:2:0 limitations can sometimes be more visible than with filmed content. But again, 4:2:0 artifacts can have a multitude of causes, from bad encoding to bad chroma upsampling up to the "chroma upsampling error". Color ringing is a sign of either bad encoding or bad chroma upsampling. The one key disadvantage of 4:2:0 is that chroma resolution is lower, so chroma detail can appear blurred. But that's it. It is possible to do 4:2:0 without ringing or similar artifacts.
madshi is offline  
post #43 of 51 Old 10-29-2013, 09:50 AM
AVS Addicted Member
 
mark haflich's Avatar
 
Join Date: Dec 2000
Location: brookeville, maryland, usa
Posts: 20,246
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 627 Post(s)
Liked: 588
J.K. says he shows visitors to his home a picture using 4:4:4 and then one using 4:2:0 and they screm what did you do, it looks horrible in comparison. In the podcast, J.K. discusses how in color bar test patterns, one sees black lines between certain colors when 4:4:4 is not employed. because of the les frequent sampling there is no color to present and the projector projects no color, no color being black. The gray train with red strop because of the lower subsampling shows black outlining the stripe horizontally and vertically at the stripe edges.

Mark Haflich
markhaflich@yahoo.com
call me at: 240 876 2536
mark haflich is offline  
post #44 of 51 Old 10-29-2013, 09:55 AM
AVS Addicted Member
 
mark haflich's Avatar
 
Join Date: Dec 2000
Location: brookeville, maryland, usa
Posts: 20,246
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 627 Post(s)
Liked: 588
Madshi. Could you share your most innermost feelings about being wrong re predicting what J.K would choose in an either or case. I was devastated when I learned I was wrong, Nice of you to question my reporting of what J.K. said A or B, when asked an A or B question. smile.gif

I thought what should I do. I was intellectually devastated. Then I took the cure. A cigar and a Coke Zero.

Mark Haflich
markhaflich@yahoo.com
call me at: 240 876 2536
mark haflich is offline  
post #45 of 51 Old 10-29-2013, 10:13 AM
AVS Special Member
 
Manni01's Avatar
 
Join Date: Sep 2008
Posts: 2,583
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 267 Post(s)
Liked: 173
Quote:
Originally Posted by mark haflich View Post

Madshi. Could you share your most innermost feelings about being wrong re predicting what J.K would choose in an either or case. I was devastated when I learned I was wrong, Nice of you to question my reporting of what J.K. said A or B, when asked an A or B question. smile.gif

I thought what should I do. I was intellectually devastated. Then I took the cure. A cigar and a Coke Zero.

I don't think Madshi would (or should) be devastated. It's not because Madshi predicted the wrong answer that Joe Kane was right to answer what he did. Unless you believe Joe Kane is some kind of god who is always right.

And I don't think he questioned your reporting out of distrust, only out of disbelief smile.gif.

I disagreed with quite a few of Joe's statements in this podcast, but that doesn't make me wrong or him right either (or vice-versa), as much of what he said were his opinions, conjectures, beliefs, hopes, but not necessarily facts.

He was making very few predictions and answered very few questions about the future. He merely stated what he was pushing for, which is what made it interesting, whether one agrees with him or not, and whether it has good chances to happen or not.

As it's difficult to know if Joe was holding back because he was trying to remain accessible to the podcast general audience, I would really like to hear him and Madshi have a 100% technical chat. Like a hard core geek talk, not the gentle soft core of the podcast, where 90% of the time is used to define concepts we already know about. I think we would learn a lot more smile.gif.
Manni01 is online now  
post #46 of 51 Old 10-29-2013, 10:31 AM
AVS Special Member
 
madshi's Avatar
 
Join Date: May 2005
Posts: 5,482
Mentioned: 9 Post(s)
Tagged: 0 Thread(s)
Quoted: 70 Post(s)
Liked: 140
Quote:
Originally Posted by mark haflich View Post

J.K. says he shows visitors to his home a picture using 4:4:4 and then one using 4:2:0 and they screm what did you do, it looks horrible in comparison.

Which downsampling algorithm was used to create the 4:2:0 image? And which upscaling algorithm used to bring it back to 4:4:4? I bet he didn't say. Maybe he doesn't even know.

Joe seems to like artificial test patterns a lot. I've a different point of view there. Artificial test patterns can be very useful to draw some specific conclusions. But what really counts is real movie content. And there 4:4:4 and 4:2:0 is usually a wash. While banding is a problem I'm seeing in many of today's Blu-Ray releases.

Quote:
Originally Posted by mark haflich View Post

Could you share your most innermost feelings about being wrong re predicting what J.K would choose in an either or case.

Haha! I was surprised. So surprised that I found it hard to believe. That's why I questioned it. I'm sorry about doubting you, Mark... smile.gif Recently I find I quite often disagree with Joe.
madshi is offline  
post #47 of 51 Old 10-29-2013, 11:22 AM
AVS Addicted Member
 
mark haflich's Avatar
 
Join Date: Dec 2000
Location: brookeville, maryland, usa
Posts: 20,246
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 627 Post(s)
Liked: 588
Quote:
Originally Posted by Manni01 View Post


As it's difficult to know if Joe was holding back because he was trying to remain accessible to the podcast general audience, I would really like to hear him and Madshi have a 100% technical chat. Like a hard core geek talk, not the gentle soft core of the podcast, where 90% of the time is used to define concepts we already know about. I think we would learn a lot more smile.gif.

As Darin will attest, it is very very hard to get Joe to engage in such a conversation. He really avoids that hard core technical stuff. He is more evangelistic in nature rather than trying to prove beyond any doubt the existence of god. smile.gif

I remember the old day when the perfect world in video that Joe was trying to obtain was a gray scale calibration of projectors to d65.

And the sea parted and we got that and it was due to Joe's crusade. Many of us anointed him with saint hood after that.

Mark Haflich
markhaflich@yahoo.com
call me at: 240 876 2536
mark haflich is offline  
post #48 of 51 Old 10-29-2013, 11:56 AM - Thread Starter
AVS Special Member
 
Scott Wilkinson's Avatar
 
Join Date: Sep 2001
Location: Burbank, CA
Posts: 1,350
Mentioned: 7 Post(s)
Tagged: 0 Thread(s)
Quoted: 358 Post(s)
Liked: 1198
Quote:
Originally Posted by biliam1982 View Post

Ths again for a great show, Scott!

Was thinking why hasn't HDMI transitioned to some kind of wireless standard?

Is the speeds not fast enough or is it political?


Glad you enjoyed it! There is a wireless standard for HDMI—two of them, in fact. WHDI operates in the 5 GHz range, giving it greater range than WirelessHD, which operates in the 60 GHz range and is therefore limited to in-room transmissions. WirelessHD version 1.1 has a top speed of 25 Gbps; I've reached out to the WHDI folks to ask about its top speed, but I haven't heard back in several days. I'll update this post when I do.


Scott Wilkinson
AVS Editor
Scott Wilkinson is offline  
post #49 of 51 Old 10-29-2013, 12:07 PM
AVS Addicted Member
 
mark haflich's Avatar
 
Join Date: Dec 2000
Location: brookeville, maryland, usa
Posts: 20,246
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 627 Post(s)
Liked: 588
Once again Scott, really great interview. I did a little research and reducing chroma subsampling is indeed data compression because it uses less data than the source information. Unfortunately the type is lossy rather than lossless.

Mark Haflich
markhaflich@yahoo.com
call me at: 240 876 2536
mark haflich is offline  
post #50 of 51 Old 12-03-2013, 05:07 PM
AVS Special Member
 
work permit's Avatar
 
Join Date: May 2001
Location: ny,ny,usa
Posts: 2,574
Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 2 Post(s)
Liked: 18
Quote:
Originally Posted by mark haflich View Post

Once again Scott, really great interview. I did a little research and reducing chroma subsampling is indeed data compression because it uses less data than the source information. Unfortunately the type is lossy rather than lossless.

Exactly. Why not use a lossless compression algorithm instead? HuffYuv has been around for a long time, can be done cheaply on the fly, and results in roughly a 2:1 compression ratio.

Alex
work permit is offline  
post #51 of 51 Old 12-06-2013, 01:27 AM
AVS Special Member
 
madshi's Avatar
 
Join Date: May 2005
Posts: 5,482
Mentioned: 9 Post(s)
Tagged: 0 Thread(s)
Quoted: 70 Post(s)
Liked: 140
Quote:
Originally Posted by Scott Wilkinson View Post

Just got off the phone with Joe Kane, who maintains that bit depth is, in fact, directly related to dynamic range, though it has nothing to do with the capabilities of a given display, and it doesn't push the gamut outward or the peak luminance upward.

Just stumbled over your SMPTE 2013 UHD Symposium report, which directly contradicts what Joe Kane told you on the phone:

http://www.avsforum.com/t/1496765/smpte-2013-uhd-symposium

Quote:
Originally Posted by Scott Wilkinson View Post

Of course, humans can perceive a much wider dynamic and color range than is reproduced on any current screen. In terms of dynamic range, most of the discussion was aimed at increasing peak brightness from the currently common 100 nits (about 30 footlamberts) to as much as 10,000 nits (nearly 3000 fL)! Of course, a full-screen white field at that brightness would sear your retinas, but that's not what the presenters meant—they were talking about small "specular" reflections, such as the sun reflected in a wine glass, as opposed to diffuse reflections from most surfaces. If specular reflections on a video display could be that bright, the entire image would look much closer to what we see in the real world.

That's exactly what I've been saying all along...
madshi is offline  
Reply AVS Forum Podcasts

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