Official Sony XBR HX909 owners' thread (52HX909, 46HX909, HX9, HX900) - Page 41

Easiest way to getting those Lenses is to contact Sony by calling Sony 239.768.7547 and ask for the "Polarizing Lenses" for the Sony TDG-BR100 3D glasses and they will ship it to you 7-10 business days.

But take note that by applying those additional lenses the 3D picture will become darker.

Btw. My question about the RGB settings for Seientific Atlanta 8300 HD seem to have gone unanswered, if anybody can share some input I'll appreciate it

Thanks

Just a heads up i also found the rgb setting on x360. Its under "system settings" and something like video output. I honestly didnt see a difference after changing it but i left it on rgb output anyway.

Should we assume then that sony will be upgrading future glasses? I'm looking to buy some in a week or so. Should I wait?

PLUGE patterns used to set the Brightness control have a blacker-than-black bar. Whiter-then-white test patterns are on some test/setup discs, and most video signal generators.

Yes, sending RGB to the HX909 looks obviously better. It's not a matter of more data, it's about what the TV does with the input signal. Clearly, the HX909 does SOMETHING better when you send RGB compared to YCbCr.

Lets say you want a pint of beer. But you have an 8-ounce glass that's already full. Pouring 8 ounces of beer into a 1 pint glass doesn't mean you have a pint of beer. You still have 8 ounces of beer, but now it's in a 1 pint glass. Same thing with data... RGB is a 1 pint glass that holds 8 ounces of YCbCr data. But the beer tastes the same whether you drink it out of the big glass or the small glass. The images will be the same whether you send YCbCr or RGB... UNLESS the TV does something different with the 2 formats. The HX909 clearly does something diifferent with RGB that makes images look better.

There are exceptions to everything... the HX909 displays are the first I've seen in a LONG time where RGB and YCbCr don't look identical. Most of the time you'll get the best results sticking with YCbCr 4:2:2 all the way to the TV - but checking the other options is always worth doing if somebody (reliable) hasn't already done it.


But if the signal originated as 4:2:0 YCbCr, it would never had had "4:4:4" worth of data in it... it will take more space to store & higher bitrates to transmit in 4:4:4 format, but it has no more resolution/information than the 4:2:0 images on the Blu-ray disc.


Let's say you have 10 pixels encoded in 4:2:0 YCbCr and you convert them to RGB. When you are done, you will have 10 RGB pixels, each with a value between 16 and 235 for red, green, and blue. Pixel #1 from the YCbCr stream converts to pixel #1 in RGB. There will be only 1 set of RGB coordinates for each of the 10 pixels. A formula will be used to convert the YCbCr 4:2:0 data to RGB... 3 coordinates in, 3 coordinates out in a different color space. There is no more information in the 10 RGB pixels than there was in the original 4:2:0 pixels. (this all assumes both signals are in 8-bit space, but converting 8-bit YCbCr to 16-bit RGB wouldn't produce any more data either.


There won't be if the conversions are all done equally well or if both streams of data are processed "equally" because 4:2:2 YCbCr has 16-235 levels for Y, 16-240 levels for Cb, and 16-240 levels for Cr. 4:4:4 also has the same number of levels for each coordinate. 4:2:2 has HALF the color data removed. 4:4:4 has none of the color data removed. Human vision cannot see the difference between 4:2:2 and 4:4:4 provided both images are displayed and "processed" the same. We see about 200x200x200 colors... 8 million. Any more than that and we don't see the transitions. Full range RGB produces 256x256x256 colors, about 16.8 million. All those additional 8 million colors are UNDETECTABLE so they are meaningless. 4:2:0 has even MORE color information removed, but we're still slightly over the 8 million color threshold.

That said... those 8 million colors have to be used correctly or we see contouring or other issues... though contouring is more often a result of losing bits than it is anything else. That can happen in non-robust processing paths. You can lose a whole bit of resolution if you are not careful... then there are not enough steps to make smooth-looking fades without visible contouring.


It doesn't matter how much data you have. What matters is what you can SEE... you can SEE 8 million colors... no more. YCbCr 4:2:0 has enough data to produce more than 8 million colors. Anything more than that accomplishes nothing since we can't see more data.

1-No it doesn't. YCbCr is 8 bits whether it is 4:2:0 or 4:2:2 or 4:4:4 (in our consumer home theater world). RGB is also 8 bits

Y is allocated values of 16-235 within the 256 value range of 8-bit data
Cb and Cr are allocated values of 16-240.

That means YCbCr in all 3 formats can produce 219x224x224 colors.
RGB will be limited to 16-235 for red, green, and blue so you'll get 219x219x219 colors from RGB. YCbCr actually has more color data because more bits are allocated to Cb and Cr.

2-"Create intermediate values" When you convert 10 pixels of YCbCr to RGB you get 10 pixels. There are no more "intermediate values". There are 1920x1080 pixels for red, green, and blue in the TV... just over 6 million for 1 frame. There are 16 million YCbCr pixels per frame regardless of whether they are in 4:2:0 or 4:2:2 or 4:4:4 format. The TV can only display red, green and blue in 16-235 or 0-255 format. There are no more pixels in the image than just over 6 million and those are defined by the 6 million YCbCr or RGB individual pixels worth of 8-bit data.

You would only get "more data" if you changed from 8-bits to 10 or 12 or 16 bits, but you would never see the difference and the panel still works in 8-bit space. You see 8 million colors and no more. If you send billions of colors, it is meaningless because only 8 million colors (or so) can be detected by human vision.

Now... if the processing of 4:2:0, 4:2:2, 4:4:4 and RGB are not all identical, certainly one format or another could look better. That seems to be the case with the HX909. But everything being equal, most of the time, the optimum path would be to maintain the YCbCr 4:2:2 data that comes out of the disc player all the way through the playback chain to have the fewest numbers of conversions - which SHOULD produce the fewest errors.

10 bits, 12 bits and 16 bits really only improves anything in that they make it more difficult to make mistakes that become visible. In 8-bit space that we have for consumer video, if everything is 'perfect' we see no contouring or posterization or anything else. If you ever see contouring or posterization or other digital artifacts, they were PUT there either in the master or by some processing step in the signal chain. They are MISTAKES... make a mistake while using 8-bit data and it will almost always become visible. For example, cable, satellite, and broadcast TV have only 6 or 7 bits of equivalent resolution because even though they START with 8-bit video, they compress it so much that the result is equivalent to about what you'd get from 6 or 7 bits of data.

In 10 bit space, it is more difficult to make a mistake that becomes visible, but it can still happen... you can still get contouring or posterization or visible block artifacts. It's harder to produce, but still possible. 12-bit data has so much "headroom" for errors that it's unlikely human vision could ever detect anything other than a fairly gross error. Video processing needs to happen in 10-bit or higher space because if you, for example, have 8-bits of data like 11111111 and to make the picture right, you have to go 1 step higher... there are no more higher steps in this 8-bit data. If you are in a 10 bit path, your 8 bits of data might look like 0011111111 or 0111111110... more than likely you wouldn't put the additional bits both in the least significant positions. So now you're +1 that you need from some control setting would become 1111111100 or 0111111111 depending on which scheme you are using. It's not that there's "more data" it's that there's more room to work with the data to get it to where you want it to be. And even though the data path may be 10, 12 or even up to 18 bits (Samsung uses 18-bit data paths), we can only see 8 million colors - easily accomplished with 8-bit YCbCr even in 4:2:0 format - if it is done without errors. You would actually round YCbCr DOWN a bit to put it into 8-bit RGB since there are more data bits in 8bit YCbCr than there are in RGB. If you're going to RGB and... AT THE SAME TIME, 10 bits or more... then the rounding errors would be smaller... But even in 8-bit space, the rounding errors SHOULD be so small they fall outside of that 8 million color limit of our vision system - unless a mistake is made - and that's always possible.

I am not a Sony official by any means, but I would like to point out something that seems to be happening here. Sony did this last year if you remember with the XBR10, which had an incredibly short life here in the USA. It was actually a later referred to as a limited production model. I think the HX900/909 series is similar to that in some ways. A very expensive TV with exotic features and limited run. Sony now has the ability to pump out products with incredibly short lifespans, or have successors way sooner than usual as we all have seen recently in the last six months.

There's also the fact that CES is just a few months away. Sony has always released new TV's at CES and I don't expect things to be any different..

Yeah the difference now is that the HX9 is significantly better than BOTH of those models so its in pretty high demand. Also, i dont recall the XBR8 being in short supply and it was certainly high end...

quick question:

Where's the hx909 made in? I heard japan but some say its mexico. which one is it?

Based upon what? And I'm not referring to the "better" part...

People here are constantly complaining about how they are trying to order them but cannot find them available anywhere. There are many potential sales of this model that Sony already lossed simply because they cannot manufacture enough of these sets (seemingly because Sharp cannot provide enough Uv2a panels for them).

If you go back a bit in this thread, you'll see page after page where people are trying to order it but cant find it anywhere...

In the UK forum (avforums.com) you can see people there are having the same issue...

And while this set is better than the XBR8 (i dont know, maybe even you agree with that?) it is not miles ahead of it, but after seeing both enough times its still a noticeable improvement.

I was merely stating that the disappearance of certain HX900/HX909 sizes reminded me of what occurred with the XBR10. I was in no way trying to compare quality of those models, but rather discuss the similarity in lifespan and availability.

Magnified 4x for clarity:


Both of these images are RGB screenshots sourced from a 4:2:0 MPEG encoded video.

The image on the left effectively has no upsampling done to it so even though the image you are looking at is RGB, it only contains 4:2:0 data.

The image on the right has been upsampled to 4:4:4 data, was converted to RGB with 16-bit precision and then downsampled to 8-bit RGB using a dither algorithm.

This is what I meant by saying 4:4:4 RGB rather than simply "RGB".

The original video was a recording of a console game which is a true 4:4:4 source. Compared directly with the original source, the 4:4:4 upsampled video is as close a match as you could hope for - it is meant to be a circular indicator telling you to press the B button.

Upsampling to 4:2:2 in 8-bit produces a middling result. Unfortunately it is difficult to find a renderer that only goes to 4:2:2, but if I use an 8-bit renderer and connect it to a display that only shows 4:2:2 the difference between it and upsampling to 4:4:4 with high precision is obvious.


But YCbCr values do not directly correspond to RGB values. You could end up with results like 45.5 and 46.4 for two pixel values in a gradient both of which would be rounded to 46 creating banding in the image. When you do YCbCr or RGB conversions they need to be done with significantly more precision to avoid introducing banding. When you then have that data with a high amount of precision you can then use dither algorithms and downconvert to whatever output is necessary eg 10-bit. (most displays are not even true 10-bit)


There are not 1920x1080 pixels of data in the source for MPEG data. There are 1920x1080 pixels for luma and 960x540 pixels for chroma.

If you are watching lower resolution video, everything has to be upsampled, which creates intermediate values that require greater than 8-bit precision.


Some panels are capable of 10-bit gradation. In most cases, whatever you send to the display is still going to be processed by it to some degree. You want to give it as much data to work with as possible.

It is possible to do a great job of upsampling in the player in 16-bit and dither down to 8-bit perfectly but then when you send that to the display it is only working on 8-bits of data and any processing it does to the image can only reduce the quality.

If you were to send 10-bit or greater data to the panel at least then it has 10-bits of data to work with and after it has done its own processing you might end up with at least 8-bits effective data being shown.



I also disagree with your claims that we cannot even see 8-bit data. DCI uses higher bit-depths after a lot of testing was carried out and it was found that there were noticeable improvements from using more than 8-bit data. I seem to recall those tests actually finding that even 10-bit data was not enough for some viewers but the number of them was small enough (around 10% if I recall correctly) that it was decided to go with 10-bit instead because a large enough number of viewers did not find 8-bit to be sufficient.

Tried calling them, but they didn't have any idea what this was. Are these replacement glasses or an attachment that goes on the existing glasses?

You have to insist that you want the Polarizing lenses for the 3D glasses, and if he/she plays dumb tell them to please research it further, that's the way I was able to come through.

From my understanding it's just add-on lenses that will snap into the existing glasses.

I'll keep you posted

Do you know if there's a part number associated with this? I called again--the guy seemed to know what it was. Transferred me to the parts dept but they had no clue and their "research" didn't bring up anything. Did you deal with tv support, parts, or ? Thanks.

TV support, the person I spoke to wanted to transfer me and I told him not to, I was put on hold then the same person came back and placed the order for me.
Sorry I don't know the part no.

Good luck I'll keep you posted when and if' I get it

Hello- My 52/909 is being delivered today & I'm all fired up. Does anyone that's in the know have a link regarding optimal viewing settings/tweaks for this panel? Thanks in advance!

Good Day!

Here's a place to start.

http://www.avsforum.com/avs-vb/showthread.php?t=1261223

Long time lurker here just needing to confirm that I mad the right choice. Is it the general consensus that the HX909 is a better set than the NX810 if my main concern is picture quality. I was given a choice between the two through the Sony replacement program and with only $147 difference between them I opted for the HX909. The size difference was not of much importance since I am upgrading from a 50 inch. Are the main differences that the HX909 has locally dimming back lighting where the NX810 does not and the NX810 had the integrated wifi where the HX909 does not.

I'm no expert here, but based on everything that I've read & heard through the grapevine. I think that you made an excellent choice with what's available within Sony's 2010 HDTV lineup. I personally gave up 8" of edge lit panel for the 52/909's full array led w/ local dimming, aka better PQ.

Cheers!

I have had my set on order since August and it still has not arrived. Apparently there has been 2 minor shipments since then but all stock has gone to specialist Sony dealers. (here in oz)

What I am concerned about is Sony suddenly bringing out a HX910/920 with the crappy samsung panel. Nobody wants to be showhorned into a new model that is missing out on the UV2A panel, because it will only be a compromise.

I ordered mine 2 weeks ago through the sxrd exchange program. It is being delivered on Monday. I looked at the shipment history, it came from Long Beach, CA to Chicago O'hare then to my house. I don't know why yours is taking so long.

I prefer the screen treatment of the XBR8 over the monolith so with that measure I would say the XBR8 is better. I've never done a side-by-side so I can't fully say I'm sold on the differences. As far as how one actually measures demand and/or sales, it's fair to say that when BB did have both the 46" and 52" available, feedback on the website was practically nil. Just go into any local metropolitan BB and talk to a sales mananger and ask em straight up what they are selling the most. Once you get past the Samsung-love fest you'll see just how dismal the HX909 penetration has been. Sony has sold boat-loads more of the LX, NX and probably to some degree HX8's compared to the HX909 sets. Feedback on all the relevant sites are thick with postings compared to a relatively shallow going on the HX909's. Sure you can excited about lots of folks complaining about availability of HX909's on a site such as this but these numbers are really squat compared to the over-all size of the market.

That type of screen treatment made the picture look much duller than the glass, and the overall constrast suffers from it more as well, but ok to each his own. I doubt you'd come to that conclusion seeing them side by side in a low-lit room as i have.

When they had both sizes available whatever they had was sold pretty fast and there were not enough manufactured to cater to any more people who wanted them. Im sure alot of those people just got sick of waiting and went for one of the other models you mentioned. Lets not forget that just as with any flagship usually the lower end models sell ALOT more because they are ALOT more affordable...

Its not like Sony sold more XBR8's than XBR6/Z's either...

When I said the demand is high before, i did not mean to say its AS high as the cheaper models as that is never the case. I meant that Sony is just not able to supply enough models to meet the demands for this set.

And as mentioned in an earlier post, its the same problem they had with the XBR10.

Sony has been trying to get more of these UV2A panels from sharp but they are not being supplied with enough of them. If demand was THAT low they wouldnt be trying so hard to get more panels in order to manufacture as many as possible...

+1

I returned an LG 55" LE8500 for the 52" HX9 and couldn't be happier. PQ on the Sony is amazing (to my eyes anyway) and absolutely no banding!

Hello everyone.
My 52hx9 was delivered yesterday.
We've watched Bolt yesterday and the picture was amazing. The blacks are very deep. I think everyone knows the pros of this model.

Today I've checked backlight uniformity and was quite dissappointed. I've chosen the game mode (or you can turn off Dynamic LED) and watched the black screen. There are two quite big clouds near the center. I've attached the image. The photo was taken with 2'' exposure, ISO 400, f/3.5. The TV backlight was set to min. So these clouds a little brighter. I can also see vertical line between 1st and 2nd row of LED zones. This line a little brighter than other "black" areas but less brighter than clouds so it's hard to see on photo.
Then I've filled the screen with white colour. There are issues also. It's not uniform. It's white with some grey clouds. I can clearly see it when I connect PC and move white window with the mouse.
When I watch the movies clouding is impossible to notice because of local dimming and uniformity issue is almost unnoticible at all.
Anyway could you please check your tv for clouding and white colour uniformity.
Thank you and sorry for my mistakes.

Uniformity does not look much better than the HX700 I have. Thank god for local dimming eh?

I suggest you look for real content where this becomes an issue. Or, keep doing the test on TVs in the showroom until you find one perfect one.