Sony VPL-VW285ES/385ES Thread

[Dylan Seeger]

haha that image is great.

Are these artefacts still there if you turn off HDR so it send SDR Rec709?

Yes, that posterization is there no matter the video output settings.

[lygren]

Have not noticed this reported posterization, but there are certainly some other issues - that luckily is possible to correct via a proper calibration. First, of the six 285´s we´ve calibrated thus far, the gamma response is not in accordance to input. Basically, a steep "S" curve is applied that needs heavy correction. Also, colour space and greyscale needs correction. Before and after calibration has rarely had as much effect as with these new ones from Sony...

[lygren]

Also, average on/off contrast levels for the 285´s are about 12-15k:1, while the two 385´s we´ve calibrated thus far is at about 22k:1. We´re calibrating using a K10A, profiling with a CR spectro (and we also have a PR670 at hand, but as this one does not support <5nm, needed for profiling lasers, we have started using the CR for profiling instead).

[lygren]

...and by the way, EXCELLENT review @Dylan Seeger...!

[Dylan Seeger]

Have not noticed this reported posterization, but there are certainly some other issues...

Posterization is not something you can easily see from a seated distance as this particular artifact is small. The next time you have a 285ES or 385ES playing video. Pause the movie on a close up face shot and walk up close to the screen. You should be  able to make out this artifact if you're standing very close to the screen.

[Dylan Seeger]

For anyone looking to see some of the color limitations on the Sony 4Ks, playback this Spears and Munsil 10 bit gradation quantization pattern. It shows both 8 bit and 10 bit patterns next to one another. You can download it here:

https://drive.google.com/file/d/0B68jIlCvW85gMW5aX21CbU1IOGs/view

Unfortunately banding is pretty excessive on the Sony's. My RS400 does not exibit this banding. The whole point of having the 8 bit next to the 10 bit is to show you how much less banding there is with a 10 bit color gradation. Sorry about the flicker, it's added by my phone. Pay attention to the rotating patterns:

[Dylan Seeger]

So it seems the key is to turn off HDR with a 4K60 10bit 4:2:0 signal to get rid the excessive banding that this type of signal adds to the image. But that kind of defeats the purpose of sending this signal to the projector as 10bit content is almost always going to be HDR in the consumer video world. From the "other" forum:

http://screenshotcomparison.com/comparison/120539

Not the best photo, but you can see what happens with HDR enabled. I'll see if I can get some clearer photos later tonight if I get some time.

[Dylan Seeger]

I still want to see some comparison of actors faces on the 385 vs. the JVC 540/640.  I use the face of the hotel desk worker in the Lucy 4k UHD to compare my Panasonic UB900 and Oppo 203.  The Panasonic was superior in conversion of HDR to SDR (1080P) and it had a much more graduated application of sharpness that looked better.  The Oppo sharpness adjustment almost seemed like it is broken as a single click up was almost too much and introduced noise (or enhanced it too much).

I read one of your comments on the other forum regarding eshift vs 4K SXRD. The 285ES does provide single pixel information on this single pixel black and white test pattern:

Sony VPL-VW285ES/385ES Thread

So in that sense it is "native 4K". The issue I talk about where 4K SXRD not providing full 4K resolution has to do with when color information is added in to the image. This happens (obviously) when we watch real video content. So while it may be able to do black and white just fine on a per pixel basis to do real native 4K, it cannot do that when we add complex color into the image. I'll reference this comparison again:

http://screenshotcomparison.com/comparison/120394

As you can see, 4K SXRD has issues doing color information properly. The result is same colored blotches (posterization) and banding. It does not show whats actually in the source properly as far as color is concerned. The JVC by comparison can simply show more subtle color information than the Sony's can. This issue is always there (HDR or not, 4K or not). So unless you're just watching black and white single line information like those test patterns, I don't think the Sony 4Ks are actually true 4K as color performance goes and this is more important than how it performs on that single pixel test pattern shown above because we all watch video with complex color in it. The 4K60 10bit 4:2:0 issue is a little different in that it makes the banding a LOT worse, to the point where it's obvious to see from a seated distance. The posterization and banding that I reference to is more subtle and is hard to see from a seated distance. Hopefully Sony can address this excessive banding with a 4K60 10bit 4:2:0 image with a firmware update.

Part of the reason I talk about this so much is because these Sony 4K's are supposed to have more resolution than anything else out there (except the JVC DLA-Z1/RS4500) and I feel this artifact combined with the non-defeatable noise reduction takes away a layer of fine detail that almost takes away the benefit of native 4K over 1080p + eshift. If Sony fixed these color issues completely I think these 4K SXRD models would show an obvious advantage in apparent resolution. But from what I've read and been told privately it seems most of these issues are hardware limitations on the 4K panel and panel driver level, so a new 4K SXRD panel and driver is most likely needed. This would be nice to see as Sony has been rocking these mostly unchanged 4K SXRD panels since 2012 when the 1000ES was released. Hopefully we see something new out in the next year or two. .74" 4K SXRD Revision 2.0!

[farsider3000]

I read one of your comments on the other forum regarding eshift vs 4K SXRD. The 285ES does provide single pixel information on this single pixel black and white test pattern:

Sony VPL-VW285ES/385ES Thread

So in that sense it is "native 4K". The issue I talk about where 4K SXRD not providing full 4K resolution has to do with when color information is added in to the image. This happens (obviously) when we watch real video content. So while it may be able to do black and white just fine on a per pixel basis to do real native 4K, it cannot do that when we add complex color into the image. I'll reference this comparison again:

http://screenshotcomparison.com/comparison/120394

As you can see, 4K SXRD has issues doing color information properly. The result is same colored blotches (posterization) and banding. It does not show whats actually in the source properly as far as color is concerned. The JVC by comparison can simply show more subtle color information than the Sony's can. This issue is always there (HDR or not, 4K or not). So unless you're just watching black and white single line information like those test patterns, I don't think the Sony 4Ks are actually true 4K as color performance goes and this is more important than how it performs on that single pixel test pattern shown above because we all watch video with complex color in it. The 4K60 10bit 4:2:0 issue is a little different in that it makes the banding a LOT worse, to the point where it's obvious to see from a seated distance. The posterization and banding that I reference to is more subtle and is hard to see from a seated distance. Hopefully Sony can address this excessive banding with a 4K60 10bit 4:2:0 image with a firmware update.

Part of the reason I talk about this so much is because these Sony 4K's are supposed to have more resolution than anything else out there (except the JVC DLA-Z1/RS4500) and I feel this artifact combined with the non-defeatable noise reduction takes away a layer of fine detail that almost takes away the benefit of native 4K over 1080p + eshift. If Sony fixed these color issues completely I think these 4K SXRD models would show an obvious advantage in apparent resolution. But from what I've read and been told privately it seems most of these issues are hardware limitations on the 4K panel and panel driver level, so a new 4K SXRD panel and driver is most likely needed. This would be nice to see as Sony has been rocking these mostly unchanged 4K SXRD panels since 2012 when the 1000ES was released. Hopefully we see something new out in the next year or two. .74" 4K SXRD Revision 2.0!

That is great information.  I have decided to wait out this round as I don't want to buy into old 4k panels, 13Gbps and mediocre lens (Sony) or fake 4k (JVC).  Maybe Epson will release a true 4k projector at CES in January.  My JVC 4910 still throws a heck of a nice image that is very sharp and I can get the benefit of DTS:X and Atmos by buying 4k UHD discs and down converting to SDR and 1080P (I don't like the eshift buzz sound).

[DavidHir]

That is great information.  I have decided to wait out this round as I don't want to buy into old 4k panels, 13Gbps and mediocre lens (Sony) or fake 4k (JVC).  Maybe Epson will release a true 4k projector at CES in January.  My JVC 4910 still throws a heck of a nice image that is very sharp and I can get the benefit of DTS:X and Atmos by buying 4k UHD discs and down converting to SDR and 1080P (I don't like the eshift buzz sound).

Chad B calibrated my RS4910 unit last weekend and created custom gamma curves for HDR support.  I have to say, I am pretty impressed even at 20ftL in high lamp (my room is a complete black pit).  The wider color gamut also turned out much better than I expected (most dEs under 3).  I've only done limited comparisons, but I prefer it to the "SDR conversion" of the UB900 although as Kris Deering says, there is really no such thing as conversion as it's all tone mapping.  My main concern is my lamp only has 185 hours and as I fade, I will probably run into brightness issues sooner than I would like.  eshift3 still does a pretty good job though (and a bit sharper than my former LS10000 it seems) and I'm very curious to hear about the new eshift5 on the RS440 which I have my eye on. 

Cal report for HDR: https://drive.google.com/open?id=0B0W-jUeeDw5hTTl6NWwxc3ZvVzQ

[AVSCraig]

Chad B calibrated my RS4910 unit last weekend and created custom gamma curves for HDR support.  I have to say, I am pretty impressed even at 20ftL in high lamp (my room is a complete black pit).  The wider color gamut also turned out much better than I expected (most dEs under 3).  I've only done limited comparisons, but I prefer it to the "SDR conversion" of the UB900 although as Kris Deering says, there is really no such thing as conversion as it's all tone mapping.  My main concern is my lamp only has 185 hours and as I fade, I will probably run into brightness issues sooner than I would like.  eshift3 still does a pretty good job though (and a bit sharper than my former LS10000 it seems) very curious to hear about the new eshift5 on the RS440 which I have my eye on.

An RS440 would be about as bright in low lamp as your RS4910 is in high lamp, so there is that consideration - more lumens for HDR ( same with the RS420's on close out ).

[DavidHir]

An RS440 would be about as bright in low lamp as your RS4910 is in high lamp, so there is that consideration - more lumens for HDR ( same with the RS420's on close out ).

Yep, the brightness would be a nice plus for sure. If I hear eshift5 is really not much different than eshift4 on content in motion from regular seating distances, the RS420 closeout would be tempting.

[Dylan Seeger]

The RS4910 is an excellent projector and I still think it holds up well despite being several generations old now, especially if you have an HDFury product which would help when connecting it to a Panny UB900 or Oppo 203.  But I do think the RS400 or RS420 is a definite step up from the RS4910 with HDR though. Eshift4 does do a noticeably better job than Eshift3. We'll have to see what Eshift5 does over 4. The brightness difference between the RS4910 and RS4xx series units is substantial enough to make a noticeable difference on HDR. With that said, I wouldn't be surprised if HDR looks excellent on your RS4910 as ChadB is well regarded and I'm sure he's making the most out what the RS4910 has to offer.

[Andreas21]

Posterization is not something you can easily see from a seated distance as this particular artifact is small. The next time you have a 285ES or 385ES playing video. Pause the movie on a close up face shot and walk up close to the screen. You should be  able to make out this artifact if you're standing very close to the screen.

I just saw the VW760 laser model and it has the same banding and posterization as all the Sony 4K models.

[Dylan Seeger]

I just saw the VW760 laser model and it has the same banding and posterization as all the Sony 4K models.

I can't say that I'm surprised. I will admit it's a little better than it used to be on older 4K SXRD projectors, but it's still not gone. Given that JVC's eshift technology has evolved so much since the first generation, I find that these artifacts are really starting to hold 4K SXRD back from achieving a genuinely noticeable advantage in on-screen resolution. These artifacts combined with the non-defeatable noise reduction wipes away a layer of fine detail that would otherwise give Sony a visible advantage in resolution. I have a feeling that Sony will only fix this issue when JVC has more native 4K competition out. From what I've been told, 4K DiLA does not exhibit these same artifacts, so when it's brought down to cheaper prices and direct comparisons are made showing JVC to be superior in both test patterns and real world video content, I think JVC will have forced Sony's hand into fixing these problems.