JVC RS3000

[DLCPhoto]

I understand what you are asking now. According to Kris in his most recent post on the " other forum ", auto seems to pick medium. I don't think there are two levels of DTM going on.

Good.  And I just read that post, and replied with a follow-up question.

As I said there, while I have no reason or basis to doubt Kris's observations, it just doesn't seem to make sense for JVC to have the "Auto" option if it is consistently the same as "Medium."  This would suggest that there is something more going on.

But what that is, I have no clue at this point, and seems like something JVC could quite easily clarify.

[AVSMike]

Interesting.  Many, including myself, had the impression that a single value (Low, Medium or high) was chosen when using Auto mode, based on whatever criteria JVC built into the process, and it stayed that way throughout.

But if the selection when in Auto mode may vary from one frame or scene to another, depending on the incoming signal, that could potentially provide optimal results, regardless of the content.  I'll have to explore using Auto, and see what it does, or doesn't do, with a variety of content.

Still, it would be extremely helpful if JVC put out a PDF of some type, not necessarily a formal 'White Paper' as such, but at least providing some level of detail as far as what the different options do, how they interact, more precise guidelines for their usage, etc.

Thanks for the reply, Craig.

Let me see if I can get a PDF from JVC.

[Manni]

Good.  And I just read that post, and replied with a follow-up question.

As I said there, while I have no reason or basis to doubt Kris's observations, it just doesn't seem to make sense for JVC to have the "Auto" option if it is consistently the same as "Medium."  This would suggest that there is something more going on.

But what that is, I have no clue at this point, and seems like something JVC could quite easily clarify.

The main reason might be to simply take the guesswork out of the user's mind: you choose auto, no need to think about it anymore, because you think the PJ will find the best option... which happens to be medium.

It's like allowing to choose the colorspace with a 12bits input: although you can choose YCC444, YCC422 or RGB, any choice always selects the same as auto, which is YCC422, due to some limitation in the PJ with 12bit content.

I personally think that AUTO should use the metadata to select low, high or medium, except when the metadata is absent or invalid, in which case it should use medium, and then use dynamic tonemapping within that range.

Ideally what you want is to be able to tell the PJ what your peak brightness is, so that the DTM makes the best choices based on that, and not any metadata. This is how madVR (and I think the Radiance Pro) does it, though there are a lot more options as well to get the best results automatically.

[Franin]

Cheap but seems to work -
Dr.Meter LX1330B Digital Illuminance/Light Meter, 0 - 200,000 Lux Luxmeter

Better but more expensive -
Sper Scientific 840020C Lux Light Meter

You will use one off and on as long as you own a projector.

Just ordered the Dr Meter cheers

[Manni]

Let me see if I can get a PDF from JVC.

You could ask them to check the existing PDF, that states that AUTO is using metadata to determine which DTM level to use. See page two of attached, it clearly says that MaxCLL,MaxFALL are being used by AUTO. Then maybe expand a bit on each option, for example what it does when the metadata is invalid (pick medium) or if it uses metadata at all with AUTO.

[Manni]

Looks like it's working as described in the PDF after all, and how it should be: AUTO uses the metadata to select Low, High or Medium when the metadata is valid, or Medium when it isn't. Kris had been fooled by a bug in the Radiance that was changing the metadata.

Hopefully at some point JVC will allow the user to input they peak brightness (like madVR and the Radiance Pro do) so that they don't have to rely to a level that doesn't work for all titles or relies on unreliable metadata to choose a range.

[Dylan Seeger]

I take back my comments about the lens not making that much of a difference on the RS3000. What I ended up doing is overlapping the RS2000's image with the RS3000/NX9's image and by pressing the Hide button I seamlessly swapped the two images instantaneously. Looking at test patterns, I notice not only are the pixels far more delineated, but there seems to be more dynamic range within the image as well. That is, the pixels who need to be black are a much darker shade of black on the NX9 in these test patterns. This gives the impression of more sharpness as adjacent pixels can be portrayed in starker contrast to one another. Really impressed so far. I might just have to find a way to keep this projector!

[AVSMike]

I take back my comments about the lens not making that much of a difference on the RS3000. What I ended up doing is overlapping the RS2000's image with the RS3000/NX9's image and by pressing the Hide button I seamlessly swapped the two images instantaneously. Looking at test patterns, I notice not only are the pixels far more delineated, but there seems to be more dynamic range within the image as well. That is, the pixels who need to be black are a much darker shade of black on the NX9 in these test patterns. This gives the impression of more sharpness as adjacent pixels can be portrayed in starker contrast to one another. Really impressed so far. I might just have to find a way to keep this projector!

The 2000 is a good projector, but that lens really separates the two. Really like my 3000. At one of the shootouts where Kris Deering set up and measured both an RS2000 and an RS3000, Kris commented on how they measured the same, contrast wise. But the 3000 looked like it had better contrast, due to the lens.

[Dylan Seeger]

Dunno if they cherry picked a unit for me, but this RS3000 is measuring in quite a bit higher in contrast than my RS2000.  The least amount of contrast I'm measuring is 38,500:1 (Iris open, max moom). Under the same conditions my RS2000 measures in around 32,000:1 iirc. Dynamic contrast is quite a bit different too. I see about 175,000:1 vs 475,000:1. Fades to black when the Iris clamps down is noticably darker on the RS3000.

[AVSCraig]

Dunno if they cherry picked a unit for me, but this RS3000 is measuring in quite a bit higher in contrast than my RS2000.  The least amount of contrast I'm measuring is 38,500:1 (Iris open, max moom). Under the same conditions my RS2000 measures in around 32,000:1 iirc. Dynamic contrast is quite a bit different too. I see about 175,000:1 vs 475,000:1. Fades to black when the Iris clamps down is noticably darker on the RS3000.

Maybe later production units have improved over time. Or, maybe you got a Golden Sample. 

[Franin]

The RS3000 has a very good lens on it. I get edge to edge sharpness as I did with my Sony 995es lens. 

[AVSMike]

The RS3000 has a very good lens on it. I get edge to edge sharpness as I did with my Sony 995es lens.

Yes, those two lenses have about the same performance. Both are excellent, top quality lenses.

[AVSCraig]

Dunno if they cherry picked a unit for me, but this RS3000 is measuring in quite a bit higher in contrast than my RS2000.  The least amount of contrast I'm measuring is 38,500:1 (Iris open, max moom). Under the same conditions my RS2000 measures in around 32,000:1 iirc. Dynamic contrast is quite a bit different too. I see about 175,000:1 vs 475,000:1. Fades to black when the Iris clamps down is noticably darker on the RS3000.

When can we expect a review of your RS3000 Dylan ? 

[Dylan Seeger]

I've only had the NX9 for a couple weeks now. With that said, I'm hoping it will be finished before the end of this month. But I have no control over when it actually gets published. I'm guessing late January or February. 

[AVSMike]

Some B-stock are coming in.