EDID thread (RTD1619DR players)

Perfect. thank you

 

@Markswift2003 I am trying to follow this discussion with some interest, as I am still using your 1000nit edid. Is there merit to what @Reset_9999 is posting here?
If so, can I have a version of your both original 200nit and 400nit BT2020 edid with HDR10+ enabled so I can compare with your original 1000nit bt2020 with hdr10+ enabled that you made for me a few months back?
My PJ can output over 200nits in certain modes... so thinking 200 or 400 would be the ones to try here?
Thanks!

 

Attached are 100 - 10,000nit LLDV EDIDs with HDR10+

Remember, for this to work you need to match the HDR10 SEI to TMaxPQ in the Dolby Block so MaxCLL and MaxMDL need to the set to the same as the EDID and MaxFALL set to a sensible value (depends what the PJ uses, if any!!). Can't remember if you have one, but the only way I know to do this is with an HDFury device to spoof the SEI (It could be done with a custom gamma and BT.2020 colour profile if the PJ supports but that's hellishly tricky!)

 

Thanks Mark!
On how to use these, I now feel lost. I have previously followed your own instructions of sending the HDR metadata when using the EDID LLDV trick. There as you know we specify 4 values: Max/min luminance and MaxCLL and MaxFALL nits... But it seems you are not referring to that? I thought you just set those appropriately for which of teh EDID files I am testing? I set Max luminance to 1000 before therefore... maxCLL was already 1000 and MaxFALL 400 by default... and I did not alter those.
So what would I do here for a 400 or 200 nit EDID file? Set the max luminance, and maxCLL to 400 to 200? What would I set the MaxFALL to ?

 

Ok, same numbers, different nomenclature - so for the 200nit file, set Max Luminance to 200, MaxCLL to 200 and MaxFall to say 80. I'm kind of sticking my finger in the air for the MaxFALL figure but it's not crucial.

With the 400 set accordingly to 400 and MaxFall to say 120 or 150 ish?

 

Thanks Mark, perfect. I'll be trying this out soon.
As I do not understand the discussion above so much, do you think there is merit to lowering the choice of nit edid like this ?
Previously it was thought 1000 was the sweet spot, and the PJ would do the scaling... would there be any downside to lowering to a 400 or 200 file in your opinion here?

 

the difference of using a proper/closer edid to your actual display brightness is that you use Dolby Vision (authored by a professional colorist) to do the tone mapping instead of your default PJ internal tone mapping curves.
No matter how good your PJ, it cannot compete with Dolby Vision pre-measured L1 and manual L2 trims.

So for example, if your PJ is 200nits with a 400nits edid and the content is 800nits, Dolby Vision will do the tone-mapping from 800 to 400 nits, and then your PJ will tonemap the remaining 200nits to its capabilities.
When you use a 1000nits edid for that same content, because the Dolby engine thinks it is a 1000nits display, it doesnt tone map and your PJ is doing almost all the tone-maping from 800 to your 200nits PJ

 

It depends what does a better job - Dolby Vision was never meant for projectors so although the VSDB spec allows for 100nits, VS10 is not very good at such aggressive tone mapping so it's generally better to hand off to the display. The lowest value of TMaxPQ I've ever seen in a direct display is 300 and that was a Sony which although it had Dolby Vision capability, was never released to the public like that - this was maybe a year or so before Sony actually released Sony Vision, around the time Dolby screwed everything up by repurposing LLDV for video because of Sony - and I can tell you it wasn't very good!

I know what you're saying about the trims and in an ideal world that would be correct, but projectors are a different matter. When I was first helping HDFury et al with all this we had exactly the same thought - set TMaxPQ to 100nits and let the source tone map the lot, but because of the low nit nature of a PJ, if you use an unmodified PQ curve (as you should if no tone map is needed) then the picture is just too dull - you need an aggressive PQ curve and to do that you need to roll off, which means tone mapping.

I'm sure there's an optimum value in there somewhere between the PJs actual performance and 1000nits but I've always found 1000nits to be the sweet spot when dealing with the plethora of DV material out there on a PJ.

But every setup is different, so experimentation is king - you just have to be a bit loose with the rules.

 

Best advice is try it and see - be critical, use content you know well, and decide which looks best - I'd be really interested to hear what you find.

What PJ is it?

 

Personally the best results on the screen on my NZ7 are:
edid LLDV 3000 nits and 'FrameAdapt -WIDE'.
the balance obtained allows me to view everything in LLDV with a very balanced image: black/white!!

 

I disagree. Most of the movies nowadays are graded in a single delivery workflow and DV was designed from day one to tone map SDR or lower HDR nits. They grade HDR and then they do the SDR 100nits / 600 and 1000 trim pass and then they can deliver any trims they want instead of grading the movie again for SDR. All the trims work in interpolation with your TV edid.
FYI ,theres actually a warning in the dolby metafier when you deliver an RPU without an SDR trim and you can export the 100nits trims pass from the RPU and the video will match 100% the SDR 1080p bluray release. There's a workflow for that in the DoVi_Scripts.

The whole point of using LLVD is to use the dynamic metadata and using a 1000nits or worse 3000 nits edid on a 100-200nits display makes absolutely no sense and it's the same as static HDR10 when the content is below those edid.
With DV your are assured that the tone mapping will be as intended, with your PJ/TV, no matter how good it is, the tone mapping will never be as good.

The only thing that can compete with DV tone mapping is madvr envy or a madvr htpc. Anyway, I provided screenshots that show exactly how DV tone map according to different EDID. You guys can do whatever you want

Code:

Copyright (c) 2013-2023 Dolby Laboratories, Inc. All Rights Reserved
12/08/2023/07:08:40.794000000 Metadata: Validation Summary:
=========================
Metadata version: "2.0.5"
Algorithm version: 2 1
Aspect Ratios (Canvas - Image): 1.77778 - 1.77778
Frame Rate: 23.976fps
Mastering Monitor: 1000-nits, P3, D65, ST.2084, Full (ID 20): Bit Depth: 16bit - Diagonal: 42.00in - Application Type: ALL - Color Info: "pq(0.0001,1000) rgb computer p3d65"
Mastering Monitor Validation Test: PASS
Target Displays Validation Test: PASS
Level6 (MaxFALL - MaxCLL): 0.00 - 0.00
Color Encoding: "u16 444 pq(0,10000) rgb computer bt2020"
Color Encoding Validation Test: PASS
Overlapping Shots Validation Test: PASS
Gap between Shots Validation Test: PASS
Negative Shot duration Validation Test: PASS
Per-Frame Data out-of-range Validation Test: PASS
Number of Shots: 62
Frame Range: 0-4566
L2 Trim Count (TargetID, count): (27, 61)
L1 Metadata Validation Test: PASS
L2 Metadata Validation Test: PASS
=========================
12/08/2023/07:08:40.795000000 Metadata: WARNING List of WARNINGS:
=========================
WARNING: shot: UniqueID = 3e168701-db96-44b2-98b7-3b9701ad1180 @[0-71] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = 6966d105-07a9-46d6-8fc7-9aecf402e276 @[72-143] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = 52616116-4877-440a-9129-13e1d85ff30a @[144-215] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = 219539ba-e7b9-4986-8300-75dc983c9057 @[216-287] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = c96c718f-8b79-4cec-ae85-b0dd91be055b @[288-359] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = 975ba8d4-8b18-4ef1-bd81-3bbb2ebce282 @[360-431] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = b1606607-e00f-44ce-af29-8c2f224d6d44 @[432-503] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = d49a8b6f-40fb-4d3b-9058-9fc6c01ddaa3 @[504-575] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = e6ee2b96-49c0-4141-b2be-0c099cb01a30 @[576-647] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = 607b7994-8e5c-4f75-bcc1-a41cac4d2445 @[648-719] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = 81d894ab-9c04-428e-83d8-7afd97e90bd8 @[720-791] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = 299c6761-c6b2-4c08-ab5e-12e2e602fcdd @[792-863] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = 79893cfe-d53d-41af-9443-40d2988ded65 @[864-935] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = fe670dab-c8d7-4881-8544-d9dcab0e144a @[936-1007] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = abecffe3-56dc-4acc-834a-ebbf6c2e58a0 @[1008-1079] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = fe31a8e2-7587-4b2d-a812-4db9098edcd1 @[1080-1151] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = 8c7212af-6487-4bdc-9ca9-5ab512d36fcd @[1152-1223] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = 6d990c5b-1df3-4bdf-9f80-52c922eb7251 @[1224-1295] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = bc1ef0c2-81aa-4e40-9dfa-3b653108e356 @[1296-1367] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = 2d8334e4-ce58-4a74-a9e0-eaedef20c65a @[1368-1439] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = 6a456bab-9c8d-4d97-9a03-f3dc747802d1 @[1440-1511] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = 39baf782-ed7f-4f6b-bb7a-1332c7688b0b @[1512-1583] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = bba9bf91-54b2-46a8-9c0f-805dfa8d0e7e @[1584-1655] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = b44fc178-f103-4969-b344-b1a57d194fee @[1656-1727] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = 536c1b3c-8cbe-4899-a439-b2d96376cdca @[1728-1799] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = bf0a542a-659a-492b-b7b3-06441de917ec @[1800-1871] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = fa0f1ad0-cfee-498d-977f-1a19d6096f18 @[1872-1943] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = d90a3975-d2bc-43ec-bbd7-848efb94a804 @[1944-2015] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = 7ce04ad0-7312-40fd-9055-a5c780951bcb @[2016-2087] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = 7d243220-3d90-4a91-9e77-a3ba9ec2b8c0 @[2088-2159] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = e08122a8-f813-45b4-86d0-8c1b1572aa51 @[2160-2231] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = a7df29e7-c7cf-451b-beea-f77c6689e987 @[2232-2303] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = 6b5a206d-55b1-4da2-bafb-aa71a8c92c33 @[2304-2375] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = ee1c2722-5de8-4ac5-9df9-7b8cbef0384f @[2376-2447] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = ea9839d0-612c-46ce-9d55-eb25519190cc @[2448-2519] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = bb616972-85f3-4528-9edc-b2cb4844d931 @[2520-2591] contains target trim(s) without a 100-nit, Rec.709 trim.
WARNING: shot: UniqueID = 3b5819c2-9a74-4677-82d9-a21aec7a2071 @[2592-2663] contains target trim(s) without a 100-nit, Rec.709 trim.

 

I know the workflow, but what's missing is the fact we're in PQ gamma and that however beautifully tone mapped from 1000nits to 100nits an image may be, if the gamma can't keep up, the image on a projector will be as dull as ditch water - try it - you'll see all the specular highlights but you'll be squinting in the dark.

The only way around it is to be aggressive with gamma, but then you need to at least soft clip and all the highlights disappear again. You could hard clip as soon as you get to the ceiling which gives more headroom, but you still blow the top end.

If you want to use the metadata as intended, you're much better off letting VS10 map to Rec.709. To be fair, it's very very good at that, particularly the new boxes.

It's a shame VS10 won't map to SDR in BT.2020.

 

I'm not saying that you should use a 100nits edid.
I said you should not use an edid higher than 400nits on a PJ and whoever (sorry if it is you) came up with the idea of using a 1000nits edid on a 100-200nits PJ has no idea how DV tone mapping works.

 

Yes I can see that there's definite merit to using say 400nits, but it's always a compromise with a projector - balancing APL and detail and generally with HDR, one suffers at the expense of the other - there's no less of an HDR display than a projector. It's an SDR device at best and most won't even reach anywhere near the level required for diffuse white at 200nits never mind anything further.

From memory the 1000nits originally came from HDFury when we were trying to decode the VSDB for a Sony EDID (The A1) which had TMaxPQ of about 1000nits. Then there was a lot of experimentation and regardless of what the Dolby Vision tone mapping does, 1000nits was always a sweet spot - as I say, a compromise.

Anyway, experimentation is the key to this, after all, to all intents and purposes it's a hack, and certainly not what the guy wiggling his trackball at DaVinci at 2AM ever envisaged being done with his content.

 

BenQ HT4550i

 

I've no experience with BenQ but if the specs are anything go by, with 100% coverage of DCI-P3, its seems a decent beamer. No mention of BT.2020 coverage but there's no way that'd be on the cards for this price point.

As I said, interested to hear what results you get at 200 and 400nits.

 

Mark,
Why is the 1000 nit file in this HDR10+ collection different to the 1000 nit file you made for me a few months back? I would have expected it to be identical? Remember, I'm using your original 1000nit bt2020 file with hdr10+ added, as something is off for me with your 'new' one. Just thinking if the 1000 files do not match, then the 200 and 400 files might be suspect ?

 

Can't remember - If I made one specifically for you that was probably your own EDID with a Dolby Block added on top. The ones in that zip are standard HDMI 2 EDIDs with full audio support plus a BT.2020 Dolby Block and an HDR10+ Block.

You'd have to send me a copy of the earlier EDID for me to look at, I do so many of the things...

 

Ok, I found it.

So no, no material difference between the two EDIDs.

The headers are slightly different - I tend to use my initials (MAS) as the manufacturer PNP and put the manufacture year as the year I do the EDID but I often forget as is the case with the second one.

The only other difference is the display speaker config is slightly different which is of no consequence at all and of course because these bytes are different, so are the checksums. But that's it - both EDIDs are identical in every other respect.

There is no possibility that one could produce different results to the other in any way.

Attached is a comparison report. EDID 1 is the original, EDID 2 the new one.

 

No it was not specially made from anything of mine; I asked you a couple of months back for this as your 'NEW' one gave me issues. It was supposed to just be you original 1000 nit bt2020 file with just the hdr10+ added, as the original file was giving me good results
Here is the file you gave me