Dolby Vision now possible through MP4 Mux.

[RESET_9999]

So I'm now even more convinced that 1, the dynamic metadata are working properly and 2 , the slight color difference with HDR10 is caused by the LLDV bug (because we can't capture tv-led)
*** appleTV files color issues is not related to this though

1- one of these sample RPU is synced properly and the other is out of sync.
you can see the brightness jump at the 4-5second mark.
https://drive.google.com/file/d/19xODV3 ... sp=sharing,
https://drive.google.com/file/d/1PepCc6 ... sp=sharing

2- Remember HDTVTEST Vincent video about tv-led vs player-led ??
well we captured the LLDV signal in PQ of the SM clip, and compared it to the HDR10 BL tonemapped in SDR. The difference is literally the same as in Vincent's video.
https://screenshotcomparison.com/comparison/16173

[HarperVision]

So I'm now even more convinced that 1, the dynamic metadata are working properly and 2 , the slight color difference with HDR10 is caused by the LLDV bug (because we can't capture tv-led)
*** appleTV files color issues is not related to this though

1- one of these sample RPU is synced properly and the other is out of sync.
you can see the brightness jump at the 4-5second mark.
https://drive.google.com/file/d/19xODV3 ... sp=sharing,
https://drive.google.com/file/d/1PepCc6 ... sp=sharing

2- Remember HDTVTEST Vincent video about tv-led vs player-led ??
well we captured the LLDV signal in PQ of the SM clip, and compared it to the HDR10 BL tonemapped in SDR. The difference is literally the same as in Vincent's video.
https://screenshotcomparison.com/comparison/16173

What I find funny and not very scientifically thorough is that he only tried one brand, Oppo I believe, and then blamed the results on LLDV instead of trying multiple sources that could do LLDV, like the Sony X800M2 to see if it was the player itself OR LLDV causing the issue. He seemed to jump to conclusions that just because the Oppo caused this, it must mean ALL LLDV sources do this. That’s not sound science based analysis!

I had the Oppo 203 and the Sony X800M2 here shortly after I was the one who discovered this LLDV trick and the Sony was clearly the better device to use for Dolby Vision on otherwise non-Dolby Vision displays and projectors. Oppo had some sort of color bug that resulted in much less saturated colors. I believed and stated at the time that it was most likely due to the same bug that caused its tone mapping to only use Rec709 color gamut even if you chose BT2020 when using its SDR/BT2020 mode. Something to do with the mediatek chip if I recall.

I bet it also has to do with him not inputting custom Dolby Vision data block info like we do now, so the player then maps the signal to the exact specs of the display and/or projector, just as it does with TV-Led. If he didn’t use the right max nits, etc. then it results in improper tone mapping, meaning compression, which could easily have been the cause of these so called “issues”!

LLDV looks absolutely stunning here in my setup using a BenQ LK990 projector, new AppleTV 4K and an HDFury Vertex2 for the LLDV EDID spoof. I just watched Army of the Dead on Netflix last night in DV on my LK990 and it was probably the single best HDR/DV experience I’ve had!!! Pure eye candy with HDR effects that I’ve not seen on a projector before!!!

[RESET_9999]

Well, it is possible that Dolby updated their SDK on the newer players.

Our test was done on the Sony x700 which is pretty old and the result 100% matches those of Vincent. @Manix should be able to compare all his device's LLDV output but he doesnt own the Sony x800m2.
I don't know about the custom Dolby Vision data block info though.

[HarperVision]

Did you see my updated information in my post? Did anyone use custom DV data block info? Also the tone Mapping (compression) as I said could easily have been the cause for these anomalies because when you do player led you have to manually tell the player what the specs are for the display, unlike TV led which gets the raw info from the player and then uses it’s already known specs internally to do the processing.

It doesn’t sound like Vincent did any of this.

[RESET_9999]

Did you see my updated information in my post? Did anyone use custom DV data block info? Also the tone Mapping (compression) as I said could easily have been the cause for these anomalies because when you do player led you have to manually tell the player what the specs are for the display, unlike TV led which gets the raw info from the player and then uses it’s already known specs internally to do the processing.

It doesn’t sound like Vincent did any of this.

do you have a link / more info about that ?
or please tell manix which settings(DV block) exactly to use on the vertex for the Spears and Munsil video. He can capture without the limitation of his CX by using only the Atomos Shogun which has a 3000nits peak brightness and he'll be happy to test anything you wish.

[Manixx2020beyound]

Did you see my updated information in my post? Did anyone use custom DV data block info? Also the tone Mapping (compression) as I said could easily have been the cause for these anomalies because when you do player led you have to manually tell the player what the specs are for the display, unlike TV led which gets the raw info from the player and then uses it’s already known specs internally to do the processing.

It doesn’t sound like Vincent did any of this.

Yes, but this custom block sinc to the tv max
Values cx so yes it’s tone mapping to the CX
Edid sinc,
I can also do it without CX tx0 sinc & check the results.
Or no tv connected but 3000nit monitor and send the correct dv meta

[HarperVision]

Did you see my updated information in my post? Did anyone use custom DV data block info? Also the tone Mapping (compression) as I said could easily have been the cause for these anomalies because when you do player led you have to manually tell the player what the specs are for the display, unlike TV led which gets the raw info from the player and then uses it’s already known specs internally to do the processing.

It doesn’t sound like Vincent did any of this.

do you have a link / more info about that ?
or please tell manix which settings(DV block) exactly to use on the vertex for the Spears and Munsil video. He can capture without the limitation of his CX by using only the Atomos Shogun which has a 3000nits peak brightness and he'll be happy to test anything you wish.

Have you read the dedicated threads at AVS Forum and AVForums? All the info should be in those threads. Search for user Mark Swift as he is the one who deciphered the DV Data block and then based on that, HDFury included the capability to write Custom DV Data.

Here is what the Tab looks like in the Vertex2’s GUI:


I personally use settings that no one else seems to be using or trying for some reason. Maybe they’re afraid to go so extreme or be ridiculed from others saying “that can’t work”, when it actually does AMAZINGLY WELL!!!

I use a 10,000 nit Custom HDR EDID and also the DV Data Block’s Max Luminance at 10,000 nits. This works awesome and presents the least amount of compression (tone mapping) it seems. The image is so clean and detailed and you can see every nuance in dark scenes and the specular highlights are incredible without being blown out, when paired with the right HDR curve in your display/projector. It appears to work best when watching streaming content like last night with Netflix and iTunes movies with DV are amazing. When used with Infuse app on AppleTV with ripped HDR UHD blurays is the pinnacle though. I can’t even tell they’re not native DV!

I have gone back and forth with a DV Data Block between 1,000 nits and 10,000 nits and for some content 1,000 may be better, but overall I think 10,000 wins. I’ve setup different macros to switch depending on content so it’s easy if I need to.

For extreme high nit sources like The Meg chapter 8, if I set Custom HDR/AVI Tab’s Max Luminance to 10,000 nits and the DV Data Block’s max luminance to 1,000 nits (or use Custom EDID 9 or 10 on Vertex2) then it dynamically tone maps to where there are absolutely no blown out highlights or crushed and dark blacks or low level. The details you see are astonishing. Low APL movies like Blade Runner 2049 and Arrival and The Mandalorian have awesome low level details and brightness, which many complain are too dark.

[HarperVision]

Did you see my updated information in my post? Did anyone use custom DV data block info? Also the tone Mapping (compression) as I said could easily have been the cause for these anomalies because when you do player led you have to manually tell the player what the specs are for the display, unlike TV led which gets the raw info from the player and then uses it’s already known specs internally to do the processing.

It doesn’t sound like Vincent did any of this.

Yes, but this custom block sinc to the tv max
Values cx so yes it’s tone mapping to the CX
Edid sinc,
I can also do it without CX tx0 sinc & check the results.
Or no tv connected but 3000nit monitor and send the correct dv meta

See my reply above about using 10,000 nits. The lower you go, the flatter and less “HDR-Like” the image gets. I noticed once you get down to about 600-800 nits and below the image is way too flat and lifeless. 1,000 is about as low as I’ll go.

Test it with 10,000 nits and tell me what you get and see. I’d love to know what your Atomos shows. I was thinking of getting one myself! A lot of it could have to do with what the display’s characteristics add to it though. My experiences are with a BenQ LK990 and Samsung LSP9T so far. Others like the JVCs act a little differently from what I read.

[HarperVision]

Did you see my updated information in my post? Did anyone use custom DV data block info? Also the tone Mapping (compression) as I said could easily have been the cause for these anomalies because when you do player led you have to manually tell the player what the specs are for the display, unlike TV led which gets the raw info from the player and then uses it’s already known specs internally to do the processing.

It doesn’t sound like Vincent did any of this.

Yes, but this custom block sinc to the tv max
Values cx so yes it’s tone mapping to the CX
Edid sinc,
I can also do it without CX tx0 sinc & check the results.
Or no tv connected but 3000nit monitor and send the correct dv meta

Oh which EDID mode are you using with the HDFury by the way?

If you’re using a Custom EDID 0-99 (red in picture) then it isn’t using the connected display’s EDID at all.

You have to use Automix (green in picture) and add in things you want to add to the display’s native capabilities, like the Custom DV Datablock, HDR10+, etc.

[Manixx2020beyound]

Did you see my updated information in my post? Did anyone use custom DV data block info? Also the tone Mapping (compression) as I said could easily have been the cause for these anomalies because when you do player led you have to manually tell the player what the specs are for the display, unlike TV led which gets the raw info from the player and then uses it’s already known specs internally to do the processing.

It doesn’t sound like Vincent did any of this.

Yes, but this custom block sinc to the tv max
Values cx so yes it’s tone mapping to the CX
Edid sinc,
I can also do it without CX tx0 sinc & check the results.
Or no tv connected but 3000nit monitor and send the correct dv meta

Oh which EDID mode are you using with the HDFury by the way?

If you’re using a Custom EDID 0-99 (red in picture) then it isn’t using the connected display’s EDID at all.

You have to use Automix (green in picture) and add in things you want to add to the display’s native capabilities, like the Custom DV Datablock, HDR10+, etc.

Exactly what I use, did that as well in the past.

[HarperVision]

Exactly what I use, did that as well in the past.

Automix with 10,000 nits?

[Manixx2020beyound]

Exactly what I use, did that as well in the past.

Automix with 10,000 nits?

Yes I always applied the dv meta lum values,
Bt2020 10k comes out unwatchable bright scenes are over blown.
What are watching it on?
Their is nothing to watch 10k on besides the monitor it self which does look beautiful @10k nits
Bc the capture file also becomes 10knits as well
But that does not match the 1000nit hdr version
Also tv led can’t use 10k nits, well fully I’m sure it will utilize the uhdtv max lum values
Forcing 10knit don’t look rite on the CX or any UhD tv out there.

[HarperVision]

Exactly what I use, did that as well in the past.

Automix with 10,000 nits?

Yes I always applied the dv meta lum values,
Bt2020 10k comes out unwatchable bright scenes are over blown.
What are watching it on?
Their is nothing to watch 10k on besides the monitor it self which does look beautiful @10k nits
Bc the capture file also becomes 10knits as well
But that does not match the 1000nit hdr version
Also tv led can’t use 10k nits, well fully I’m sure it will utilize the uhdtv max lum values
Forcing 10knit don’t look rite on the CX or any UhD tv out there.

The key is setting the HDR/AVI Max luminance to 10,000 nits, not just the DV Data Block. This is the crucial setting that makes it not blow out on highlights, along with using the least aggressive HDR curve on your display. On the LK990 it is setting “HDR Brightness” to the -2 setting which has settings of -2, -1, 0, 1, 2 where 0 is the default for 1,000 nits.

Everyone thinks the DV Data block is the only thing that affects the tone mapping and that’s simply NOT true and I’ve proven it time and time again! It is EASILY seen and repeatable when testing scenes like The Meg chapter 8 and Aquaman chapter 6.

[Manixx2020beyound]

Don matter how u configure it.
Results do not match.
Secondly any nit value entry in the dv data block lum section will override the hdr avi lum entry
10k nit in both hdr/avi - dv- data block
Even if u create (avi meta 422 ycbr )in which I tried as well all while sending the hdr info along with the dv still no match there is a bug no denying it it’s clearly visible.
Even used max fall/cll with the dv entry.
Nop no go
It’s also on the same scene on all devices

Also to prove my point
My Shogun can output dv/rgb 8b NO LLDV
So I capture the hdr10 spares munsil with the correct colors.
Then out put in dv to my CX using the vertex
And colors match and are not affected
By tv led.
No way 10k nit work in both lum entries that’s for sure.
Also how are u comparing what your seeing to what’s its soposed be hopefully not by just a visible glance or watching,
That’s not a sure shot way of saying what your seeing matches the original??
Some comparison from the signal your getting would be nice to valadate your actual inputs in the vertex2.
Till then it’s a bug.

[RESET_9999]

I went through all the scenes Vincent HDTVtest compared The LLDV bug is real!!!(at least for the oppo, x700, shield, dune, zidoo, Chromecast) looks like a problem with yellow clipping and a darker overall image(less punch).
The SM clip is an extreme case and is built to expose those kinds of flaws, as you can see in our previous comparisons with p5 p8 and HDR10 , it can look VERY close to the HDR10 / tv-led output. I feel sorry for the Sony TV owners(or whoever getting lldv only) but at the same time, I understand why some people are not bothered by it or won't even notice the difference in most movies.

@HarperVision , my X-550R JVC projector cannot compete with your PJ but did you ever compare lldv-pq and HDR10 madVR HDR to SDR pixel shader (with dynamic target nits and tonemapping)? This looks amazing on my end, like 50x better than passthrough. I bet madvr also looks better on your high-end PJ . https://www.avsforum.com/threads/improv ... r.2954506/

eagle: https://slow.pics/c/oPMdh7qd
Red Flower: https://slow.pics/c/7U5KBDaF
Sky: https://slow.pics/c/JLOQXCUf
reptile: https://slow.pics/c/pIDLTaDN
cactus: https://slow.pics/c/0ShTtaD7
waves: https://slow.pics/c/KAo1kZw5