INTERACT FORUM
More => Old Versions => JRiver Media Center 31 for Windows => Topic started by: murray on May 05, 2023, 11:06:08 pm
-
I moved from MC30 to MC31 today and want to thank Hendrik and all the others who worked on the development of JRVR MC31 with all the new features....
For me MC30 using JRVR wasn’t quite there for my top end home theatre, JVC NZ9, Studiotec G4 curved microperf 150" scope screen, Isco IIIL on a Cineslide, automated 4 stop side masking, so I had to rely totally on madvr…. For the last few years I wanted to drop madvr PC since development had stopped and the only feature I required to move me into a 20K Envy was the geometry correction for my curved screen. But I was always reluctant paying 30K for just geometry correction.
I’ve only had a couple of hours testing the “new” JRVR MC31 and I’m just loving it, so much so I called my friend Nathan in Australia (I’m in NZ) raving about all the new features!!!
First off the new “adaptive sharpening” is wonderful, I no longer need the three settings I used in madvr for Soft/Sharp/HDR Sharpening, just the one setting is perfect and so much more simple.
BTW Im using a 3080 video card so I can use the highest settings in JRVR.
I’ve done some testing on HDR to SDR and the image is superb and clean, better than madvr HDR to SDR I was using.
SDR is also looking beautiful and the rendering times have certainly dropped compared to madvr.
Now the real “icing on the cake” for me is Geometry Correction. I’ve wanted to correct the bend in my image due to the curved screen for years but had no way of doing so without the top model Envy $$$$$$$. The only way to avoid this is to have the projector dead centre of the screen, well we all know why that isn’t possible. Well JRVR MC31 has given me the fix for just $20, this is just crazy cheap! I’ve only had to use tiny step settings to correct the top and bottom of the screen.
Guys I say thank you, all my Christmases have come at once, I cant wait to play more with this wonderful software. Time to start telling “others” who also have this curved screen distortion issue, and finally a great alternative to madvr Envy and Lumagen.
BTW some of the images look bad as I only used my iphone. Plus there are no images of the scope screen here as thats not important as the "A" lens doesnt create the distortion.
-
Great to hear, Murray! Thanks for taking time to post.
-
Nice! I'm sure JimH will take the $20K :P
JRVR is in a really good place right now as it's:
- Really efficient and will happily run from a low end iGPU like my NUC all the way to my 4090 by just choosing the right performance profiles.
- Feature rich, it even tick's all of Murray's box's.
- Under active development.
- Just works, and on Multiple Platforms.
It reminds me of the development of "Red October". Prior to that was the insane lengths we had to go to with various Direct Show Filters and hacks. Then LAV. Now JRVR. Well done.
-
Great to hear, Murray! Thanks for taking time to post.
Thank you also Jim for providing something really unique, now we can say who needs an Envy or Lumagen....
-
But you are missing out on another Rack Mounted, Brushed Aluminum device with a piercing blue LED! At the end of the day, all these devices are code running on a box... and the HTPC continues.
-
I’ve been playing around with all the new settings on JRVR MC31 again today and I’m just Lovin it!
I’ve tweaked the geometry correction a bit more and reset the angles on the Isco IIIL Anamorphic lens and now I have perfection, best its ever been.
I’m still amazed with the HDR to SDR and it's certainly superior to madvr's HDR to SDR. I haven’t even bothered to play with HDR to HDR yet as I’m replaying many of my HDR films just to wow myself with the image. These shots are just from my iphone off the screen and almost look as if they are from one of my oleds as the results are so good.
The adaptive sharpening is just the “bees knees” it just adds that three dimensional look to my large 150” curved scope screen without getting into any artifacts.
Even the SDR films upscaled to 4K look beautiful and almost as good as the HDR films.
Can someone take a look at the rendering shot and let me know if 21 is acceptable for HDR with the 3080 video card? Most of the SDR films sit round 14.
I’m over the moon!
Sorry cant post the rest of the shots as we are only allowed 6 here......
-
Hi Murray, Just add another post for attaching the additional screen shots.
If you are asking is 21ms is fine, then yes for 23.976fps material as you have up to 41ms before hitting the dropped frame limit (well something before 41ms to be safe).
-
Ok these are the SDR films
-
For some reason I'm getting Monty Python vibes!
-
And here we have some HDR to SDR shots. The one with the stats has the obvious bend corrected in 16:9 with the new geometry correction.
-
And finally some HDR to SDR 16:9 with amazing definition from an old original Vistavision feature.
-
For some reason I'm getting Monty Python vibes!
Sorry what does that mean :)
-
Life of Brian? Holy Grail?
-
Life of Brian? Holy Grail?
Ha ha, no its Ben Hur.
-
And here we have some HDR to SDR shots. The one with the stats has the obvious bend corrected in 16:9 with the new geometry correction.
do you intend to convert HDR to rec709? just checking as I think you have a jvc so DCI-P3 may be a better target for you
-
do you intend to convert HDR to rec709? just checking as I think you have a jvc so DCI-P3 may be a better target for you
I have the JVC NZ9, and I do intend to test it. However Ive been enjoying what Ive been seeing and doing lots of testing up until now.
-
I have the JVC NZ9, and I do intend to test it. However I've been enjoying what Ive been seeing and doing lots of testing up until now.
fwiw I think pretty much every single UHD out there goes beyond 709
there's an interesting thread over at https://www.avsforum.com/threads/checking-for-bt-2020-color-gamut-usage-beyond-p3-or-709-in-real-content.3256532/ which describes how you can check this for yourself if you're interested
-
Interesting thread. FWIW - Just looking at the mastering Meta Data over my BD Collection:
- 1/3rd have a HDR Peak Luminence above 1,000nits
- A bit under 10% having HDR Mastering Colourspace as 2020 (the rest are P3)
I also see that newer displays are now getting well above 1,000 Nits and exceeding P3 coverage (not that I have either), so I hope more mastering moves off 1,000/P3 to full 2020 as I can see at some point the studios shilling "Remastered LOTR Box Set in 8K 2020/10,000nits for the "ultimate viewing experience"". I don't need to buy it a 3rd (or is it 4th) time!
The interesting thing I really need to test more over a range of displays is JRVR's ability to map both Luminance and Gamut for the target display Vs inbuilt functions.
-
The BT2020 (DCI P3) parameter is somewhat confusing. It really indicates the disc was mastered to BT2020, but the monitor used during the mastering process was only capable of DCI P3.
-
It's more likely it is P3 Limited in a 2020 container which is bulk std for most UHD grades.
-
My settings using the 3080 video card if anyone wants to see, so simple and easy to setup.
-
My settings using the 3080 video card if anyone wants to see, so simple and easy to setup.
Hi Murray,
Thanks for your settings. I have an ISCO IIIL A lens as well and will try your geometry settings. I only noticed that at the most left/right of my screen that the lens is not giving me super sharp focus but looks fine in the centre.
Which tone mapping algorithm are you using in the JRVR setup? For me, for best details without blowing my white levels, I'm using BT.2446a. One of the others gives me a brighter picture but at the expense of clipping my whites. My projector has been calibrated to the DCI colorspace so I'm using DCI-P3 as my screen gamut.
I'm using a runco Q750 paired with ISCO IIIL lens on a 103" 2:35 G3 studiotek. Picture looks amazing, it's literally set and forget. Was really sick of madvr continually tweaking.
-
My settings using the 3080 video card if anyone wants to see, so simple and easy to setup.
as noted before, you're outputting rec709 for UHD content which is a smaller gamut that is on disk so you're compressing into the smaller gamut = not ideal.
you're also checking the 10bit output box which I'm not sure is the right choice for a JVC projector, my recollection is that this leads to some further processing in the JVC but I think the correct answer on this depends on the settings used on the projector (memory says you can use 12 bit if in low latency mode otherwise should use 8 bit). The correct setting here will be buried on one of those mega threads on avs.
-
Murray seems to have "Auto" selected for Gamut. Should that not keep the Gamut untouched?
-
Auto is basically BT.709 (for SDR) since information what your screen expects is not really available right now, unless you use an ICC profile which contains that information, which overrides all those settings.
-
Hi Murray,
Thanks for your settings. I have an ISCO IIIL A lens as well and will try your geometry settings. I only noticed that at the most left/right of my screen that the lens is not giving me super sharp focus but looks fine in the centre.
Which tone mapping algorithm are you using in the JRVR setup? For me, for best details without blowing my white levels, I'm using BT.2446a. One of the others gives me a brighter picture but at the expense of clipping my whites. My projector has been calibrated to the DCI colorspace so I'm using DCI-P3 as my screen gamut.
I'm using a runco Q750 paired with ISCO IIIL lens on a 103" 2:35 G3 studiotek. Picture looks amazing, it's literally set and forget. Was really sick of madvr continually tweaking.
Im pleased you are liking JRVR MC31, it certainly is great with all the new updates.
As for your your Isco IIIL out of focus or soft far left and right tells me something is wrong here....
Being an X cinema projectionist for some 35+ years I know my "A" lens stuff!
Setting up an "A" lens isnt easy for most. There are sooooo many adjustments that can make things better or worse.
You must first be 100% satisfied that the 16:9 image is perfectly focused 100% correct from left to right using a test file. If its out the "A" lens will make it worse. The problem is some projectors dont focus perfectly all across the screen to start with, if thats the case the scope will look even worse.... You may have to compromise and soften the centre a tiny amount if its an issue with the proj.
Make sure the beam of light coming out of the proj lens is absolutley centered in the rear lens element of the Isco. Tilt the Isco down to fill the screen top and bottom, make sure the light beam is still dead centre of the rear element of the Isco. Of course all this needs to be done with a test file for scope. The finals is the horizontal and verticals of the Isco.
One needs plenty of time to get this all perfect as its so easy to correct one part and make another worse.
I hope this helps.
See my posts about tone mapping, Ive decided to stay SDR all the way.
-
Auto is basically BT.709 since information what your screen expects is not really available right now, unless you use an ICC profile which contains that information, which overrides all those settings.
I'm on Auto Gamut & I see BT709 when playing SD material and 2020 when playing HDR material (but I'm in HDR mode all the time and have HDR options enabled). This makes sense to me.
-
Hey Murray, given the advice from mattkhan and Hendrik - you may want to test with Calibration --> Screen Gamut changed from Auto to either 2020 or DCI-P3-D65. Your SD material should look the same and your UHD material richer.
-
There are basically only two things we can tell Windows: We're sending SDR BT.709, or, we're sending HDR PQ BT.2020. It does not have any (supported) flagging for other types, so anything else is a manual override, which is not "Auto".
Microsoft is working on improving SDR wide gamut support as well, so maybe this will change in the future.
Note that changing the Screen Gamut setting should always be done in accordance to how your display/projector is configured. The "wrong" setting will just look wrong. If SDR is looking "right" now, then changing the setting will likely not make anything better, unless you also change a projector setting to go hand in hand with this.
-
his projector covers ~90% of DCI-P3 without the filter and all of it with the filter and the it also has a DCI-P3 mode
it means, for HDR content, the optimal setup is going to be converting to DCI-P3 in jrvr and having the PJ in a DCI-P3 colour profile. As Hendrik says, these have to go hand in hand but if the goal is top picture quality then sending it as rec709 is suboptimal.
-
...and I take it the JRVR DCI-P3 option is really DCI-P3 limited in a 2020 wrapper?
-
...and I take it the JRVR DCI-P3 option is really DCI-P3 limited in a 2020 wrapper?
No. Its pure DCI-P3.
-
Thanks for your settings. I have an ISCO IIIL A lens as well and will try your geometry settings. I only noticed that at the most left/right of my screen that the lens is not giving me super sharp focus but looks fine in the centre.
appendix D in http://www.techht.com/Downloads/CineSlideManual-2.3ce.pdf is a good reference guide for lens setup
for geometry correction, I suggest using a laser pointer that can show vertical + horizontal lines mounted on a tripod, display a crosshatch pattern & align your laser lines with this. Now put the lens in line & activate geometry correction, adjust the parameters until you get things aligned. Move the laser pointer up/down/left/right as necessary to ensure alignment across the screen.
Once you have the lines straight then measure the distance between each line and check that each section of the grid is equal across the entire screen, if it's not adjust linearity accordingly until it is. I wouldn't use too fine a crosshatch pattern for this, I think mine is a grid that is something like 10x10 which is sufficient for this purpose.
The correct values will vary according to your setup.
For reference (I have an isco iii xl), mine are attached
-
It's been a fun morning with Murray testing using Wider Colour Spaces on his JVC with JRVR. Test on real footage, he is seeing little if any difference from SDR 709. Looking at what he plays, it tends to be classics and they have very little if any highly saturated colours. I asked him to put on a Mad Max Explosion scene, or Marvel flick to find some intense green or reds but.... that is not his style. I'm sitting at an 80% confidence level he will stay with SDR-709.
-
It's been a fun morning with Murray testing using Wider Colour Spaces on his JVC with JRVR. Test on real footage, he is seeing little if any difference from SDR 709. Looking at what he plays, it tends to be classics and they have very little if any highly saturated colours. I asked him to put on a Mad Max Explosion scene, or Marvel flick to find some intense green or reds but.... that is not his style. I'm sitting at an 80% confidence level he will stay with SDR-709.
Just tried to call you and sorry you are wrong, ;D I’ve decided to stay with 2020 on the proj and on JRVR. The colours are much more intense and saturated, its a lot better!
I think where I was going wrong with hours of testing, I was pausing the film between proj and JRVR changes, this takes about 20secs.... In the photos I was taking of the paused film there didn’t seem to be much change. However when I was about to pack it all in I did the 20sec change from 709 to 2020 and noticed a big change in saturation (out of pause), I’m not sure why when in pause I noticed little to nothing?? Maybe the film in paused isn’t being rendered correctly.
So guys you were right, its better keeping the proj and JRVR all in 2020 for HDR to SDR and straight SDR films. Thank you for the thumbs up!
One last question, is it more correct to stay like this all in 2020 or all in P3?
BTW I do notice with proj and JRVR all set to 2020 I tend to have my posters on theatre view a bit over saturated is this normal?
-
Glad to be wrong and another bump in quality! FWIW, Murray is going to check out P3 settings next.
-
This is a calibration question so the results will just depend on how critical your eye is and how you calibrate the projector. Essentially bt2020 is going to clip if there is content outside the native gamut of the projector. There is not so much content though at this time (though it is reportedly growing slowly).
-
This is a calibration question so the results will just depend on how critical your eye is and how you calibrate the projector. Essentially bt2020 is going to clip if there is content outside the native gamut of the projector. There is not so much content though at this time (though it is reportedly growing slowly).
So are you saying there wouldnt be any clipping if I was to just remain on HDR to SDR 709 for everything?
-
So are you saying there wouldnt be any clipping if I was to just remain on HDR to SDR 709 for everything?
this is just compression which, by definition, is not accurate
nearly all UHDs have content outside 709
relatively few have content outside DCI-P3
it makes no sense to me to buy a high end projector then compress the signal into it for no reason
-
I agree with Matt, set P3 for now on both the PJ and JRVR as that is what your PJ can cover. For the few 2020 mastered disc and JRVR with map it down to P3. When you get your new JVC X9999 that covers more than P3 then using 2020 may make sense.
-
Exactly, I'd agree 100%, this is the way to go.
My display (Panasonic TX-55JZN1508) is very close to 100% DCI-P3, but doesn't reach 80% Rec 2020. Therefore, I have it locked on DCI-P3. I had a professional calibrator here, and he calibrated my video chain for DCI-P3; pre-calibaration of the TV, and 3D LUT for the Radiance Pro. The Radiance Pro is configured to output DCI-P3. It'll convert everything to that colour space. So e.g. Rec 709 content gets mapped into the DCI-P3 colour space and looks exactly like it should, but other content can still use the entire DCI-P3 colour space.
In my opinion it is the best approach to use the colour space for which your display device has the best coverage. For now it is DCI-P3, but it might be something else with the next device.
However, keep in mind that you might get a more accurate calibration when targeting the colour space which is almost completely covered by your display device. In my example, I targeted DCI-P3 deliberately, although the TV covers ~75% Rec 2020, because I wanted the calibration to be as accurate as possible. I wouldn't target Rec 2020 unless my display device were close to 100% Rec 2020 (not ~75% like in my case).
-
I want to say "Thanks" to everyone contributing to this thread. I've personally learnt a lot working with Murray to config up JRVR with him remotely. My lessons learnt are:
- You need to think through the posts you read, as many just say "do this" and many are wrong, or wrong for your context
- Start by thinking about what the capabilities are of your HW (Display, GPU) and from that, look at setting up JRVR to suit
- Stop twiddling knobs and enjoy the content
It sounds obvious, but with madVR there is a lot of focus on the technical settings (queues, flush settings etc etc) and hence the renderer itself becomes the focus. JRVR design is very clever, as the technical components are all "under the hood" and just work. All you need to do as a user is match your output settings to your device, and set the "quality" settings to be within the capabilities of your GPU. Job Done.
Thanks again.
-
I’ve been raving so much about MC31 JRVR over the last week I thought I better go back to madvr and just double check I wasn’t going crazy, nuts or delusional. Sometimes its easy to forget what the old was like without going back on the same material to double check.
So last night I did the test going back to madvr and checking my test material on HDR and SDR….. this was a very worth while thing to do and it only took me 5 minutes to realise I wasn’t going crazy, yes JRVR MC31 is heaps better than madvr for me! I’ve done my tests and comparisons and feel 100% satisfied I’ve made the right move. Image is the most important thing to me and JRVR is better on all material and to have crazy low rendering times is a massive bonus on my 3080 card.
-
Thanks, murray!
-
Hey friends,
I´ve got a problem with the geometry correction in JRiver 31. I have a small horizontal bow at the bottom and cannot fix that. When changing the values, nothing happens.
PC is Windows10, grahiccard is GTX180ti. From there the signal goes to my Denon 4500 and then to my N5. Any clues?
Thanks, Christoph
-
What is your version of MC and are you using JRVR? What setting did you change?
-
I´m using Red October. Settings I changed were Distortion Factor, Distortion Center and Distortion vertical bow. Values were 0,50, just to see if anything happens. But as said, nothing changed. BTW: what means "MC"?
-
MC is Media Center.
Red October has more than one possibility. Standard, JRVR, madVR. This thread is only about JRVR.
-
Aah, okay. MC is JRiver 31, last update. And it´s JRVR.
-
how are you testing? i.e. what test pattern are you using to verify?
are you clicking apply each time you make a change?
are you using jrvr profiles? if so, are you updating the right profile?
-
how are you testing? i.e. what test pattern are you using to verify?
are you clicking apply each time you make a change?
are you using jrvr profiles? if so, are you updating the right profile?
I just use the JRiver Screen. It can be seen there. In the middle of the bottom there´s a bow about 20mmon my 120" screen. Can also be seen in the JVC Test pattern.
And yes, i clicked apply each time, but nothing happened.
I updated the profile Denon. That´s my AVR which is connected to the PC and the N5. The other Profile is the Acer monitor that´s connected to the PC.
-
The correction only applies to video, not the interface. You need to test with actual video playback.
-
Thanks a lot to all of you. Works like charm. No lumagen or envy needed, JRiver does it!
-
Can you guys help me out with where to start? None of the adjustments I use seems to do what I need.
I have a bow top and bottom where the middle of the screen is bowed towards the center. This is caused by my 1.33 anamorphic lens.
-
Please see Hendrik's comment a couple of posts above yours.
-
Can you guys help me out with where to start? None of the adjustments I use seems to do what I need.
I have a bow top and bottom where the middle of the screen is bowed towards the center. This is caused by my 1.33 anamorphic lens.
assume you're playing a pattern through MC not trying to use some built in pattern?
here's what I use for my lens, mine is placed pretty much centrally but with some vertical lens shift, perhaps it gives you an idea of how to approach it
-
Can't really tell from your photo because its at an angle, but from your description start with changing the Curvature Distortion Factor - Vertical (Y). Both positive and negative values are possible.
-
I've got it now thanks. It's hard to distinguish the picture from the projector beam so I have to concentrate on the actual picture to fine tune it.
-
Having spent a lot of time now correcting my screen I want to verify I am doing it correctly. I understand the software and how it works but want to make sure I am using it correctly. Pictures are captures that I am using to try and show what is going on, obviously things look different on my screen through my 1.33 anamorphic lens.
Picture names with explanations -
Before - This is how my screen looks uncorrected with a bow in on top and bottom.
Before Masking - I have been overscanning to make the center of the top and bottom touch the screen with the bow on left / right of top and bottom overscanned into the borders. This results in no black bars around the screen but non-level lines on test pattern.
After - This is with geometry correction, the picture now negative bows outwards. Through the anamorphic lens lines on test patterns are now perfectly "straight", meaning parallel to the screen borders. Even if screen is not perfectly level (it is very close) I chose to square everything with the screen.
After Borders - This is my question, picture is now perfectly "straight" and square but I still have picture that has to be overscanned into the borders. My assumption is that software is correcting the picture but obviously cannot correct the shape of the beam of light coming through the lens. I think when you define screen borders with MadVR / Envy it is creating a new screen inside of the beam and just placing the picture into that new defined area while ignoring everything that is outside (which I believe also costs you resolution depending on how much you cut off the beam). The JRiver correction is fixing the geometry but not cutting off any picture so still need to hide the bow of the light beam, correct?
Does this look / sound like I have done it correctly? My picture is now "straight" but I still have overscan. My screen is 2.35 so I have no issue with overscan, I have no choice with 2.39 films anyway, just want to make sure I have done this correctly.
-
This might make the question easier, it may be difficult to see in the picture but everything is square and level, is the little bit of overscan normal?
-
You should be able to pull that last little bit out I believe. I dont have any of my sides being overscanned into the black masking.
-
You should be able to pull that last little bit out I believe. I dont have any of my sides being overscanned into the black masking.
But your screen is curved correct? Mine is flat.
-
But your screen is curved correct? Mine is flat.
Correct but I certainly dont think that changes anything....
-
Correct but I certainly dont think that changes anything....
The light beam itself coming out of the a lens is bowed, I believe I can fix the picture itself but not the beam. Your curved screen is made for the anamorphic light beam. At least I think all that is accurate? Maybe someone with flat screen can confirm.
-
Here is what uncorrected looks like.
-
Here is what uncorrected looks like.
Correct the bend even more and see what happens you should be able to take that out. Then you might be able to zoom the picture a bit smaller from the proj.
-
My situation is different to most and I have an issue also.
I really need two separate settings of geometry correction for my setup, 99% of others wouldn’t as they leave their “A” lens in place permanently.
As I use a cineslide and the “Isco IIIL” lens moves in front of the proj for scope and out for 16:9, plus a curved scope screen with automated masking 4 stops.
When the “A” lens is out of the light beam for 16:9 I have distortion at the bottom of the screen as the proj is level with the top of the screen, this distortion only happens on a curved screen. A flat screen produces no distortion at the bottom using lens shift.
So I need one set of geometry correction for 16:9, and another set for scope films. This could be done with two profiles but I want to operate these two settings manually (which we cant do in MC at present) so I can build them into my commands using Command Fusion.
-
Correct the bend even more and see what happens you should be able to take that out. Then you might be able to zoom the picture a bit smaller from the proj.
I'm not sure you're understanding me. It's not physically possible for the software to alter the shape of the light beam, all it can do is manipulate the picture inside of the light beam. Nothing I do can make the light beam straight, all I can do is make the picture straight.
Its easier to look at the second line because then the light beam bend is out of play.
First pic is corrected second is not, you can see how the warping fixes it to perfectly straight, there is no other correction left its already straight.
-
Looks fine to me. FWIW, I've a flat screen and I make sure that the edge just sits in the masking as in many setups, things will not be perfectly level and square and that is what the masking is designed to do. Also, everything outside the edges of the picture with geometry correction applied (but in the beam) should be black anyway (or as black as your display can do) and fall into the masking.