Official Lumagen RadianceMini 3D thread. - Page 4

LG; so long as they don't make a machine that watches the films for us, I'm happy.

At first learning about calibration was a challenge and a steep learning curve. Then when I got Chromapure it was like the mists cleared somewhat and then it was 'just' the pain of reitterating the greyscale/gamma points over and over until I was happy. I'd stay up until 4am sometimes just trying to nail the settings, then I'd do a 15 minute CMS adjustment (using my VideoEQ Pro) and that was it. I know I'm still learning new things after nearly 3 years, but while it's technically interesting I've got better things to do with my time (like actually watch films ).

I've found that my preferance (at least in my current room and with the HD350) is to have a 2.3 gamma that drops to 2.2 for the 10 IRE and below. If an autocalibration got me to a flat 2.2 and then all I had to do was change the gamma setting in the M3D +0.1 and tweak the 10 IRE point down 0.1 then I'd consider that fantastic. I'd still do a 'final' run to check everything was where I wanted, but to save the pain of hours of adjustment would be worth it for me. I know I could still have an inaccurate calibration if my sensor wasn't properly setup/off calibration/dark readings done, but at least that would still mean I had some input to ensure a good result.

Having said that the last time I had a Lumagen was before I got Chromapure, so maybe the two combined will make my manual calibrations less painful. I hope so as for now I'll have to setup the new M3D by hand anyway, at least Ricky is 'calibrating' my i1LT against his referance sensor so it'll do until I get some practice and can rent the i1Pro again.

Wish I had the M3D already...

I received the Mini yesterday, set it up, but did not calibrate.

I am very happy so far. I had forgotten how good the Radiance's scaling and sharpness controls are. I didn't buy it for those features but they are well worth the modest (for Lumagen) price of admission.

I was watching some Dirty Jobs on the Discovery Channel. It looked so sharp and likelike (on my RS20) that I would have thought that I was using frame interpolation. I had never before had that impression with this projector. I think I was using too much sharpness for my taste but I could easily see that someone would like the way it looked.

I plan to calibrate this weekend and am really looking forward to that.

The US is back ordered until early March. The Spectracal Lumagen interface is highly automated but there are still things for the calibrator to do. The Chromapure interface should be even more automated but no matter how automated it gets it will not be idiot proof so all end users and pro calibrators can relax or er not relaX.

I got mine this morning: Just like LG I haven't fully calibrated yet (I'm still using my VideoEQ Pro for CMS at the moment) but using it tonight for the vertical stretch on a BluRay and also on a DVD earlier this evening and the upscaling and general sharpness, etc is already winning me over.

Later this week I plan to run through the whole (manual) calibration process so I can remove the VEQ from the loop...so far I'm very happy with the purchase. Even off air HD at 1080i looks better than I've seen before via the Edge plus VEQ.

On physical input 1 if you have say 3 devices with same resolution are you out of luck? I thought I read inputs 1 and 2 have all these resolution choices but can you choose the same resolution more than once on each input? As I understand it the MEMA-D come after that choice.

I just got the XS and realize this is the Mini thread, but I assume they are basically the same as far as 3D and the question would apply to both. I have the Acer H5360 and hooked it up with the Lumagen XS to my PS3 for 3D. I can get an image that looks like it is a 3D image, although it is pretty jumpy. I had been assuming that I could use the Xpand X103 glasses I have, but I couldn't get them to work and from looking at the instructions and the Xpand site it looks like they aren't compatible with anything but IR systems. So, does anybody know what glasses I need for this configuration? Is it just DLP-Link glasses?

Thanks,
Darin

Hi Darin,

I think you need the nVidia glasses for the Acer, as the Acer is using the nVidia proprietary system (and not variations of 3DTV like the Xpand or the JVC). Don't quote me on this, hopefully someone who owns the Acer will chime in, but that's the direction I would explore.

Edit: here is what they say on the Amazon page for the product: "Stereoscopic visuals add depth and feeling to games, movies and photographs. The Acer H5360 projector is certified for leading 3D solutions: NVIDIA® 3D Vision and DLP 3D. NVIDIA® 3D Vision kits transform hundreds of PC games to full stereoscopic 3D. DLP 3D Ready kits with high (120 Hz) refresh rates render a flicker-free 3D experience."

So it looks like it supports DLP-3D too, on top of the nVidia vision kit.

I assume that you have not color calibrated yet? I did over the weekend. I am now seeing very obvious banding on DVDs and on some HD sources. It goes away when I turn the CMS off. There are some other issues as well. I have not gotten to the bottom of this yet and it may simply be user error. But, I had a Radiance in the past and did not experience any issues like this. I am kind of out of ideas and have e-mailed Lumagen's tech support.

keep us updated on this ! I have a XE running on one display and thought about adding a Mini to another...

I've done a very quick colour calibration on my TV, but this suffers from banding anyway, so not the best display to test out the Mini3D (I may as well use it though as the signal passes through it via a HDMI splitter). I'm using the 013011 firmware (I think, the latest as of last Friday). It certainly improved the gamma on that TV as it is low for most of the IREs and too high at the bottom end: I was able to quickly dial in a flat 2.2 bar a lowering to 2.1 at the bottom end for improved shadow detail (was always poor on this older LCD set).

I will be very dissapointed if I spot banding on my projector though (HD350). I'm currently using a VideoEQ Pro which outputs at 10 bits after processing the 8 bit input and I do this specifically to avoid banding (which I've not seen on the VEQ). I wanted the Mini3D to replace both my Edge and VideoEQ Pro with the advantage of having one less box and much better support (than the VEQ anyway)...being able to place it next to my HDMI splitter also meant I could benefit from using it to correct two TVs as well as the projector.

Perhaps it is issues such as using the wrong HDMI levels or some other setting related cause. Please let us know what you find so that we can benefit from the correct settings (if it is settings related).

EDIT: I just checked and the Lumagen firmware page for the Mini3D says that there is a new update 'due on Monday' which is anytime now I guess. It says that with the current 013011 there is 'Video noise with this update on some units' so whether this is your issue but it's possible.

Hopefully this coming update solves my problem. I would not describe what I am seeing as noise. But, here's hoping.

If not, I'll try calibrating again. It took me a little while to get the hang of calibrating again with RGB controls. Plus, I went back and did some touchups. I think if I started from scratch after now being reacquainted with the controls, I might have better luck.

Using the RGB controls is actually pretty easy. I use Calman and can zoom in on each color on the CIE chart as I go. So, I can see which direction I need the color to move.

Things must really be screwed up because my wife, who never says a word about picture quality, told me today that my calibration must have gone wrong. She was watching the Grammys and she said that everyone's face looked wrong.

Ooops . . .

"which outputs at 10 bits after processing the 8 bit input and I do this specifically to avoid banding (which I've not seen on the VEQ)"

In the Radiance you can set it to output 8, 10 or 12 bit. Match it to what is appropriate for your display.

Shawn

Where do you set the output to 12 bit? I've seen 8 and 10, but not 12.

If you are not applying dither it is 12 bit output. You add dither when you are truncating down to 10 or 8 bit output.

Shawn

I need to check this tonight as I couldn't see where to set this output...probably in a menu I haven't yet discovered.

You are calibrating RS40 or RS50, right? I noticed similar thing (very very obvious color banding) with my X3 + Radiance XS when I tried to use "Wide1" color space and have the problem narrowed to 2 different options
1. Could be caused when X3 color saturation is brought down a LOT. Wide1 in X3 is very oversaturated so the colours are tweaked a lot when Wide1 is brought down to REC709. I have a wild theory that this somehow might cause banding due to color tracking issues or something with new JVC DLA
2. My Chroma 5. It seems to read quite considerably different than my Eye One pro. This has been collecting dust a bit too long so it may have to go to Spectracal for recalibration. Dunno if a bad meter could cause banding, not that experienced with this stuff yet.

I now have very nice calibration using Radiance, Calman 4.1 Enthusiast + E1 Pro but I had to use Standard color space where green is slightly undersaturated. But gamma/greyscale are tracking quite perfectly and everything is spot on on colours (best skin tones etc. I've ever seen with any projector) except very slight green undersaturation. CAnnot see the undersaturation on the screen, but knowing it is there pains me a bit. I'd actually propably want to create another CMS profile that is slightly and consistently oversaturated for cartoons etc.

I've been meaning to debug this a bit more (like trying bring wide1 to REC709 with Calman + E1P as I used C5 last I tried) before contacting Lumagen, but I've been super busy and doing a full calibration with E1 Pro a few hours every time. I try to watch a movie every now and then instead of always tweaking

The RS50 went back so I am am keeping the RS20.

I solved the problem. Had I done it more scientifically, I might have some insight about what is going on. As it is, I only have theories.

I had been calibrating the RS20 using the User1 Preset with the CMS turned off. This is probably the widest color space that the RS20 can muster. I changed this to the Cinema 2 preset. If I recall correctly, Cinema 2 is not as wide as User1 with the CMS turned off. Using Cinema 2, I now have no issues. This experience supports your theory that the largest color spaces are too wide to be brought in without banding. But, I don't think that the wide color space is the cause for me.

When I first began correcting my calibration, I measured what I had done earlier. For whatever reason, the Y of some of the colors was way too low - xy positions looked okay. So, I had added way too much correction. I am still not completely sure why I did this but I have a theory. After I zeroed out my existing calibration and started from scratch, Everything was proceeding as it should. When I got to Yellow, I got some really weird readings. I was sure that they were not correct. So, I exited and restarted Calman. When I measured what I had done, R, G and B measured fine but C, M and Y did not. So, I was getting incorrect readings on these colors. I finished the calibration, correcting C, M and Y. Everything now measures correctly. I checked 100% stim and it is also fine.

I went back and looked at the material in which banding was obvious. It was gone. I checked for banding on cable and it was also gone.

So, I think the cause of my banding is that I added too much correction because I was getting bad readings. I am not sure why this happened. It could be that I was not diligent enough in reinitializing the i1pro over time and it was drifting. So I am thinking what happened to me is a measurement issue. The cause of it is probably me.

In your case, I would not accept an undersaturated green. If you saved your "Wide" Calibration, I would go back and remeasure it. Measure it with both the i1pro and the Chroma 5 and attempt to draw some conclusions. If it was truly accurate when you did the calibration, it should still be accurate now (but I always get slight measurement differences when I got back and remeasure).

I don't see any general options to select the bitrate of the output. There is an option in the dither feature to select 10 or 8 bit, but I do not understand that this affects the output. Could be wrong. At one point, I was using 10 bit dither and outputting RGB and the RS20 reported 8 bits. I think that RGB is limited to 8 bits. When I output YCbCr, the RS20 does not report any bitrate.

The manual reports that processing is 10 bit by the calibration pipeline is 12 bit. I assumed that the 12 bit processing was internal.

I read somewhere that:

Hi LG,

Something must be wrong with your driver (or altered by the Radiance), if I select YCbCr 4-2-2 I get 8 bits reported on the JVC. If I select 4-4-4 I get 10 bits (this is with an nVidia 6870 Catalyst 1.1). If I select 4-4-4 with my Sony BD player I get 12 bits on the JVC. There is no Radiance in the chain though. RGB Studio is limited to 4-2-2 (8 bits), but RGB Enhanced is 4-4-4, reported as 10 bits on the JVC.

Do you go through an AVR or direct from the HTPC to the Radiance to the JVC?

I read in the Lumagen documentation that the recommended input colorspace was YCbCr 4-2-2. but I wonder if this also applies for HTPC, as selecting RGB Standard might avoid a double conversion.

I do not know if with recent drivers when YCbCr 4-2-2 is selected in the driver, it is output as such (no conversion), or if the YCbCr 4-2-2 from the bluray is first converted to RGB to be displayed locally, then back to YCbCr to be sent t the HDMI output, which would mean a double conversion. I know this used to be the case, but I'm not sure it is still the case with recent drivers and HDMI present on the graphics adapter.

There may be a double conversion being done on the PC (YCbCr -> RGB -> YCbCr). I don't know. In the Radeon thread here on AVS, someone from Spectracal posted some pictures of greyscale ramps where YCbCr looked the best and showed the least amount of banding. It could just be a good conversion. Maybe there is no conversion.

But, the Radiance is outputting YCbCr directly to the projector. It is the last thing in the chain before the projector (so after the PC and AVR). I checked yesterday and the projector does not show the bit depth but does show YCbCr. Maybe this was a fluke so I'll check again.

I've seen no evidence that the Radiance outputs 12-bit unless you send it 36-bit Deep Color from a source (the only source I have that does that is my Oppo through processing). Every other time my RS-50 reports 8-bit input from my Radiance.

This post from Jim Peterson says that the Radiance is 10 bit from input to output.

I think that the YCbCr 4:2:2 output is 10 bits.

Thanks, I was searching for that post!

Here is another valuable thread where Jim explains why it is best to feed the Radiance and output YCbCr 4:2:2.

I remember reading somewhere in the manual (I think) that they recommend setting the output to YCbCr 4:2:2 though I wasn't sure what the reason behind this was, the link above explained it well enough so thanks.

I'm currently still using my VideoEQ with my projector until I can spend some time recalibrating my HD350 using the Mini3D. With the M3D set to YCbCr 4:2:2 will I be feeding my HD350 with 10 bits (I suppose it will show up as 10 bit in the 'deepcolour' info menu if so)? Maybe I'm overthinking this as I guess Lumagen have already thought about how processing an 8 bit signal will effect it WRT rounding errors/truncation...I should just calibrate and get on enjoying the picture.

If it does, let me know. I have no info on there on the RS20.

Hi LG,

Sorry, my last post was incorrect.

I double checked tonight, and here is what I get reported on the JVC depending on the pixel format on my AMD driver:

YCbCR 4-4-4 : 10 bits
YCbCr 4-2-2 : nothing (like you, which means no deep colour)
RGB Studio Limited 4-4-4 : 10 bits
RGB PC Standard (full) 4-4-4 : 10 bits

And of course bluray outputs 4-2-0 not 4-2-2

So I don't know what I was doing to post som much wrongness in one single post, but it looks like what you report is normal and the Radiance or your AVR isn't interfering.

I guess it just shows I've sadly reached the age when I should stop trusting my memory.

Sorry!

I had a quick play only using my TV which can't accpet anything other than 8 bits. I changed to YCbCr 4:4:4 and the dither setting to 10 bits but I didn't lose the picture on the TV, which makes me think it wasn't 10 bit output. (In a similar setting on the Edge I would only get a picture on the HD350 which would show 10 bit in the deepcolour menu).

Have to wait until tomorrow to try the different options to see if I can get the HD350 showing 10 bits in the information menu. Another (slightly perverse) idea is to temporarily connect my Edge after the Mini3D just to see what it's information screen says about the input.

That quote is before they went to 12 bit in the cms. Check out the release notes of Nov. 2007 software for the XD.

Shawn

That's weird because I am pretty sure that when I output RGB from the Radiance, it came up as 8 bit.

I have to admit that I don't yet really understand what this means or why/if it is important.

I guess I can understand and accept that the VXP in the Radiance likes internally processing YCbCr 4:2:2 data. What I don't know is what the bit depth is of the YCbCr that is being output. It seems to me that more bits is better than less, but I suspect that this may be simplistic thinking. If you are processing data at 10 or 12 bits I would think that something gets lost if you convert it back down to 8 bits to be output. Again, I just don't know the answer.

I can do what I am told to do but I would rather understand why before doing it.