or Connect
AVS › AVS Forum › Video Components › DVD Players (Standard Def) › Explanation of Color Space: HDMI and DVI
New Posts  All Forums:Forum Nav:

Explanation of Color Space: HDMI and DVI  

post #1 of 210
Thread Starter 
First off, all DVDs are made in SD, and this does not change to HD when you upscale a DVD; it's still SD color space. To get proper colors, the output of your player must be SD, AND the input of your projector, HDMI or DVI, must process the signal as an SD input, not an HD input.

Because HDMI players output the digital YCbCr signal right from the DVD, they all output the correct SD color space. This is not the case with DVI which must be converted into an RGB signal, and can be done two ways, one right and one wrong. This ALSO means that when using an HDMI out player the fact that it uses proper SD color space over HDMI does NOT mean that it is using proper SD color space if you are going HDMI to DVI.

Since DVI inputs are inhenently RGB, they just pass what they get, SD or HD, so the player MUST output SD over DVI to have correct color space. There is no way to correct for this issue. For example, according to the WSR of the Sony 975, it outputs HD RGB over DVI, so the color space will be wrong, and there is no way to correct it.

With HDMI players, they, too must properly output SD color space and inherently do so. The problem arises when the HDMI input on the projector sees a 72op or 1080i signal, thinks it's an HD signal, and automatically switches to HD Color Space. Wrong color space.

If your projector has a color space setting on HDMI, you can correct it. The control would be marked Rec. 601, which is SD color space, and Rec. 709 which is HD color space. Use 601. Note that some DVI in projectors have this color space choice, but it applies to the component inputs only, not DVI which is RGB.

Rec. 601 and Rec. 709 are YCbCr to RGB converters when used in a projector, or in a player to get a DVI RGB signal. 601 is SD. 709 is HD.

If you feed an HD Color Space signal, for example HDTV, into an SD Color Space input, or an SD color space signal into a HD color space input, the colors will be wrong, principally green, with either green push or depression. HD into HD for a DVD is also wrong since DVDs are SD, but I do not know what color anomolies are produced.

As confusing as this may sound, it is really quite simple. Since DVDs are SD you want your player to output an SD signal, whether upscaled or not, from any output, HDMI, DVI, or component for that matter. And you want your display, regardless of the input used, to handle it as an SD signal to get proper color space.

What is NOT simple or easy is finding out the color space which a player is outputting. The projector is easy: for DVI, it's a pass through of what it's fed so it's entirely dependent on the player's output which MUST be SD. For HDMI, your player will output an SD YCrCb signal, and your projector must give you the option of overriding the automatic switch to HD color space in the presence of a 720p/1080i signal and force it into 601.

Also, note that if selecting HDMI RGB, the SD DVD signal is going to have to pass through a YCbCr to RGB decoder, and if the decoder is an HD Rec.709 it will output incorrect HD RGB color space. Thus, it is entirely possible to have an HDMI DVD player which outputs the correct Rec. 601 SD color space over HDMI YCbCr, but the wrong Rec. 709 HD color space over HDMI RGB, or when hooked up to a DVI in display.

I guess the simplest answer is that if your greens are screwed up over DVI or HDMI RGB, you know the player is improperly outputting HD color space, and if your greens are screwed up over HDMI YCbCr, it means your projecor should be forced into 601 instead of 709, if possible.

If you have the DVI problem outputting HD 709, there is no solution. If your HDMI projecor does not allow you to force it into SD 601, there is no solution.

For an excellent explanation of these issues see the Jan WSR review of the Sony 975. Any suggestion on how to determine SD/HD color space beyond looking for screwed up greens is welcome. Also, any reports of color space on players, or the ability to force an HDMI projector into SD 601.
post #2 of 210
Quote:
First off, all DVDs are made in SD, and this does not change to HD when you upscale a DVD
You wish! At least one studio leaves their DVDs in 709. They are created from an HD downconvert and they don't twist to 601.

Since this thread is on CSC, I have attached the images with one half in 601 and the other in 709. I decodec twice, once using 601 and the other 709. You can see what happens when the decode is incorrect.

 

cb-601-50.zip 292.5654296875k . file
post #3 of 210
Quote:
Originally posted by sspears
You wish! :) At least one studio leaves their DVDs in 709. They are created from an HD downconvert and they don't twist to 601....
Which one do you know does not convert?
post #4 of 210
Robert and Stacey,
Can we quantify this a bit?

For example the Avia Color Decoder test pattern shows Red, Green, and Blue errors as a percentage above or below perfect decoding.

If you have HDMI to DVI operating at 1080i, and the HDMI is incorrectly sending RGB converted via 709 to the DVI which naturally treats it as 601 (if I follow Robert's explanation), this should result in Green being too dark (among other things) -- presumably a negative value in the Green Color Decoder column on the Avia test screen. Right?

So just how big would this error be on the Avia chart? Would it be clearly distinguishable from normal manufacturing/calibration variants in the displays? More negative than -20% on that chart for example? Or would it be a small error, such as -3% or -5%?

----------------------------------------------

Also the Pioneer Elite DV-59avi has a special item in its HDMI picture adjustment settings called "HDMI Color Adjust" with a choice of "Standard" or "Enhanced". The factory default for the HDMI output is "Standard". The choice is not offered on the component outputs.

The manual only describes this as, "Adjusts the overall color intensity in the HDMI video signal," and goes on to warn, "You may find that selecting the Enhanced mode in HDMI Color Adjust causes picture problems on your TV. You can correct this if your TV has an RGB mode that you can switch on when this occurs."

Now changing the factory default from Standard to Enhanced on an HDMI to DVI hookup at 1080i (to a Fujitsu P50-30 series plasma) has the apparent effect of increasing white levels and decreasing black levels dramatically (you have to lower Contrast and raise Brightness on the plasma significantly to compensate), but once you do that the Color, Tint, and Decoder tests are essentially unchanged from those achieved with the Standard setting -- no need to re-adjust Color or Tint, and no change on the Decoder chart.

Do you have any clue what the Enhanced setting might be trying to accomplish? So far I've found no circumstances where I'd want to change from the default Standard setting. Is it perhaps something that would only be useful in an HDMI to HDMI hookup?
--Bob
post #5 of 210
Thread Starter 
Bob-

Sorry if I was unclear, but a DVI input simply passes through the digital RGB signal it gets. It neither treats the signal as HD or SD. If correct, an HDTV signal, which uses HD color space, will simply pass through the DVI input as a REC. 709 HD signal. If correct, since DVDs are SD color space (or are supposed to be), the player will output an SD signal which simply passes through the DVI input as a REC. 601 signal.

As to the quantification of the error when REC. 709 is used when REC. 601 should be used for an SD signal, read Paul Bigelow's excellent first post in the Sony 975 thread. He states that this results in a -20% green on the Avia decoder chart, resulting in a significant green depression. The actual calculated differences between REC.601 SD and REC. 709 HD are +.0864R, -.1282G, and +.0418B, where Y=RGB of 1.

Finally, as to the normal/enhanced color control on the DV-59AVi, notwithstanding its name, from your description, it sounds like a 0IRE/7.5IRE control as it appears to not affect colors, but brightness or contrast.

Bob
post #6 of 210
Excellent post Robert to clear up this confusing issue. I found this first hand on my Zenith DVB-318 which pushes or depresses green because of the wrong color space on most HDTV thinking 1080i is a HD signal and therefore apply HD colorspace into a SD signal (albeit upconverted) from the Zenith player. Luckily my old PJ (NEC LT150 allow colorspace choices and indeed choosing SD colorspace (to match Zenith's SD colorspace) correct a lot of the green problem. Digital PJ using the cheaper bulbs tend to push green due to the imperfect light source so there is always a slight green tint which I also tried to correct using a photographic filter (Hoya MMC FL-D) to reduce the green and blue light and allowing more reds through. This maybe why some owners are not as happy with Zenith as others like myself.
post #7 of 210
If you have a display device capable of receiving HDMI YCbCr, and you have a DVD player outputting YCbCr, then is the color space irrelevant, that is always correct? Are there HDMI capable display devices that can only accept HDMI RGB, and not accept HDMI YCbCr?
post #8 of 210
Robert,
Thanks for the pointer to Paul's test. The -20% figure should be sufficient to clarify this for anyone who has a player they want to try via HDMI to DVI and with access to an Avia calibration disc.

[Note to newbies: You have to adjust Color and Tint properly first (which will, necessarily, zero out the Blue column in the Color Decoder chart) before the Red and Green column results from the Color Decoder chart become meaningful.]

I feel pretty comfortable, then, in asserting that the Pioneer Elite 59avi does *NOT* have this problem when hooked up HDMI to DVI at either 720p or 1080i and when using the default factory settings for "HDMI Direct" picture mode -- which includes a "Standard" setting for the mysterious "HDMI Color Adjust" item.

And since switching that item to Enhanced (and then re-adjusting basic levels on the display) doesn't seem to have any significant effect on the Color Decoder test either, I'd have to believe this color conversion problem doesn't get turned on if you switch it to Enhanced either -- whatever else the heck Enhanced does.

-----------------------------------------------------------

I'm still leaving that "HDMI Color Adjust" setting in the, "I have no clue what it is trying to do" column. It doesn't make sense for it to be a black levels adjustment (there are other menu items for that). And has been discussed before, the whole 0 vs. 7.5 IRE distinction is meaningless in the digital video signals anyway, since "Black" should be digital 16 regardless.

Ah well, just leave HDMI Color Adjust set in Standard and be happy.

So it appears to me the 59avi can safely be put in the camp of both "proper" Studio RGB and "proper" 601 color conversion when hooked up HDMI to DVI at 720p or 1080i.
--Bob
post #9 of 210
Thread Starter 
uzun-

No. On the HDMI input, when a projector sees a 720p or 1080i signal, it automatically switches into HD color space. HDTV signals use HD color space, and the switch is entirely appropriate.

But DVDs are SD, and inherently 480i. When you upscale them to 720p or 1080i you are making a phony HD picture in a way, but the color space does not change. It's still SD. Your proj. doesn't know this, and automatically switches to HD color space resulting in severe green depression.

The ONLY way around this is IF your projector has a setting on the HDMI input where you can choose Rec. 601 SD color space, or Rec. 709 HD color space, and can force the projector into SD color space for upscaled DVDs over HDMI YCbCr.

No, there are no devices which can only accept HDMI RGB. But do not use HDMI RGB as an output for your DVD player as the YCbCr to RGB convertor in the player may be a REC. 709 HD convertor. Then, when you fed the signal into the display and it sensed an HD signal and switched the projector to HD color space, you would have double improper HD color space and I think your greens might disappear altogether.
post #10 of 210
Quote:
If you have a display device capable of receiving HDMI YCbCr, and you have a DVD player outputting YCbCr, then is the color space irrelevant
If the source color info (601/709) is not passed across HDMI to the display, it will guess which color space to use when it converts to R'G'B'.

So, if you have a 59AVi and you output 720p Y'CbCr over HDMI and there is no color info, the display will assume it is HD and apply the HD color matrix when it should have applied the SD color matrix.

The DVDO scaling in the Denon 5910 has the ability to twist the color info from 601 to 709. The HQV part has this capability as well.

The color info is supposed to be part of the MPEG2 data that is carried on DVD. Of course there is nothing that says you have to add it when you create the DVD. :)

In the case of HD, like WMV HD, there is no way to store this info in WMV9 MP. Some of the titles have actually been twisted to 601 but you will never know. For DVE Pro, I created both 601 and 709 discs.

WMV9 AP (and VC1) have the ability to store the date like MPEG2. The problem at the moment is that there are no tools to fill in the data.
post #11 of 210
Quote:
Originally posted by sspears
If the source color info (601/709) is not passed across HDMI to the display, it will guess which color space to use when it converts to R'G'B'.

So, if you have a 59AVi and you output 720p Y'CbCr over HDMI and there is no color info, the display will assume it is HD and apply the HD color matrix when it should have applied the SD color matrix.

...The color info is supposed to be part of the MPEG2 data that is carried on DVD. Of course there is nothing that says you have to add it when you create the DVD.
Oh, for crying in a bucket! What next!? ANOTHER issue with upscaling a DVD source!

Stacey, do you have any idea how often DVD movies are encoded without color info?

Gary
post #12 of 210
Thread Starter 
Actually, with the color space issue and the PC remapping issue, I have decided to abandon upscaling players, having owned the Samsung 931, using the Denon 1910 with the Toshiba and Pan S97 on order, and use my old RP82 and let the Faroudja chip in my InFocus 7200 upscale.

So there are extra D/A and A/D conversions and no pixel to pixel mapping. Big deal. At least I don't have to worry if I'm crushing blacks and whites, if I'm getting the correct color space, if my greens are depressed or too high.

In one post on color space, I started, "As confusing as this might sound, it's actually easy..." I was wrong. It isn't easy at all. It's hopelessly complicated.
post #13 of 210
Quote:
The color info is supposed to be part of the MPEG2 data that is carried on DVD. Of course there is nothing that says you have to add it when you create the DVD.
Oddly enough, if sequence_display_extension is not present in the MPEG-2
video bitstream (as is the case with probably 99.99% of DVD's, although now
that I think about it, I'll have to check a few DVD's), the default color space is
ITU-R Rec. 709.

However, it's a moot point. Since an MPEG-2 decoder is not responsible for
YCbCr to RGB conversion, the sequence_display_extension information is
ignored and typically not available externally for any YCbCr to RGB conversion
device.

Ron
post #14 of 210
Quote:
Originally posted by Robert Whitehead
...As to the quantification of the error when REC. 709 is used when REC. 601 should be used for an SD signal, read Paul Bigelow's excellent first post in the Sony 975 thread. He states that this results in a -20% green on the Avia decoder chart, resulting in a significant green depression. The actual calculated differences between REC.601 SD and REC. 709 HD are +.0864R, -.1282G, and +.0418B, where Y=RGB of 1...

Bob
I remember reading the post referenced above. I found his explanation of green depression/push very enlightening.

However, I'm pretty sure his explanation of the 975 (via HDMI) stated that it incorrectly output SD color space -- and that's why the 975 exhibited a green depression. To be clear, Paul's explanation did not pertain to a DVI input.

According to you, this is what should be happening (e.g., color space should be SD). Since you both seem to know what you're talking about, who's wrong? It sounds like you'd should be contradicting Paul's assessment that the 975 has a green depression via HDMI (due to its use of SD color space). Furthermore, Paul said the S97 outputs HD color space and looks correct. (Should I conclude his display also applies the HD color space to a 720/1080 signal and THATS why the Sony looked depressed)?

Also, since my Sony HS20 does not change color palates when I choose an upconverted signal, I guess it must automatically detect the correct color space (vs automatically switching to HD color). You didn't seem to include the possibility that color space could be automatically detected (or did I mis-read you)?

gp
post #15 of 210
may i ask,

is REC 601 is the same as SMPTE-C?
post #16 of 210
What I'm reporting on is the final, visual results.

Using the same display (Panasonic TC-22LH1), with the same settings, with the same DVE, with the same filter, with the same HDMI connection, with the same HDMI cable, with the same 1080i resoluiton, with the same fixed HD color matix (whew ;) ) :

1. The Panasonic has no issue with green depression
2. The Sony has noticible green depression

At 480p resolution (where the matrix can be selected at the display) with the SD matrix:

Neither player has a problem with green depression.

At 480p resolution with the HD matrix:

Both players have green depression.

Paul
post #17 of 210
Thread Starter 
Waterbug-

No. SMPTE-C refers to a particular set of phosphors used in CRT TVS. REC. 601 is NTSC (analog TV, VCRs, DVDs). REC. 709 is ATSC (HDTC, etc.)
post #18 of 210
Thread Starter 
Gregg-

First, HDTV displays have both 601 and 709 color matrixes and DO automatically switch between them based on the input rec'd 480i/p>SD 601 and 720p/1080i>HD 709.

DVD is inherently a 480i standard definition NTSC signal encoded as YCbCr. HDMI YCbCr simply passes this SD signal out, so ALL HDMI out players correcly output SD. (Note: I have heard that at least one studio improperly makes its DVDs using 709 color space; in that case the signal would be HD but would be wrong.)

I don't understand what is happening with your Sony. If it is designed to display HDTV, then it should have both SD REC.601 and HD REC.709 convertors, with switching determined by the signal format (480i, etc). So when you switch from an SD signal format to an HD signal format, the display should switch from 601 to 709, and the colors should change, for the worse. Is this setting possibly manual in your projector?

Gregg and Paul-

From WSR; "However, automatic selection of the YCbCr decoding matrix [by the display] won't work in the case [of upscaled DVDs]. When the DVD player upconverts the standard-definition DVD to 720p/1080i, the display will incorrectly use the HD decoding matrix. To obtain accurate color, the display must provide the ability to manually select the standard definition decoding matrix. If the high definition decoding marix is used, the colors will be shifted. Green and yellow appear darker, cyan becomes too blue and so forth. This is easy to see by simply displaying color bars and changing the DVD player's output format. At 480i/p they are correct, but at 720p and 1080i, they aren't"

Paul-

ALL of your results, with one exception, are entirely consistent with this. What all of this says is simply that if you put an upscaled SD DVD YCbCr signal into an SD color matrix, the colors are right, and if you put it through an HD color matrix, the colors are wrong.

When you put SD 480p from either player into SD color space, both were fine. When you put the same SD signal into HD color space, you got green depression.

When you fed the Sony at 1080i into HD color space, you got green depression which is exactly what you should have gotten, since you were passing an SD signal through HD color space.

The anamolous finding was that there was no green depression on the Panasonic at 1080i when you passed its SD signal into HD color space. You should have had green depression. You did get it at 480p, so by all rights, you should have gotten it at 1080i too.

Upscaling an SD DVD into 720p/1080i does NOT change the color space from SD to HD; it merely changes the resolution. Nor do HDMI out upscaling players change the color space from SD to HD when they upscale a signal. As I said above, the HDMI YCrCb outputs what has been encoded digitally on the DVD as SD YCbCr.

I have no explanation for this anamolous result. It is NOT that the Panasonic is outputting HD color space at 720p/1080i, and SD at 480i/p. For it to do this, it would have to contain a REC.601 SD Color Space to REC. 709 HD Color Space converter and I don't know if they exist. Your results with the Panasonic at 480p clearly show it is correctly outputting SD color space. You've got me. Try Kris or Stacey.

Bob

P.S. This discussion is limited to HDMI>HDMI, and not HDMI>DVI which presents a whole different set of issues.
post #19 of 210
Quote:
Originally posted by Robert Whitehead
What all of this says is simply that if you put an upscaled SD DVD YCbCr signal into an SD color matrix, the colors are right, and if you put it through an HD color matrix, the colors are wrong.
Is that ALL of this thread is about? Has not this subject already been discussed?

Can we try to a bit more objective in the future?
For example, what are the requirements for which HDMI implementations should meet? What is optional?
http://www.hdmi.org/manufacturer/nda...&submit=Submit

What capability does the specifications provide for? Lets take a look at a few. Take video level regardless of what the expert here try to push:
"Video Quantization Ranges black and white levels for video components shall be either Full Range or Limited Range. YCBCR components shall always be Limited Range while RGB components may be either Full Range or Limited Range. While using RGB, Limited Range shall be used for all video formats defined in EIA/CEA-861B, with the exception of VGA (640x480) format, which requires Full Range."
Notice the terms Limited and Full. BOTH ranges are valid according to the specification. Please accept it or work to change the specification.

AVI InfoFrame Packet Contents
- RGB or YCBCR indicator. See EIA/CEA-861B table 8 for details.
- Active Information Present. Indicates whether field R0…R3 is valid. See
EIA/CEA-861B table 8 for details.
- Scan Information (i.e. overscan, underscan). See EIA/CEA-861B table 8 for
details.
- Colorimetry (ITU BT.601, BT.709 etc.). See EIA/CEA-861B table 9 for details.
- Picture Aspect Ratio (4:3, 16:9). See EIA/CEA-861B table 9 for details.
- Active Format Aspect Ratio. See EIA/CEA-861B table 10 and Annex H for
details.
- Video Format Identification Code. See EIA/CEA-861B table 13 for details.
- Pixel Repetition factor. See EIA/CEA-861B table 14 for details.
- Non-uniform Picture Scaling. See EIA/CEA-861B table 11 and paragraph on page 58.

As HDMI is an intelligent protocal the color space is defined. Is it being set properly? (Hint:Only a bus analyzer can provide the answer)

DVI/HDMI Device Discrimination
In order to determine if a sink is an HDMI device, an HDMI Source shall check the E-EDID for the presence of an HDMI Vendor Specific Data Block within the first CEA EDID Timing Extension.
Any device with an HDMI VSDB of any valid length, containing the IEEE Registration Identifier of 0x000C03, shall be treated as an HDMI device.
Any device with an E-EDID that does not contain an HDMI VSDB of any valid length shall be treated by the Source as a DVI device (see Appendix C).

So the HDMI source can tell if the destination is DVI and react accordingly.

Now what are the issues with these devices in this thread?
Is the HDMI specification incomplete?
Or is it a issue of non-compliance?
Or is it an optional feature which the manufacture has chosen not to implement?
Again an HDMI bus analyzer with Compliance Test Specification software provides the only definitive answers. Everything else is personal feelings.

What is the Compliance Test Specification?
The Compliance Test Specification (CTS) was developed by the HDMI Founders to assist manufacturers in ensuring the compliance of their products with the HDMI Specification. It consists of numerous tests designed to check for compatibility of various HDMI-related aspects of a product, including audio, video, EDID, electrical signaling, protocols, etc. The HDMI Compliance Test Specification represents the minimum compliance testing required for HDMI products, and it is up to each manufacturer to fully test product functionality and interoperability.

My recommendations again:
1) Deemphasize the legacy analog video tests, and also the ones that are rarely if ever actually seen in current releases.
2) If a player generates MACROBLOCKING fail it immediatly. Don't even test it. Consider revising past rankings and scores as to not mislead newbees.
Even Consumer Reports has a rating for products such as these. They catorgize it as "Unacceptable".
3) Use a HDMI/DVI analyzer/generator with Compliance Test Specification software to answer virtually ALL of the questions now and forever (HD Dvd coming soon) more. Do you accept?
post #20 of 210
Thread Starter 
reincarnate-

The portion you quoted deals with how a player converts YCBCr into RGB, either Full Range (usually known as PC) RGB or Limited Range (usually known as Video or Studio) RGB. Only Limited is correct. If Full is used you get crushed whites and blacks, as with the Samsung 931. This is a very important issue for those with DVI in displays, whether the player is HDMI or DVI out.

This thread, however, deals with Color Space, which is a different but equally important issue for those with DVI in displays. Done improperly, the colors will be incorrect, notably with a green depression.
post #21 of 210
The term limited is misleading. It means where the reference points are. I discussed this with SI a couple of years ago and they did not think it would cause any confusion and I disagreed.

The limited range mode is not supposed to clip values outside of reference, but it does. They will fix it, but for now CE companies are working around it. I can confirm that the Pioneer 59AVi and the HD TiVo both work around it.

Ron, I will start checking for SDE, to see what is out there. I know Meridian is passing the color info over HDMI. Not sure what the info is and where they are getting it at the moment.

HD-DVD and Blu-Ray are going to make things a lot more complicated and there will be plenty of room to mess it all up. :)
post #22 of 210
I checked a few recent release DVD's last night. Hero, Spiderman 2, I, Robot,
Kill Bill 2, Hellboy and Star Wars IV. The FOX releases (I, Robot and Star Wars
IV) do have sequence_display_extension, but none of the others. So it's not
as bad a I thought, but still a low percentage.

The FOX titles signal SMPTE 170M color space, which is the same as Rec. 601.

Ron
post #23 of 210
Robert,

There is an obvious difference in how the color is reproduced at 1080i between the Sony and Panasonic. The Sony (and certain upconverting Zenith/LG players) has the green depression, the Panasonic S97, the Momitsu V880, the Denon DVD-1910 do not.

Have a look here at the AD converter for the DVP-NS975V on page 42. There is both HD and SD color control:

http://www.analog.com/UploadedFiles/...7300A_1A_a.pdf

I don't think it's far-fetched to have control over the HD/SD color matrix within the Sony player.

Paul
post #24 of 210
Paul,

First off thanks for all the information you present here:)

I simply want to know if the DVP-NS975V is a good short term (until HD DVD) choice for up-converting dvd player for a DLP RP TV? (Samsung HL-P5674W)

How obvious is the lack of BTB over HDMI going to be? Will space scenes in Star Wars (for example) look washed out? Or will the BTB appear mostly when using a calibration disc?

Sorry about the 1000 questions. Reading all this information is just mind-boggling.
post #25 of 210
Thread Starter 
Paul-

Perhaps my understanding of what happens when a player outputs HDMI YCbCr is wrong. As I understand it, the digital YCbCr signal from the DVD passes through the MPEG decoder and then directly to the HDMI out for a 480i signal. If the signal is upscaled, it goes from the MPEG decoder to the deinterlacer/scaler and then to the HDMI out. So the player does nothing to the color space as encoded on the DVD, which should be SD REC. 601. There is no processing of the signal on the DVD beyond MPEG decoding, deinterlacing and scaling.

Because the signal is YCbCr, an HDMI player has no opportunity to screw up the color space on HDMI out. (Different result for HDMI to DVI out when the player must convert the signal to RGB and does have a chance to get the wrong color space.) Nothing is done to the color space on the DVD. It's just passed out HDMI.

The Analog Devices video encoder 7300A does have SD and HD color space. But, if my understanding is correct, no such encoder or decoder is used in the HDMI YCbCr signal.

If I recall your JVC uses HD color space only at 1080i, and passing an SD signal through HD results in green depression. This is exactly what WSR found with the Sony: when the display used SD REC. 601 the colors were fine; when the display used HD REC. 709, the colors were off.

Again, unless my understanding of HDMI YCbCr is completely wrong, I am at a complete loss as to explain the absence of green depression in the players you mentioned. I wish Stacey would read these posts and offer an explanation of the absence of green depression on the Pansonic, et al. STACEY!!! HELLO!!!

But, hypothetically speaking, if those players were taking SD Color Space DVDs and somehow outputting them in HD color space, then those players would be outputting a sigal that already had green depression, and the HD processing in your JVC would make the green depression from those players even worse than it is on the Sony. So the explanation can not be that the Panasonic et al are outputting HD Color Space, because the green depression would be worse, not absent.

Bob
post #26 of 210
Thread Starter 
Go Spur-

If your Samsung DLP has DVI in the Sony is not a good choice because the Sony improperly uses an Hign Defintion Color Space REC. 709 YCbCr to RGB converter, and you will get screwed up colors with severe green depression. (DVDs are SD or Standard Defintion.)

If your Samsung DLP had HDMI in, it will be fine IF the Samsung has a menu option to allow you to chose between SD REC.601 and HD REC.709 Color Space. If it does not, then when the Samsung senses a 720p signal from the Sony, it will automatically switch into HD REC.709 color space (which is correct for HDTV), and you will again get screwed up colors with severe green depression.

This is the case with all HDMI out players. Whether this is a problem is entirely display dependent. If you can not switch between the two color spaces, then you are better off getting a good 480p player, as that will cause the display to go into the correct SD color space for DVDs which are made in SD.

Paul can far better answer the question on BTB on the Sony if your Samsung DLP allows the use of an upscaled DVD signal with SD Color Space.
post #27 of 210
Eeeek! What a mess. I just read this thread from top to bottom. I think I grasp most of it, but who knows ! I don't use HDMI and realize most of the info here relates to that but its quite the read. Now I know why some users that used HDMI to DVI converters find blacks clipped or other issues.

Makes me wonder what type of potluck performance is going on with everyone's shiny new upscaling dvd player purchases out there!

My pioneer 503 plasma uses pc monitor type setting I think. I use a Bravo D1 via DVI and Denon 3800 via component and a 2910 via DVI. The plasma only has DVI so its always DVI to DVI, no HDMI involved.

I have posted several times I felt the 2910's finer shades of some soft colors/depth to be off, I wonder if this has anything to do with all this or its just what Denon has been working on with the last 3 firmware updates for the 2910.

We have gone 3- to -4 to -5 firmware version 5 in just 2-3 months.


The 3800 is known for being better for film and being a great player so it could just be the quality of that model and the SIL chip. Its color palette to my eye is more accurate and more depth on the subtle tones than the 2910.

I find I need to adjust the IRE or the black level settings to enhanced to get true blacks on the plasma but otherwise the palette looks 95% or so correct to me except for the depths of some of the finer softer colors and shades of some of them on the 2910 via DVI when compared to the 3800 as mentioned above via component.

I think these companies should put more info or a small primer out on what types of displays require what settings. I confess I like to think of myself as tech savy but following this thread for the average joe must be hard.

My hat is off to you guys. Happy Holidays!

;)
post #28 of 210
Great thread Rob, along with the others on dvi/hdmi. Worth pointing out that I am somewhat annoyed that some display manufacturers have chosen to disable color matrix control for HD resolutions. Like auto-full aspect ratio with HD resolutions I'd really like to see this go away on displays that allow user menu controls for these items.
post #29 of 210
Thread Starter 
Paul-

Interestingly, I just read in another thread that HDMI YCbCr at 480i is as pure and unadulterated a signal as SDI. The signal comes right off the MPEG decoder carrying the Color Space of the DVD with it. There is no processing of the signal of any kind, and no alteration of the SD color space on the DVD.

Now, when you upscale the signal, the only thing additional is the deinterlacer/scaler. Most commonly used these days is the FLI23xx, and the only thing that chip does is deinterlace and scale. It does nothing to the color space.

So, as I believed, there is no alteration or processing of the SD Color Space on a DVD over HDMI YCbCr at any resolution. This does NOT explain why you are not seeing green depression with the Panasonic, Bravo or Momitsu. Since at 1080i your display is operating in HD color space, you should get green depression on EVERY HDMI YCbCr player, but you don't and I can't figure out why.

Bob
post #30 of 210
A service manual of all players and data sheets on all involved ICs would definitely help -- otherwise were just looking at the viewed results and speculating.

I've looked for FLI-23xx and Sigma EM-8500 datasheets and have come up empty so I can't really say what the exact capabilites of the these chip are.

Since the Sony (even at 480i HDMI), Panasonic, Momitsu, and Denon players have control over various picture parameters (including color) via HDMI and/or DVI "something else" is in the digital chain rather than just pure digital output of the DVD.

Paul
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: DVD Players (Standard Def)
This thread is locked  
AVS › AVS Forum › Video Components › DVD Players (Standard Def) › Explanation of Color Space: HDMI and DVI