ATI HD 5000 Series Known Problems (5870 5850 5770 5750 5670 5570 5450) - Page 28 - AVS Forum

AVS Forum > Video Components > Home Theater Computers > ATI HD 5000 Series Known Problems (5870 5850 5770 5750 5670 5570 5450)

Home Theater Computers

Andy o's Avatar Andy o
04:16 AM Liked: 11
post #811 of 2707
03-03-2010 | Posts: 4,382
Joined: Jan 2007
Quote:
Originally Posted by rjyap View Post

Can you point me how to override the video part of EDID? I doubt it's EDID problem as HD 4850 had no issue with same setup but no harm trying.

I guess I or someone else with confirmed Full RGB pixel format working would have to extract their EDID and pass it on to Tulli or someone else that can build up a proper EDID for you. I'll get on with it tomorrow.
sanderdvd's Avatar sanderdvd
04:38 AM Liked: 24
post #812 of 2707
03-03-2010 | Posts: 2,206
Joined: Nov 2001
Quote:
Originally Posted by somy View Post

There is no perfect solution now, and to be safe just choose RGB full.

But like I said there are many things that can be changed so I assume there must be some things to tell about all these settings?
somy's Avatar somy
06:11 AM Liked: 11
post #813 of 2707
03-03-2010 | Posts: 283
Joined: Oct 2008
Sotti,

One more question. When you test YCC output, what DR setting did you use? Limited or full? Do you see any difference?

Regards
Wiggies666's Avatar Wiggies666
06:24 AM Liked: 10
post #814 of 2707
03-03-2010 | Posts: 42
Joined: Mar 2008
In the supported formats tab of the ATI HDMI Output, I just see DD, DTS, sample rates till 96KHz and bit depth till 24 bit (lseems ok).
Shouldn't I have more than this?
Is this connected to not having a TrueHD/DTS-MA compatible receiver? Im using a Marantz SR7001, does this have anything to do with it?

Reason I am asking this, is because I have no clue to what the pc is outputting to the amp.
I am playing Blu-ray and HDDVD with TMT3 (.170) through lpcm. How can I see what my pc outputs besides using Reclock, cause this is not working with the latest patch of TMT? Gives the error "the audio format is not supported by the hardware/driver. 48/24bitpcm/6channels" when playing HDDVD with TrueHD.

Used card is the Sapphire HD5770 1GB
Driver is the ATI HDMI from Catalyst 10.2.
jong1's Avatar jong1
06:41 AM Liked: 12
post #815 of 2707
03-03-2010 | Posts: 1,688
Joined: Jun 2007
Quote:
Originally Posted by somy View Post

Sotti,

One more question. When you test YCC output, what DR setting did you use? Limited or full? Do you see any difference?

Regards

I know you didn't ask me But here Full DR is the only way to go. Limited DR in combination with YCbCr output over compresses all video affected by the DR setting (PVP and "normal").
Andy o's Avatar Andy o
06:46 AM Liked: 11
post #816 of 2707
03-03-2010 | Posts: 4,382
Joined: Jan 2007
Yeah, that was my problem as well, and when you do Full DR, banding gets introduced just like when you do full DR in RGB pixel formats.
somy's Avatar somy
06:48 AM Liked: 11
post #817 of 2707
03-03-2010 | Posts: 283
Joined: Oct 2008
Sotti mentioned in his post:
"One thing I did check and notice is that using DR of Limited does pass BTB and WTW for BD. But of course it jacks up the levels for DVD. "
Choose DR limited pass BTB and WTW basically means he is getting 16-235 all the time on this TV.
In my setup I get 0-255 output as you can see:
http://www.maimuzi.com/HTPC/1.JPG
somy's Avatar somy
06:51 AM Liked: 11
post #818 of 2707
03-03-2010 | Posts: 283
Joined: Oct 2008
Quote:
Originally Posted by Andy o View Post

Yeah, that was my problem as well, and when you do Full DR, banding gets introduced just like when you do full DR in RGB pixel formats.

OK, I assume you're talking about MPC-HC.
PowerDVD never gives me banding no matter what combination I choose (DR limited/full, pixel formate limited/full).
cbemoore's Avatar cbemoore
07:26 AM Liked: 10
post #819 of 2707
03-03-2010 | Posts: 20
Joined: Jan 2007
Andy,

Now that the black level problems have been confirmed by lots of people, would it be a good idea to update your first post?
Andy o's Avatar Andy o
07:29 AM Liked: 11
post #820 of 2707
03-03-2010 | Posts: 4,382
Joined: Jan 2007
Yeah, I'm gonna mention the YCC problem as well. The thing is that it's difficult to explain clearly what's happening, and where/when it's happening with these issues.
nickpicker's Avatar nickpicker
07:38 AM Liked: 10
post #821 of 2707
03-03-2010 | Posts: 126
Joined: Dec 2004
Reading how the same graphics card gives different results, I'd like to point out that EDID also contains information about pixel formats. Sample blocks from my Panasonic plasma tv (captured with SwitchResX):

Quote:


Display type:
-------------
RGB 4:4:4 & YCrCb 4:4:4 Color Encoding Formats
Display is non continuous frequency
Default color space is not sRGB standard
Preferred timing mode includes Native Pixel Format

CEA-EXT: CEA 861 Series Extension:
----------------------------------
Revision:..............................3
First DTD block at offset..............33
Display Supports:
basic audio
YCbCr 4:4:4
YCbCr 4:2:2

Perhaps the ATI driver will compress/tweak the output based on this information.
Andy o's Avatar Andy o
07:47 AM Liked: 11
post #822 of 2707
03-03-2010 | Posts: 4,382
Joined: Jan 2007
Yeah that's what we've been speculating as well. I'll upload my KRP's EDID later today.
somy's Avatar somy
07:48 AM Liked: 11
post #823 of 2707
03-03-2010 | Posts: 283
Joined: Oct 2008
Quote:
Originally Posted by nickpicker View Post

Reading how the same graphics card gives different results, I'd like to point out that EDID also contains information about pixel formats. Sample blocks from my Panasonic plasma tv (captured with SwitchResX):



Perhaps the ATI driver will compress/tweak the output based on this information.

Yeah, this is what we're trying to figure out now.
somy's Avatar somy
07:55 AM Liked: 11
post #824 of 2707
03-03-2010 | Posts: 283
Joined: Oct 2008
Quote:
Originally Posted by somy View Post

I noticed the HDMI output of 5770 is very different from my old 4850:
HD4850: YCrCb 4:4:4 pixel format output from 4850 is perfect, I get both WTW and BTB, the colour is accurate and the grey rump looks very smooth. I've tested both PowerDVD(pdvd default DXVA decoder) and MPC-HC(ffdshow as video decoder) with overlay renderer and both worked fine. RGB 4:4:4 produces accurate colour and smooth transition B/W but I didn't get WTW and BTB information.
HD5770: YCrCb 4:4:4 also gives WTW and BTB in overlay just as 4850, but only in PowerDVD. I've checked the grey rump, and it looked horrible: the transition isn't smooth at all. Besides, the colour is way off and I think something must be wrong in the video pipeline. Overlay in MPC didn't give me BTB and WTW anymore, and the output was compressed to something around 50-220 (grey black issue). I then switched to RGB 4:4:4 and tried EVR renderer, everything works fine again except that I didn't get WTW and BTB anymore.
All colour and contrast settings in CCC is disabled and my display panny AE3000 comes with a waveform monitor for RGB colours, so I think the result of my test is reliable. I've also double checked with my DVE blu-ray disk.

Another problem I noticed with powerplay is that when the TV is turned on, it raises freq to the 500/1375, and it never goes back to 157/300 even when TV is turned off later. I have found a work around for this issue: change profiles.xml and set lowest freq to be 157/300, and that will make sure when you have TV connected powerplay won't raise the freq.

Finally my observation brings enough attention now
nickpicker's Avatar nickpicker
08:13 AM Liked: 10
post #825 of 2707
03-03-2010 | Posts: 126
Joined: Dec 2004
Well, the actual issue could be a trivial one.

In an Intel GMA thread here it was mentioned that Intel's driver engineers had a hard time in the past simply for the fact that in their lab they only had a few AVRs to test their HDMI parts of the driver. And, to be honest, with the market situation (vital parts of graphic cards produced and sold at penny cost) there is no way ATI or Intel/nVidia would put enough money in their driver R&D departments to buy a few dozen quality AVRs and tv sets to test enough permutations.

What I have observed is that many of you have above-average configurations, and our displays certainly should be capable of accepting RGB 4:4:4 & YCrCb 4:4:4, YCrCb 4:2:2. What if ATI only tested with a few displays, and actually tried to compensate for their deficiencies, not "knowing" that as a result they mess up other displays, such as those used in this forum?

Don't attribute to incompetence that which could simply come from lack of funds.

EDIT:

Also, and this is purely anecdotal:

A few days ago I downloaded one of tulli's pre-made .INF files (for a smaller version of my Panasonic plasma) and replaced the default "Plug&Play Monitor" driver. I also have applied an .ICC (generated by SwitchResX). This is under Windows 7 x64, using a Radeon 4650, 10.1 driver.

It *seems* to me that desktop colors got "better" after this change. Could it be that ATIs driver actually reads and interprets the EDID from the display "driver", instead of sniffing it on their own?
Andy o's Avatar Andy o
08:44 AM Liked: 11
post #826 of 2707
03-03-2010 | Posts: 4,382
Joined: Jan 2007
The changes in desktop are not subtle at all. If desktop is not normal, it's either severely clipped or compressed, exactly the same difference you see when you switch AVIVO DR from/to full and limited when watching a video.
nickpicker's Avatar nickpicker
08:55 AM Liked: 10
post #827 of 2707
03-03-2010 | Posts: 126
Joined: Dec 2004
@Andy, I clarified that this is currently on a 4650, the 5570 will arrive in a few days.

Anyway, isn't it interesting that the monitor driver affects colors/luma? Until then I was under the assumption that, of course, the ATI driver gets the EDID information from the GPUs HDMI bridge itself.

Could this be the difference in user's experience, i.e. some use the default "Plug&Play Monitor" driver, and some use a custom one matching their display's EDID?
Andy o's Avatar Andy o
09:22 AM Liked: 11
post #828 of 2707
03-03-2010 | Posts: 4,382
Joined: Jan 2007
Pixel format in the 4000 series has the same job to do as the 5000 series, so if any changes are done, they should be clearly visible. If it messes subtly with colors and not just levels, something is much more wrong than we're speculating.

I'm not sure about the distinction you're making about EDID and the driver. Maybe Tulli or vladd can weigh in on that.
nickpicker's Avatar nickpicker
09:37 AM Liked: 10
post #829 of 2707
03-03-2010 | Posts: 126
Joined: Dec 2004
Thinking further about it, things start to make sense:

So what was the original reason why we needed tulli's EDID override? Answer: Certain receivers sent invalid information, so the monitor/HDMI-out driver needed to incorporate EDID information itself.

Turning this around means: The EDID information sent on the HDMI is ignored if the monitor driver already incorporates EDID blocks.

So the question is: What happens when you don't have a monitor driver installed with EDID information? Does the ATI driver sniff it from the HDMI bus, or does it assume "default" EDID information?

Another thing about AVRs in the equation:

Even though an AVR might be able to pass-through HDMI, it certainly is not a given that the AVR will repeat EDID blocks from tv set to graphics card. Because, essentially, this would mean that the AVR were spoofing the display device, and that's certainly forbidden under HDMI licensing terms.

So another question: Do AVRs used in this thread actually repeat valid EDID blocks from display to graphics card?
Tulli's Avatar Tulli
09:52 AM Liked: 87
post #830 of 2707
03-03-2010 | Posts: 5,219
Joined: Nov 2001
Video problems occur for sure when the AVR does not repeat the display device EDID base block, and I've seen that happening, but it's the exception, not the rule. If it happens an EDID override fixes it of course.

OTH I have no idea about the 5xxx's ATI driver just ignoring the AVR's EDID base block specs, so this would be an open question for further investigation.
nickpicker's Avatar nickpicker
10:06 AM Liked: 10
post #831 of 2707
03-03-2010 | Posts: 126
Joined: Dec 2004
Not sure about the spread of the issue. The Onkyo 60x/80x line does not repeat HDMI correctly (there is a firmware upgrade for 80x to fix HDMI issues), the entire Denon 1/2/3x10 line did not (starting with 2310 there is a firmware update, below there is none).

And in any case, repeating EDID is acually spoofing and forbidden under licensing terms. Certain HDMI "adapters" lost their license in the past, and mainstream KVM switches with DVI/HDMI do not repeat EDID but send their own.
Tulli's Avatar Tulli
10:15 AM Liked: 87
post #832 of 2707
03-03-2010 | Posts: 5,219
Joined: Nov 2001
Quote:
Originally Posted by nickpicker View Post

Not sure about the spread of the issue. The Onkyo 60x/80x line does not repeat HDMI correctly (there is a firmware upgrade for 80x to fix HDMI issues), the entire Denon 1/2/3x10 line did not (starting with 2310 there is a firmware update, below there is none).

And in any case, repeating EDID is acually spoofing and forbidden under licensing terms. Certain HDMI "adapters" lost their license in the past, and mainstream KVM switches with DVI/HDMI do not repeat EDID but send their own.

Very interesting information that explains many of the issues that I've seen. But again, a large majority of cases that've I've checked, even with Denons, repeat the base block.

But by "repeat" I mean untouched transmission of all base block data except:
  1. Vendor/Product Identification Descriptor (bytes 8-17).
  2. Monitor Descriptor.
No spoofing here.
nickpicker's Avatar nickpicker
10:24 AM Liked: 10
post #833 of 2707
03-03-2010 | Posts: 126
Joined: Dec 2004
Another thing:

With my Dell 2407, and also on an Eizo 2231W, in CCC on the 4650 (also on 4770) I have the option "Use EDID". Checking this changes desktop colors a bit on the 2407, and dramatically on the Eizo because it's a wide-gamut display.

This means: The ATI driver can be instructed to adjust colors/luma/gama according to EDID information, and do it correctly.

However, when running against the Panasonic plasma, there is no such checkbox even though the display sends correct EDID information (verified by SwitchResX, it even tells me EDID has a valid checksum).

CCC identifies the Dell/Eizo as a computer display, and the Panasonic as a tv set.

Does this mean that for tv sets EDID is ignored unless explicitly supplied in the monitor driver?
Tulli's Avatar Tulli
10:38 AM Liked: 87
post #834 of 2707
03-03-2010 | Posts: 5,219
Joined: Nov 2001
I've checked that option for my Dell 2407 too but haven't tested for significant changes carefully really.

Your point is very persuasive !, but if you look at the registry entries (at Computer\\HKEY_LOCAL_MACHINE_SYSTEM\\CurrentControlSet\\Enum \\DISPLAY) every "Generic PnP" monitor has an EDID stored/associated (even without an override). So I take that, unless there's a bug with current ATI drivers on 5xxxs, the EDID is not ignored for the active display device.
sotti's Avatar sotti
10:49 AM Liked: 175
post #835 of 2707
03-03-2010 | Posts: 6,640
Joined: Aug 2004
Quote:
Originally Posted by somy View Post

Sotti,

One more question. When you test YCC output, what DR setting did you use? Limited or full? Do you see any difference?

Regards

DR Full
nickpicker's Avatar nickpicker
10:55 AM Liked: 10
post #836 of 2707
03-03-2010 | Posts: 126
Joined: Dec 2004
@Tulli, my interpretation is slightly different: even though the ATI driver might know about the correct EDID information via Windows registry, it has to be instructed to actually use it. Over DVI, you have to manually check that option in CCC, otherwise the EDID information is not taking effect. Over HDMI, the option is not even exposed.

I would not have noticed the difference myself were it not for the wide-gamut Eizo. Unchecking the option certainly gives wrong gamut, even when I have the Eizo monitor driver installed. Which leads to the conclusion that Windows might store the EDID in that registry key, but the ATI driver is not picking it up unless instructed to do so.

Perhaps this explains the slightly off green colors that somy/sotti measured?

Does PowerStrip work on the ATI 5xxxs? If I recall correctly there is an option in PowerStrip "Use EDID", as well.

(By the way, the same Eizo with an nVidia 9600GT also gives an overexposed red, typical for wide-gamut displays, in Windows desktop and DirectX games, which further confirms the fact that Windows might store the EDID but the ATI/nVidia drivers don't use it by default. ATI at least can be instructed to use it when running over DVI.)
Tulli's Avatar Tulli
11:45 AM Liked: 87
post #837 of 2707
03-03-2010 | Posts: 5,219
Joined: Nov 2001
Again, very persuasive, @nitpicker. Now we only need a volunteer to put the theory to test.

It would be just a matter of:
  1. Capturing both EDIDs (display device and AVR ... see section 3 in my sig.).
  2. Compare that the AVR EDID base block is actually repeating critical EDID data of display device (I can check that just in case).
  3. Override the EDID (section 2 in the link in my sig.)
  4. Check for color problems and compare (with non-EDID override results so far).

somy's Avatar somy
11:51 AM Liked: 11
post #838 of 2707
03-03-2010 | Posts: 283
Joined: Oct 2008
Quote:
Originally Posted by Tulli View Post

Again, very persuasive, @nitpicker. Now we only need a volunteer to put the theory to test.

It would be just a matter of:
  1. Capturing both EDIDs (display device and AVR ... see section 3 in my sig.).
  2. Compare that the AVR EDID base block is actually repeating critical EDID data of display device (I can check that just in case).
  3. Override the EDID (section 2 in the link in my sig.)
  4. Check for color problems and compare (with non-EDID override results so far).

Very exiting, I would very much like to be the volunteer, but unfortunately I cannot really connect my PC directly to my projector (cable is too short)
Tulli's Avatar Tulli
12:25 PM Liked: 87
post #839 of 2707
03-03-2010 | Posts: 5,219
Joined: Nov 2001
@somy, which projector do you own? Is it listed on the EDID Overrides/bitstreaming thread 1st. post?
somy's Avatar somy
12:47 PM Liked: 11
post #840 of 2707
03-03-2010 | Posts: 283
Joined: Oct 2008
Quote:
Originally Posted by Tulli View Post

@somy, which projector do you own? Is it listed on the EDID Overrides/bitstreaming thread 1st. post?

Hi Tulli,

My projector is Panasonic PT-AE3000, and it's not on your list. Last time I tried to post EDID from my receiver (onkyo 906) but that didn't help.
The strange thing is, HD4850 with CCC9.12+ works fine with my projector, there was no colour issue though I still didn't get RGB full to work. This leads me to think this is not EDID related. Anyway, it's a good idea to give a try.

Reply Home Theater Computers

Subscribe to this Thread

Powered by vBadvanced CMPS v3.2.3