View Full Version : ATI Radeon HD 2X00 (2400,2600,2900) series owners thread
Exactly. I guess it is time for trouble ticket. 1080p50 worked also for me until I put a HDMI splitter in between my HD2600Pro and display. Now, as 1080p50 is not in HDTV support there is no easy way to force it.
If it works without the splitter then it's likely that ATi will suggest removing the splitter :rolleyes:
I guess that the splitter may well be messing with the PC's ability to obtain the EDID.
Have you tried producing a report (similar to SugoE's) to see if the EDID is correct?
Wo0zy
But it's confusing me below in the box it means:
These options are intended only for displays that report incomplete or
incorrect capabilities in their EDID information.:confused:
Anyway 1080p50 works perfectly for me, which was not the case with CCC 8.5 with the same settings.
I'm fairly sure it's only the top box on the page that this message refers to. In other words "here's a list of other standard timings but your display isn't advertising that they are supported so make them selectable at your own risk". Once you select (tick) one of the timings from the top box, it becomes available i.e. is "moved" to the bottom box, where you can apply/edit it along with the supported timings.
Anyway, at least it works :)
Wo0zy
If it works without the splitter then it's likely that ATi will suggest removing the splitter :rolleyes:
I guess that the splitter may well be messing with the PC's ability to obtain the EDID.
Have you tried producing a report (similar to SugoE's) to see if the EDID is correct?
Wo0zy
I was hoping that adding one resolution would be an easy task and they would get a nicer looking release note that way...
I believe without checking that the splitter does not have great EDID handling and hence 1080p50 goes missing from resolutions in display manager.
I was hoping that adding one resolution would be an easy task and they would get a nicer looking release note that way...
I believe without checking that the splitter does not have great EDID handling and hence 1080p50 goes missing from resolutions in display manager.
If you post the EDID you do get we can check. Also Powerstrip can probably be used to force the issue.
Wo0zy
I've realised I know nothing about EDID. Is there a simple guide out there?
According to Moninfo my TV supports 48-62Hz and the only two resolutions mentioned are 720x480@50Hz and 1920x1080@60Hz. However, in practice the card drivers seem to realise it can use just about any resolution and refresh between the two including 800x600, 720p, 1080i etc.
I've realised I know nothing about EDID. Is there a simple guide out there?
According to Moninfo my TV supports 48-62Hz and the only two resolutions mentioned are 720x480@50Hz and 1920x1080@60Hz. However, in practice the card drivers seem to realise it can use just about any resolution and refresh between the two including 800x600, 720p, 1080i etc.
Hi jong1,
I learnt everything I know from Archibael and being involved in the DTDCalc thread (it's a tool for creating custom resolutions for systems with Intel Graphics). One of DTDCalcs features is an EDID Interpreter. Simply copy your EDID into it and click the "interpret" button and it will list all supported video and audio modes contained within.
Alternatively, http://en.wikipedia.org/wiki/EDID is a look introduction to learning EDID structures and values.
Your can get DTDCalc Here (http://www.clevertec.co.uk/productsfree.htm). Obviously only the "interpret" function will work on an ATi-based system but it saves a lot of reading. The EDID needs to contain the data only (no line numbers). In other words 8 (or 16) lines or 16 Hexadecimal values
For example
02 03 1e 77 4f 94 13 12 16 11 15 1f 05 04 03 07
02 06 10 01 23 09 07 07 65 03 0c 00 30 00 01 1d
00 bc 52 d0 1e 20 b8 28 55 40 c4 8e 21 00 00 1e
02 3a 80 d0 72 38 2d 40 10 2c 45 80 c4 8e 21 00
00 1e 01 1d 80 18 71 1c 16 20 58 2c 25 00 c4 8e
21 00 00 9e 8c 0a d0 8a 20 e0 2d 10 10 3e 96 00
c4 8e 21 00 00 18 02 3a 80 18 71 38 2d 40 58 2c
45 00 c4 8e 21 00 00 1e 00 00 00 00 00 00 00 71
Is good but
Raw EDID extension (CEA-861)
00: 02 03 1E 77 4F 94 13 12 16 11 15 1F 05 04 03 07
10: 02 06 10 01 23 09 07 07 65 03 0C 00 10 00 01 1D
20: 00 BC 52 D0 1E 20 B8 28 55 40 C4 8E 21 00 00 1E
30: 02 3A 80 D0 72 38 2D 40 10 2C 45 80 C4 8E 21 00
40: 00 1E 01 1D 80 18 71 1C 16 20 58 2C 25 00 C4 8E
50: 21 00 00 9E 8C 0A D0 8A 20 E0 2D 10 10 3E 96 00
60: C4 8E 21 00 00 18 02 3A 80 18 71 38 2D 40 58 2C
70: 45 00 C4 8E 21 00 00 1E 00 00 00 00 00 00 00 91
Won't work.
If you have a 16-line EDID (most modern TVs will), the first 8-lines contain the basic EDID information while the second 8 lines contain the Extended (Version 3) data.
Within the EDID you will find video modes defined in three ways
"Established Timings"- Codes which relate to VESA standard timing specifications.
"SVDs" (Short Video Descriptors)- A shorthand method of specifying which CEA/EIA-861B standard timings are supported
"DTD" (Detailed Timing Descriptors)- Blocks of 18 Hex values (bytes) which define all of the information regarding specific timings supported by the display.
Most displays will have at least one DTD (but not always) which is used to define the detailed timing characteristics of the displays "native" or preferred resolution(s).
Obviously theres a stack more information as well but understanding the video timings and audio support parts are the most useful to us.
I'd suggest having a play with DTDCalc for a quick answer to your EDID question then reading the Wiki for a detailed description then PM me or Archibael with any specific questions. I think Ive got a reasonable grasp of it now but, like I said, Archi is the real expert.
Hope this helps,
Wo0zy
PS Sorry for the long post.
Thanks for that Wo0zy,
I will definitely have a play later.
HT Slider
06-01-08, 02:49 PM
I've realised I know nothing about EDID. Is there a simple guide out there?
According to Moninfo my TV supports 48-62Hz and the only two resolutions mentioned are 720x480@50Hz and 1920x1080@60Hz. However, in practice the card drivers seem to realise it can use just about any resolution and refresh between the two including 800x600, 720p, 1080i etc.
In reality, unless the video card interprets your HDTV as a generic PC monitor, it actually uses HDTV formats for everything. Most likely, as well as 720x480 and 1920x1080, the driver also probably assumes it supports 1280x720@60Hz giving it 720x480, 1280x720 and 1920x1080 to work with.
When you are using 800x600 for example, it scales all output into 1280x720 and sends it to your HDTV in 720p. When using 1280x1024, it scales all output into 1920x1080 and sends it to your HDTV in 1080i (or p if your HDTV support it).
Also, if the video card decides that your HDTV provides an overscanned image, instead of scaling everything into the native 720p or 1080i/p resolutions, it scales everything into the "custom resolution" for 720p or 1080i/p instead. In other words, lets say your HDTV has 4% overscan in horizontal and 3% in vertical so you have created custom resolutions of 1232x702 and 1848x1048. The driver now is smart enough to use these custom "window" resolutions (I'm calling them a "window" because it still uses the 720p or 1080i/p full resolution to send to the card, but only the "window" in use is the custom resolution you created) for every single PC resolution.
In other words your 800x600 gets scaled into 1232x702 and then the center 1232x702 pixels are used within the full 1280x720 using 720p timing. Your HDTV still receives 720p, but it looks like 800x600 from your software's perspective. As far as you are concerned you see what looks like a perfect 800x600 displayed on your HDTV, even though the HDTV doesn't actually receive 800x600 at all.
The EDID that your HDTV sends to the video card during handshaking is what tells the driver which standard HDTV formats your HDTV supports (and if it supports true PC resolutions and timings; I'm not aware of any HDTV that does actually support true PC resolutions and timings though). By using Powerstrip to create a custom driver or creating your own .inf driver for the HDTV and then installing it, you can, to some extent, override the EDID data and allow the video card to drive the HDTV using additional resolutions and timings. These days you don't usually need to create a custom driver since CCC allows you to force most standard HDTV formats (like 1080p/50, 1080p/60, etc.), but you can still create a custom driver to work around certain limitations.
I played around with custom drivers for my Toshiba 51H83 quite a bit and always found it difficult to get multiple HDTV formats to work at the same time. In general, if you provide a range, the video card sometimes thinks it can actually use the range of timings when in reality an HDTV can only handle very specific timings. You can add a small, specific frequency range and specific resolutions and then tell the video driver to treat it as an HDTV. This worked in some situations for me (but I still couldn't get all of the non-standard formats that my HDTV supports like 540p, 960i, and 1440i to work without using Powerstrip). Powerstrip has the ability to completely override video timings so you can drive your HDTV with absolutely anything if you use it in combination with a custom HDTV driver.
Yeah. Good spot on 800x600. I knew that. I don't know what made me think earlier it was supported natively.
The TV seems to declare it supports 480p, 576i,720p (50/60hz), 1080i(50/60Hz), 1080p (50/60Hz) at least and I have got 1080p/48Hz working using Powerstrip. I'm just surprised hardly any of these appear in the Moninfo report. i'll be interested to see what DTDCalc manages to find!
miami_sin
06-01-08, 04:21 PM
No, that's only for PDVD Bluray/HDDVD playback. You can't do anything about accelerating older 1080p x264 files - the problem is the files break the playback standards. Ones made recently are tending to work though.
Quote:
Originally Posted by miami_sin
I found your explanation thread...
Do I have to adjust the corresponding value?
Quote:
Originally Posted by arfster
No, that's only for PDVD Bluray/HDDVD playback. You can't do anything about accelerating older 1080p x264 files - the problem is the files break the playback standards. Ones made recently are tending to work though.
Oh, OK... Didn't know that! Thanks for your quick reply though!
Then I'll go for 720p And hope newer 1080p are working!
-------------------------------------------------------------------
Any news on the above? :confused:
I'm still trying to get the 1080p stuff H/W-accelerated in Media Player Classic...
(720p is working perfectly in MPC, 1080-BluRay-Untouched are also working in Power DVD with 5% CPU)
Thanks in advance!
If you post the EDID you do get we can check. Also Powerstrip can probably be used to force the issue.
Wo0zy
OK, here goes
$ffffffffff ----------------------------------------------------------------
$ffffffffff Monitor EDID
$ffffffffff ----------------------------------------------------------------
$ffffffffff +00+01+02+03+04+05+06+07+08+09+0a+0b+0c+0d+0e+0f
$0c00000000 +00 00 ff ff ff ff ff ff 00 0f 30 22 91 00 00 00 00
$0c00000001 +10 32 10 01 03 80 3c 22 78 2a 03 20 a7 55 45 96 24
$0c00000002 +20 11 49 4b 1f df 00 45 59 a9 40 81 80 31 59 01 01
$0c00000003 +30 01 01 01 01 01 01 01 1d 80 18 71 1c 16 20 58 2c
$0c00000004 +40 25 00 5b 56 21 00 00 9e 8c 0a d0 8a 20 e0 2d 10
$0c00000005 +50 10 3e 96 00 5b 56 21 00 00 18 00 00 00 fc 00 43
$0c00000006 +60 41 54 2d 36 30 36 36 0a 20 20 20 20 00 00 00 fd
$0c00000007 +70 00 38 4b 1f 32 08 00 0a 20 20 20 20 20 20 01 e5
$0c00000008 +80 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
$0c00000009 +90 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
$0c0000000a +a0 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
$0c0000000b +b0 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
$0c0000000c +c0 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
$0c0000000d +d0 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
$0c0000000e +e0 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
$0c0000000f +f0 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
I have RivaTuner installed for adjusting fan speed and it seems to have some resolution creation functionality. Have to look into that. I also tried to use moninfo, but it returns silly numbers and wont let me select display.
arfster
06-01-08, 04:49 PM
Any news on the above? :confused:
I'm still trying to get the 1080p stuff H/W-accelerated in Media Player Classic...
Like before: if the files aren't encoded properly, they won't work.
In reality, unless the video card interprets your HDTV as a generic PC monitor, it actually uses HDTV formats for everything. Most likely, as well as 720x480 and 1920x1080, the driver also probably assumes it supports 1280x720@60Hz giving it 720x480, 1280x720 and 1920x1080 to work with.
When you are using 800x600 for example, it scales all output into 1280x720 and sends it to your HDTV in 720p. When using 1280x1024, it scales all output into 1920x1080 and sends it to your HDTV in 1080i (or p if your HDTV support it).
Also, if the video card decides that your HDTV provides an overscanned image, instead of scaling everything into the native 720p or 1080i/p resolutions, it scales everything into the "custom resolution" for 720p or 1080i/p instead. In other words, lets say your HDTV has 4% overscan in horizontal and 3% in vertical so you have created custom resolutions of 1232x702 and 1848x1048. The driver now is smart enough to use these custom "window" resolutions (I'm calling them a "window" because it still uses the 720p or 1080i/p full resolution to send to the card, but only the "window" in use is the custom resolution you created) for every single PC resolution.
In other words your 800x600 gets scaled into 1232x702 and then the center 1232x702 pixels are used within the full 1280x720 using 720p timing. Your HDTV still receives 720p, but it looks like 800x600 from your software's perspective. As far as you are concerned you see what looks like a perfect 800x600 displayed on your HDTV, even though the HDTV doesn't actually receive 800x600 at all.
The EDID that your HDTV sends to the video card during handshaking is what tells the driver which standard HDTV formats your HDTV supports (and if it supports true PC resolutions and timings; I'm not aware of any HDTV that does actually support true PC resolutions and timings though). By using Powerstrip to create a custom driver or creating your own .inf driver for the HDTV and then installing it, you can, to some extent, override the EDID data and allow the video card to drive the HDTV using additional resolutions and timings. These days you don't usually need to create a custom driver since CCC allows you to force most standard HDTV formats (like 1080p/50, 1080p/60, etc.), but you can still create a custom driver to work around certain limitations.
I played around with custom drivers for my Toshiba 51H83 quite a bit and always found it difficult to get multiple HDTV formats to work at the same time. In general, if you provide a range, the video card sometimes thinks it can actually use the range of timings when in reality an HDTV can only handle very specific timings. You can add a small, specific frequency range and specific resolutions and then tell the video driver to treat it as an HDTV. This worked in some situations for me (but I still couldn't get all of the non-standard formats that my HDTV supports like 540p, 960i, and 1440i to work without using Powerstrip). Powerstrip has the ability to completely override video timings so you can drive your HDTV with absolutely anything if you use it in combination with a custom HDTV driver.
I agree with your assessment regarding how HDTVs handle/scale PC resolutions and how "custom timings" are actually standard timings with a reduced number of active pixels and increased blanking interval however, don't forget that EDID version 3 data deals with HDTV timings. The SVDs define the supported EIA/CEA standard timings (including the popular 720,1080i and 1080p resolutions) and DTD's are still be used as a method of further defining a given display's preferred timing parameters for these resolutions (decode an EDID from any modern HDTV and you'll find version 3 data including both SVDs and DTDs). I don't know if ATI drivers ignore DTDs in favour of just checking SVDs and delivering standard timings but it would be a shame if that was the case. In either case, there is no assumption. The standard offerings are defined by EDID.
Powerstrip can certainly be used to override EDID parameters (that's kind of the whole point) and CCC's HDTV Support options can be used to "force" an unsupported HDTV timing if you desire but EDID still plays a major role in deciding what video timings (and audio formats) are supported and offered as standard within a driver.
Wo0zy
OK, here goes
$ffffffffff ----------------------------------------------------------------
$ffffffffff Monitor EDID
$ffffffffff ----------------------------------------------------------------
$ffffffffff +00+01+02+03+04+05+06+07+08+09+0a+0b+0c+0d+0e+0f
$0c00000000 +00 00 ff ff ff ff ff ff 00 0f 30 22 91 00 00 00 00
$0c00000001 +10 32 10 01 03 80 3c 22 78 2a 03 20 a7 55 45 96 24
$0c00000002 +20 11 49 4b 1f df 00 45 59 a9 40 81 80 31 59 01 01
$0c00000003 +30 01 01 01 01 01 01 01 1d 80 18 71 1c 16 20 58 2c
$0c00000004 +40 25 00 5b 56 21 00 00 9e 8c 0a d0 8a 20 e0 2d 10
$0c00000005 +50 10 3e 96 00 5b 56 21 00 00 18 00 00 00 fc 00 43
$0c00000006 +60 41 54 2d 36 30 36 36 0a 20 20 20 20 00 00 00 fd
$0c00000007 +70 00 38 4b 1f 32 08 00 0a 20 20 20 20 20 20 01 e5
$0c00000008 +80 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
$0c00000009 +90 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
$0c0000000a +a0 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
$0c0000000b +b0 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
$0c0000000c +c0 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
$0c0000000d +d0 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
$0c0000000e +e0 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
$0c0000000f +f0 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
I have RivaTuner installed for adjusting fan speed and it seems to have some resolution creation functionality. Have to look into that. I also tried to use moninfo, but it returns silly numbers and wont let me select display.
Ok, that EDID block contains the following Established Timings
40 x 480 @ 67 Hz
640 x 480 @ 72 Hz
640 x 480 @ 75 Hz
800 x 600 @ 56 Hz
800 x 600 @ 60 Hz
800 x 600 @ 72 Hz
800 x 600 @ 75 Hz
1024 x 768 @ 87 Hz (Interlaced)
1024 x 768 @ 60 Hz
1024 x 768 @ 70 Hz
1024 x 768 @ 75 Hz
1280 x 1024 @ 75 Hz
It also contains two DTD's
01 1D 80 18 71 1C 16 20 58 2C 25 00 5B 56 21 00 00 9E (1920x1080i)
8C 0A D0 8A 20 E0 2D 10 10 3E 96 00 5B 56 21 00 00 18 (1280x720p)
However, you may have noticed that the second 8 lines are all zeros. This means the extended EDID information isn't being parsed. This second block is likely to contain either the DTDs and/or SVD's (see my previous posts) which define your 1080p50 timing.
If 1080p60 is definitely offered with the splitter then, as per HT's post, the ATI drivers are just offering you what is considered a "standard" HDTV timing but can't build on this by reading the EDID information.
Is it possible that you can temporarily disconnect the splitter and produce another report with the PC directly connected to the TV? If so and the Extended EDID does contain a DTD for 1080p50, you can use this information with Powerstrip to produce the exact timing you need. If it only offers 1080p50 in SVD then using a generic (EIA/CEA) timing with Powerstrip will work fine. In either case you may have to adjust to compensate for overscan.
Wo0zy
edavis530
06-01-08, 11:40 PM
You'll have to dload powerstrip then. Can't remember quite how, but it was pretty easy to do - you just make a new driver with 1920*1080 (or whatever) as the max possible.
How do I make my own driver? Anyone know?
I've never used powerstrip.. Also is this a free download?
HT Slider
06-02-08, 12:36 AM
I agree with your assessment regarding how HDTVs handle/scale PC resolutions and how "custom timings" are actually standard timings with a reduced number of active pixels and increased blanking interval however, don't forget that EDID version 3 data deals with HDTV timings. The SVDs define the supported EIA/CEA standard timings (including the popular 720,1080i and 1080p resolutions) and DTD's are still be used as a method of further defining a given display's preferred timing parameters for these resolutions (decode an EDID from any modern HDTV and you'll find version 3 data including both SVDs and DTDs). I don't know if ATI drivers ignore DTDs in favour of just checking SVDs and delivering standard timings but it would be a shame if that was the case. In either case, there is no assumption. The standard offerings are defined by EDID.
Powerstrip can certainly be used to override EDID parameters (that's kind of the whole point) and CCC's HDTV Support options can be used to "force" an unsupported HDTV timing if you desire but EDID still plays a major role in deciding what video timings (and audio formats) are supported and offered as standard within a driver.
Wo0zy
I guess I'm a little out of date when it comes to typical HDTV EDIDs. I didn't realize the newer HDTVs were actually reporting support of PC resolutions and higher refresh rates. I haven't really looked at EDIDs in detail for a few years and back in ~2005, all of the EDIDs I examined from HDTVs all only supported HDTV resolutions and formats (plus their single native resolution - like 1366x768). There were no 800x600's, 1024x768's or 1280x1024's in the ones I looked at back then. I'm also still under the impression that ATI only outputs HDTV formats to anything it considers an HDTV, with the specific exception to allowing 1360x768 at 60Hz. I could be wrong.
karrih, which HDTV is that EDID from?
This is even more of a reason to consider upgrading to a modern 1080p digital display (especially if it has a fast enough pixels to benefit from above 60 Hz for video games, etc.).
karrih, with your HDTV hooked up, can you select 1280x1024 at 75Hz from CCC? If it shows 1280x1024 at only 60Hz, my suspicion is it is still using 1080p for the actual output (with the current driver).
HT Slider
06-02-08, 01:00 AM
edavis530,
Powerstrip can be used for free (at least it used to), but it pops up annoying "buy me messages" unless you pay for it. It is still a great tool though (fully functional trial) and I did pay for it myself. Have a look at http://entechtaiwan.com/index.shtm for a download and additional information.
woozy,
Maybe you could take a look at my EDID and see if there is an explanation for the grey level performance I am seeing with the HDMI dongle in use (and still requiring brightness/contrast tweaks). Personally I have no idea how to interpret the color characteristics.
Here is the interpretation as well as the EDID from moninfo (as you can see my HDTV doesn't even list 720p, but it does accept and display it; as well as 540p, 960i and 1440i if I use Powerstrip):
Monitor
Windows description......... Generic PnP Monitor
Manufacturer description.... Toshiba TV
Manufacturer................ Toshiba
————————————————————————————
Plug and Play ID............ TSB0100
Serial number............... 168290
EDID data source............ Registry (stored)
————————————————————————————
Manufacture date............ 2003, ISO week 31
EDID revision............... 1.3
Display type and signal..... Digital
Sync input support.......... n/a
Screen size................. n/a
Power management............ n/a
Color characteristics
Display gamma............... 1.00
Red chromaticity............ Rx 0.640 - Ry 0.330
Green chromaticity.......... Gx 0.300 - Gy 0.600
Blue chromaticity........... Bx 0.150 - By 0.060
White point (default)....... Wx 0.313 - Wy 0.329
Timing characteristics
VESA GTF support............ Not supported
Horizontal scan range....... 31-46kHz
Vertical scan range......... 59-61Hz
Video bandwidth............. 80MHz
Extension blocks............ 1
Timing recommendation #1.... 1920x540 at 60Hz
Modeline................ "1920x540" 74.250 1920 2008 2052 2200 540 542 547 562 +hsync +vsync
Timing recommendation #2.... 1920x1080 at 30Hz
Modeline................ "1920x1080" 74.250 1920 2008 2052 2200 1080 1084 1094 1124 interlace +hsync +vsync
Timing recommendation #3.... 720x480 at 60Hz
Modeline................ "720x480" 27.000 720 736 798 858 480 489 495 525 -hsync -vsync
Standard timings supported
640 x 480 at 60Hz - IBM VGA
720 x 480 at 60Hz - Toshiba
1920 x 540 at 60Hz - Toshiba
1920 x 1080 at 30Hz - Toshiba
Raw EDID base
00: 00 FF FF FF FF FF FF 00 52 62 00 01 62 91 02 00
10: 1F 0D 01 03 80 00 00 00 0E EE 91 A3 54 4C 99 26
20: 0F 50 54 20 00 00 01 01 01 01 01 01 01 01 01 01
30: 01 01 01 01 01 01 01 1D 80 18 71 1C 16 20 58 2C
40: 25 00 00 00 00 00 00 9E 8C 0A D0 8A 20 E0 2D 10
50: 10 3E 96 00 00 00 00 00 00 18 00 00 00 FC 00 54
60: 6F 73 68 69 62 61 20 54 56 0A 20 20 00 00 00 FD
70: 00 3B 3D 1F 2E 08 00 0A 20 20 20 20 20 20 01 E2
mariner888
06-02-08, 02:05 AM
I suspect the CEA extension block in EDID is not read into the registry by the latest Catalyst drivers. Not sure if it's by design or a bug.
No problem with the Forceware drivers.
Is it possible that you can temporarily disconnect the splitter and produce another report with the PC directly connected to the TV? If so and the Extended EDID does contain a DTD for 1080p50, you can use this information with Powerstrip to produce the exact timing you need. If it only offers 1080p50 in SVD then using a generic (EIA/CEA) timing with Powerstrip will work fine. In either case you may have to adjust to compensate for overscan.
Wo0zy
I guess I'll try direct connection too. Some kind of cut and paste EDID operation for 1080p50 would indeed be ideal.
karrih, which HDTV is that EDID from?
karrih, with your HDTV hooked up, can you select 1280x1024 at 75Hz from CCC? If it shows 1280x1024 at only 60Hz, my suspicion is it is still using 1080p for the actual output (with the current driver).
When I took the report I had Panasonic TH42-PX80 plasma and Mitsubishi HC1100 projector connected using HDMI splitter. Mits was in standby, though. I don't know how splitter handles EDID from those two displays, but we might find out soon...
Using 1080p when choosing 1280x1024 sounds odd, but then I haven't quite figured out either why choosing 1080i25 in CCC gets reported as 1080i50 by my pj.
however, don't forget that EDID version 3 data deals with HDTV timings. The SVDs define the supported EIA/CEA standard timings (including the popular 720,1080i and 1080p resolutions) and DTD's are still be used as a method of further defining a given display's preferred timing parameters for these resolutions (decode an EDID from any modern HDTV and you'll find version 3 data including both SVDs and DTDs). I don't know if ATI drivers ignore DTDs in favour of just checking SVDs and delivering standard timings but it would be a shame if that was the case. In either case, there is no assumption. The standard offerings are defined by EDID.
I know how older ATI drivers behaved. Haven't checked out the latest drivers yet, though. With older drivers ATI did pay attention to the DTDs. However, IMO it's actually ***BAD*** to do that. Why? Because the pixel clock in DTDs is just plain wrong for 24Hz and 60Hz. What we need for HTPC playback is 24/1.001 and 60/1.001 and the DTDs don't have enough decimal digits for that purpose. So a standalone Blu-Ray player actually ignores at least the pixel clock in the DTD and uses EIA/CEA standard timings instead. ATI didn't do that in older drivers and that is BAD. Furthermore ATI interpreted the SVDs in such a way that 24Hz is 24.000Hz and 60Hz is 60.000Hz. That is wrong, too. It should be 23.976Hz and 59.940Hz.
But as I said that's all with older drivers. Have yet to check the latest drivers where according to the release notes some changes were made to this stuff.
@Wo0zy,
Do you have a suggested tool for collecting the EDID data to go into DTDCalc?
Thanks
Is it possible that you can temporarily disconnect the splitter and produce another report with the PC directly connected to the TV?
Turns out that if I do this with HDMI connection, at least RivaTuner reports it can not read EDID.
@Wo0zy,
Do you have a suggested tool for collecting the EDID data to go into DTDCalc?
Thanks
With Intel Drivers the built-in diagnostic report contains the EDID (very helpful) but for other vendors I usually use Moninfo (http://www.entechtaiwan.com/util/moninfo.shtm)
I think you can also extract EDID information from within Powerstrip (not sure about that one. Will have to check).
Wo0zy
Turns out that if I do this with HDMI connection, at least RivaTuner reports it can not read EDID.
Hmmm. Try using Moninfo. I just included the download URL in my previous post.
Wo0zy
nicksan
06-02-08, 11:54 AM
Hello,
Been reading up on this thread. Took me a while and I can't deny the fact that I am indeed pretty confused about all this!
Last week, on a whim, I decided that I wanted Blue-ray playback for my Home Theater that features a Sanyo PLV-Z4 720p Projector, Sony SD DVD Player, and an antiquated P4 Northwood 2.8 PC. (Asus P4P-800, 1.5GB DDR, Creative Audigy SE)
So rather than buying a PS3 or building a new PC from scratch I decided to buy Sony's BD-ROM and the Sapphire Radeon 2600XT AGP 8X card. I also have the full version of PowerDVD 8 ultra with most recent update. (Also made sure I saved the older codec since I am aware that the latest update disables HA.)
I installed the latest ATI Catalyst hotfix (8.5) and that's the only driver I installed. This machine never had an ATI card installed so I guess it's like a fresh install in this sense. I read about the Omega drivers but then I saw the 8.5 hotfix so decided to give that a go instead. I did apply the registry hack on the guide however.
I am running audio to spdif, output via Audigy SE and into my older Pioneer Elite receiver. It's Dolby Digital/DTS 5.1 via spdif.
I bought Casino Royale Blue-ray and that actually plays back fine in PowerDVD with HA enabled. The CPU typicall shows 40%-45% with spikes that bring it up to 50%-55%...definitely not more than 60%. No stuttering at all. Mind you, this is all in 720p. (Video card resolution set to standard 720p 1280 x 720...which is the native resolution on my Projector)
So being that this was so exciting, I decided to play the demo Blue-Ray DVD that came with the Sony BD-ROM. (Has trailers for Spiderman 3, Casino Royale, Surf's up, etc...) For whatever reason this one stutters like mad. 100% CPU despite the fact that HA is enabled.
So in PowerDVD I checked the "Information" tab and compare between this and the Casino Royale BR DVD that works well and they look pretty much the same EXCEPT...the audio encoding.
Casino Royale = Dolby Digital
Sony BD-ROM Sample BR DVD = MLP
I know MLP is more "advanced" in that it is lossless but could this actually affect the HA and and cause stuttering? In PowerDVD since I send the audio to spdif, and because the receiver doesn't do MLP, I choose to mix it down to either DD or DTS. This has no effect. It stutters like crazy!
I borrowed a BR DVD from my co-worker, Oceans 13, and will see how that plays, however I am concerned about all this. The last thing I want to do is to upgrade my machine to a Core2Duo...my wife would kill me!:eek:
So gathering from what I have described, is there anything I am doing wrong? Again, Casino Royale BR DVD plays fine. So it's not a complete wash. There is something on that Sony BD-Rom sample BR DVD that is causing the CPU to spike up and cause stutter and the only difference I can tell is, again, the MLP audio vs. regular DD 5.1 on Casino Royale.
Any help would be greatly appreciated.
I am using an ATI 3850.
I already had looked at Moninfo, but that only has the 2 timings in it. Posted the raw data into DTDCalc and I get the same.
I am guessing I am not getting the extended stuff, if it exists. I know it's a bit off-topic but any tips?
I also have the full version of PowerDVD 8 ultra with most recent update. (Also made sure I saved the older codec since I am aware that the latest update disables HA.)
Casino Royale = Dolby Digital
Sony BD-ROM Sample BR DVD = MLP
I know MLP is more "advanced" in that it is lossless but could this actually affect the HA and and cause stuttering?
Any help would be greatly appreciated.
I am also running underpowered setup with HA and until PDVD Ultra 7.3 build 4102 I was not able to use Dolby TrueHD without major stutter. Now I can use TrueHD with 720p but not with 1080p due to the HA problems you mention. The secondary audio stream usually worked for me, though.
nicksan
06-02-08, 12:51 PM
I take it Dolby TrueHD is the same thing as MLP?
I also do have PDVD 7.3 installed and it stutters. I am not sure if I have 4102 though...I think it's lower than that so perhaps it's worth a try.
You would figure though that with PDVD 8 Ultra being a newer version they would have ironed out the kinks.
Perhaps it's specific to my system?
When you say secondary audio stream, you mean when I down-mix the MLP to DD 5.1 and send it to spdif?
Perhaps this problem is specific to the Sampler BR DVD in question. I am tempted to upgrade my rig, however I guess the prudent thing to do is to just buy/rent some BR DVD's and start watching them.
The thing is though, I am wondering if this problem is with BR DVD's authored more recently and that I won't have problems with BR DVD's authored in 2007 but the newer release will have this problem?
In these BR DVD's is there an audio option screen like regular DVD's? On my Casino Royale BR DVD I cound't get the menu to come up and had to screw around to get the movie to start. Is that a common problem?
Maybe I should just invest in a G35 motherboard + Core2Duo (or quad)....
I am also running underpowered setup with HA and until PDVD Ultra 7.3 build 4102 I was not able to use Dolby TrueHD without major stutter. Now I can use TrueHD with 720p but not with 1080p due to the HA problems you mention. The secondary audio stream usually worked for me, though.
Hmmm. Try using Moninfo. I just included the download URL in my previous post.
Wo0zy
Here it is, but doesn't look too promising to me. EDID source registry, eh?
Monitor
Windows description......... Generic Television
Manufacturer................ PNP
Plug and Play ID............ PNP09FE
Serial number............... n/a
EDID data source............ Registry (stored)
Manufacture date............ 1990
EDID revision............... 1.3
Display type and signal..... Analog 0.700,0.300 (1.0V p-p)
Sync input support.......... n/a
Screen size................. n/a
Power management............ Standby, Suspend, Active off/sleep
Color characteristics
Display gamma............... 1.00
Red chromaticity............ Rx 0.000 - Ry 0.000
Green chromaticity.......... Gx 0.000 - Gy 0.000
Blue chromaticity........... Bx 0.000 - By 0.000
White point (default)....... Wx 0.000 - Wy 0.000
Timing characteristics
VESA GTF support............ Not supported
Horizontal scan range....... 30-48kHz (est.)
Vertical scan range......... 48-60Hz (est.)
Video bandwidth............. 64MHz (est.)
Extension blocks............ n/a
Standard timings supported
640 x 480 at 60Hz - IBM VGA
800 x 600 at 60Hz - VESA
1024 x 768 at 60Hz - VESA
Raw EDID base
00: 00 FF FF FF FF FF FF 00 41 D0 FE 09 00 00 00 00
10: 00 00 01 03 00 00 00 00 F0 00 00 00 00 00 00 00
20: 00 00 00 21 08 00 00 00 00 00 00 00 00 00 00 00
70: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 D1
I take it Dolby TrueHD is the same thing as MLP?
When you say secondary audio stream, you mean when I down-mix the MLP to DD 5.1 and send it to spdif?
In these BR DVD's is there an audio option screen like regular DVD's? On my Casino Royale BR DVD I cound't get the menu to come up and had to screw around to get the movie to start. Is that a common problem?
Right now I can't remember if TrueHD uses MLP, but it could well be.
In addition to TrueHD there is often option to use another audio format, for example PCM or DD 5.1 (not downmixed) and those are lighter for processor. You can change audio streams using buttons in PDVD when there is no audio menu at hand.
nicksan
06-02-08, 01:24 PM
I looked it up and it looks to be the case regarding TrueHD=MLP/lossless.
I am aware of the PDVD audio stuff...what I do is select spdif as the output and it comes up with option whether to use Primary Audio or something else. (I forget)
I guess using Primary just passes the MLP straight up and my receiver only does 2.1. There is another option to down-mix to DD 5.1 / DTS. I choose this option and my receiver does 5.1.
However I am not sure when down-mixing, it is still processing MLP as well and the down mix is secondary?
Got a headache now!:eek::)
Right now I can't remember if TrueHD uses MLP, but it could well be.
In addition to TrueHD there is often option to use another audio format, for example PCM or DD 5.1 (not downmixed) and those are lighter for processor. You can change audio streams using buttons in PDVD when there is no audio menu at hand.
nicksan
06-02-08, 02:04 PM
Also I am wondering maybe there is a problem with the fact that there is no menu on the sampler BR DVD that allows me to choose the audio format. Seems to force using MLP...maybe that's what is causing the 100% spike and stutter?
In Casino Royale, even though I cannot for the life of me reach or see the main menu, it is selecting DD 5.1...maybe it's selected by default...or that it doesn't have TrueHD being that it's an "older" release?
Who knows...but it's kind of difficult to figure out a pattern unless I view more BR DVD's I suppose. I guess I shouldn't let the sampler BR DVD that came with the Sony BD-Rom drive to scare me into upgrading my PC!
For those BR DVD that have TrueHD, do they typically have "separate" DD5.1 tracks on them as well? Perhaps if that is the case, I have nothing to be concerned about. Not sure how all that works...
As for my h.264 file, it seems to run fine, whether I am using PDVD or MPC w/CoreAVC codecs. So for simplicity I may just use PDVD as my only player.
Curious to see how the others like WinDVD and Arcsoft performs. Too bad WinDVD9 trial cannot do BR. Stupid...just plain stupid. How am I supposed to evaluate it for BR?:confused:
Right now I can't remember if TrueHD uses MLP, but it could well be.
In addition to TrueHD there is often option to use another audio format, for example PCM or DD 5.1 (not downmixed) and those are lighter for processor. You can change audio streams using buttons in PDVD when there is no audio menu at hand.
I am using an ATI 3850.
I already had looked at Moninfo, but that only has the 2 timings in it. Posted the raw data into DTDCalc and I get the same.
I am guessing I am not getting the extended stuff, if it exists. I know it's a bit off-topic but any tips?
Hmm. I tried Moninfo for the first time on my ATI system and I get the same problem. It does the same thing with Intel Graphics (not that it's a problem as we can get a full EDID from the driver diags report) however, the last time I used it with an NVIDIA card it captured a complete EDID including the extended (1.3) data. Very odd.
Not sure what to suggest. The Phoenix EDID editor does the same thing and I can't get the Viewsonic editor to work properly under Vista. :(
I'll ask Archibael if he's aware of anything.
AFAIK you can't export EDID from within Powerstrip but if you right-click on the taskbar icon and select Options>Monitor Information, it lists supported timings. You could try this and see if if offers you any additional timings.
Sorry
Wo0zy
AFAIK you can't export EDID from within Powerstrip but if you right-click on the taskbar icon and select Options>Monitor Information, it lists supported timings. You could try this and see if if offers you any additional timings.No, sadly this gives the same minimal info. Maybe that IS all that is there and it guesses the rest. I really do not know. Not that I have a problem, but I am intrigued now to know what gems my TV really does divulge!
HT Slider
06-02-08, 03:32 PM
No, sadly this gives the same minimal info. Maybe that IS all that is there and it guesses the rest. I really do not know. Not that I have a problem, but I am intrigued now to know what gems my TV really does divulge!
Which HDTV do you have again?
It is very strange to see only standard PC resolutions in the EDID and no HDTV resolutions.
All of the HDTV EDID's I've personally extracted and looked at only have HDTV resolutions (as well as often a standard 640x480).
I suspect that for some reason the EDID isn't being read properly and the numbers moninfo is displaying are some sort of default.
HT Slider
06-02-08, 03:33 PM
Hmm. I tried Moninfo for the first time on my ATI system and I get the same problem. It does the same thing with Intel Graphics (not that it's a problem as we can get a full EDID from the driver diags report) however, the last time I used it with an NVIDIA card it captured a complete EDID including the extended (1.3) data. Very odd.
WoOzy, did you notice my post a page back? (http://www.avsforum.com/avs-vb/showthread.php?p=13993473#post13993473)
I wondered if you might have some insight into the grey level performance I am seeing with my HDTV.
@Wo0zy,
Do you have a suggested tool for collecting the EDID data to go into DTDCalc?
Thanks
I used RivaTuner to read out EDID, maybe it's more accurate then Moninfo.
Which HDTV do you have again?
It is very strange to see only standard PC resolutions in the EDID and no HDTV resolutions.
All of the HDTV EDID's I've personally extracted and looked at only have HDTV resolutions (as well as often a standard 640x480).
I suspect that for some reason the EDID isn't being read properly and the numbers moninfo is displaying are some sort of default.It's a Sony 46X2000 but I didn't say it only reads PC resolutions.
According to Moninfo my TV supports 48-62Hz and the only two resolutions mentioned are 720x480@50Hz and 1920x1080@60Hz. However, in practice the card drivers seem to realise it can use just about any resolution and refresh between the two including 800x600, 720p, 1080i etc.
As I said before I don't know what was going through my mind when I mentioned 800x600 there. Clearly nonsense!
In practice it seems to support 480p/i, 576p/i, 720p and 1080p/i @48-60Hz and devices seem to recognise that, but I am mystified how!
I used RivaTuner to read out EDID, maybe it's more accurate then Moninfo.No. tried that one too! But thanks.
WoOzy, did you notice my post a page back? (http://www.avsforum.com/avs-vb/showthread.php?p=13993473#post13993473)
I wondered if you might have some insight into the grey level performance I am seeing with my HDTV.Maybe I am missing something, but as per my earlier post this seems clear. Unless the dongle is "in action", switching the card to "HDMI mode", then the card outputs RGB @PC-levels. It seems clear from what is appearing in your CCC that, despite the dongle, the driver is recognising, correctly, that it is in fact connected to a DVI display, that it therefore cannot use YCbCr (which always uses video levels) and so it is treating it like a PC monitor.
Is there a way that this does not fit with the facts?
I used RivaTuner to read out EDID, maybe it's more accurate then Moninfo.
Nice one SugoE. The RivaTuner EDID report does capture all information (including 1.3 extended data).
Jong1, once you create the report it'll need a bit of notepad editing before you can run it through DTDCalc....but it does work.
Wo0zy
Edit: Just noticed that you said it didn't for you?? Once I selected "Monitor EDID snapshot" and "Monitor EDID details", I got the following
$ffffffffff ----------------------------------------------------------------
$ffffffffff Monitor EDID
$ffffffffff ----------------------------------------------------------------
$ffffffffff +00+01+02+03+04+05+06+07+08+09+0a+0b+0c+0d+0e+0f
$0c00000000 +00 00 ff ff ff ff ff ff 00 52 62 04 01 00 00 00 00
$0c00000001 +10 00 10 01 03 80 c4 8e 78 0a ee 91 a3 54 4c 99 26
$0c00000002 +20 0f 50 54 00 00 00 01 01 01 01 01 01 01 01 01 01
$0c00000003 +30 01 01 01 01 01 01 01 1d 80 d0 72 1c 16 20 10 2c
$0c00000004 +40 25 80 c4 8e 21 00 00 9e 8c 0a d0 90 20 40 31 20
$0c00000005 +50 0c 40 55 00 c4 8e 21 00 00 18 00 00 00 fc 00 54
$0c00000006 +60 4f 53 48 49 42 41 2d 54 56 0a 20 20 00 00 00 fd
$0c00000007 +70 00 31 3d 0f 44 0f 00 0a 20 20 20 20 20 20 01 eb
$0c00000008 +80 02 03 1e 77 4f 94 13 12 16 11 15 1f 05 04 03 07
$0c00000009 +90 02 06 10 01 23 09 07 07 65 03 0c 00 10 00 01 1d
$0c0000000a +a0 00 bc 52 d0 1e 20 b8 28 55 40 c4 8e 21 00 00 1e
$0c0000000b +b0 02 3a 80 d0 72 38 2d 40 10 2c 45 80 c4 8e 21 00
$0c0000000c +c0 00 1e 01 1d 80 18 71 1c 16 20 58 2c 25 00 c4 8e
$0c0000000d +d0 21 00 00 9e 8c 0a d0 8a 20 e0 2d 10 10 3e 96 00
$0c0000000e +e0 c4 8e 21 00 00 18 02 3a 80 18 71 38 2d 40 58 2c
$0c0000000f +f0 45 00 c4 8e 21 00 00 1e 00 00 00 00 00 00 00 91
$ffffffffff ----------------------------------------------------------------
$ffffffffff Monitor EDID details
$ffffffffff ----------------------------------------------------------------
$0d00000000 EDID version : 1.3
$0d00000001 Manufacturer ID : TSB (Toshiba)
$0d00000002 Model ID : 0104 (TOSHIBA-TV)
$0d00000003 Manufactured on : 2006, week 0
$0d00000004 Serial number : 00000000
$0d00000005 Input signal : digital
$0d00000006 Screen size : 196 x 142 cm
$0d00000007 Horizontal frequency : 15-68KHz
$0d00000008 Vertical frequency : 49-61Hz
$0d00000009 Bandwidth : 150MHz
The EDID block matches exactly with the report that the Intel Diagnostic report produces
I'll look at it again tomorrow. I may have been too quick to judge!
WoOzy, did you notice my post a page back? (http://www.avsforum.com/avs-vb/showthread.php?p=13993473#post13993473)
I wondered if you might have some insight into the grey level performance I am seeing with my HDTV.
Sorry HT, I missed that post.
I think Jong1 is correct. Your connection isn't end-to-end HDMI therefore the connection is treated as RGB/PC rather than YCbCr/video levels. I use my main system without the dongle and with my TV calibrated (as best I can) for RGB levels and using only the "UseBT601CSC=1" reg tweak, I'm fairly happy that all my levels (desktop, SD and HD) are consistent. At least they are acceptable to my eyes. All video looks good and photos are detailed enough for casual viewing (I don't use this system for editing anyway). Obviously, switching to a YCbCr input (such as my DVB-T tuner or Sky receiver) reveals a rubbish washed out, over bright image but this doesn't bother me as I use the HTPC for just about eveything anyway :)
Regarding your EDID report, as per other posts, it only seems to contain the first block (no extended data). Perhaps you could try RivaTuner as SugoE suggests and see if it's possible to capture the Version 1.3 extended block? It's likely that the 720p timing is in there as either an SVD or DTD.
Wo0zy
@ Arfster - If you were me, would you choose the HD 2400 PRO or the HD 3650? Just figured I'd ask since i'm tryin to decide which one is better since I only use the cards for movies especially when I bought em both at differ times of course.
Here it is, but doesn't look too promising to me. EDID source registry, eh?
Monitor
Windows description......... Generic Television
Manufacturer................ PNP
Plug and Play ID............ PNP09FE
Serial number............... n/a
EDID data source............ Registry (stored)
Manufacture date............ 1990
EDID revision............... 1.3
Display type and signal..... Analog 0.700,0.300 (1.0V p-p)
Sync input support.......... n/a
Screen size................. n/a
Power management............ Standby, Suspend, Active off/sleep
Color characteristics
Display gamma............... 1.00
Red chromaticity............ Rx 0.000 - Ry 0.000
Green chromaticity.......... Gx 0.000 - Gy 0.000
Blue chromaticity........... Bx 0.000 - By 0.000
White point (default)....... Wx 0.000 - Wy 0.000
Timing characteristics
VESA GTF support............ Not supported
Horizontal scan range....... 30-48kHz (est.)
Vertical scan range......... 48-60Hz (est.)
Video bandwidth............. 64MHz (est.)
Extension blocks............ n/a
Standard timings supported
640 x 480 at 60Hz - IBM VGA
800 x 600 at 60Hz - VESA
1024 x 768 at 60Hz - VESA
Raw EDID base
00: 00 FF FF FF FF FF FF 00 41 D0 FE 09 00 00 00 00
10: 00 00 01 03 00 00 00 00 F0 00 00 00 00 00 00 00
20: 00 00 00 21 08 00 00 00 00 00 00 00 00 00 00 00
70: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 D1
Very odd! This is when directly connected to your TV via HDM with no other display connected?
It's almost like your display reports no EDID and it's picked up a generic entry of some sort :confused:
Registry (stored) is normal but the the lack of data and the fact the connection is reported as Analog is very strange.
Perhaps try again with RivaTuner as SugoE discovered.
Wo0zy
arfster
06-02-08, 07:59 PM
@ Arfster - If you were me, would you choose the HD 2400 PRO or the HD 3650? Just figured I'd ask since i'm tryin to decide which one is better since I only use the cards for movies especially when I bought em both at differ times of course.
3650 definitely. It's more powerful, as the 3650 is a 2600 equivalent, and the 2400 cards have all sorts of daft issues (can't do decent deinterlacing for a start). ATI also have a bit of a habit of giving up on old cards too, so those bugs will never be fixed.
Best of all, prices have dropped so there's not much difference between them now :-)
Perhaps try again with RivaTuner as SugoE discovered.
Wo0zy
That was my first attempt. RivaTuner reports it can not read EDID. Perhaps ATI offers all HDTV resolutions when the connection is recognized as HDMI and there is no EDID? If my splitter takes my projectors EDID and passes it as its own then ATI thinks that it knows what is available and no more offers all HDTV resolutions as standard.
That was my first attempt. RivaTuner reports it can not read EDID. Perhaps ATI offers all HDTV resolutions when the connection is recognized as HDMI and there is no EDID? If my splitter takes my projectors EDID and passes it as its own then ATI thinks that it knows what is available and no more offers all HDTV resolutions as standard.
I guess it's possible but it's very unusual for an HDTV not to have even a basic EDID, especially a known brand like Panasonic.
Didn't you produce the first report from RivaTuner with the splitter still connected?
It might be worth clearing out the monitor entries from your registry and trying a RivaTuner report again without the splitter.
Wo0zy
For me Rivatuner also reports no extended data. It shows a table covering the extended area but all after line 7 is blank.
I cannot believe this, as having read a bit more about all this I now know that the beginning of the extended area is where an HDMI device is required to declare that it is HDMI, to distinguish itself from DVI using the "HDMI Vendor-Specific Data Block":
http://www.hdmi.org/download/HDMI_Specification_1.1.pdf
It may be to do with a "splitter". My video is routed through my AV amp (Denon 4306). But it is clear from the above document that the repeater is required to pass through the EDID data so the end device can operate correctly. Maybe it is just all these tools do not read the data correctly when a repeater is in the loop, although for compatibility reasons I cannot see why.
For me Rivatuner also reports no extended data. It shows a table covering the extended area but all after line 7 is blank.
I cannot believe this, as having read a bit more about all this I now know that the beginning of the extended area is where an HDMI device is required to declare that it is HDMI, to distinguish itself from DVI using the "HDMI Vendor-Specific Data Block":
http://www.hdmi.org/download/HDMI_Specification_1.1.pdf
It may be to do with a "splitter". My video is routed through my AV amp (Denon 4306). But it is clear from the above document that the repeater is required to pass through the EDID data so the end device can operate correctly. Maybe it is just all these tools do not read the data correctly when a repeater is in the loop, although for compatibility reasons I cannot see why.
This is very strange. It must be the tools.
If the extended data block wasn't passed back with "Repeaters" in the chain then Intel based HTPC's would not send HDMI audio (Intel are very anal about EDID ;)) and (other than the issue of malformed SAD's in some Denon AVRs) we know they do. Regardless, the extended block is where the supported audio formats are stored and it would be ridiculous to think that AVR's can't report that information :rolleyes:
Don't suppose you're able to hook up the HTPC directly to the TV just for a quick test?
Wo0zy
HT Slider
06-03-08, 12:28 PM
Maybe I am missing something, but as per my earlier post this seems clear. Unless the dongle is "in action", switching the card to "HDMI mode", then the card outputs RGB @PC-levels. It seems clear from what is appearing in your CCC that, despite the dongle, the driver is recognising, correctly, that it is in fact connected to a DVI display, that it therefore cannot use YCbCr (which always uses video levels) and so it is treating it like a PC monitor.
Is there a way that this does not fit with the facts?
While that might be a logical explanation, the fact that when I remove the HDMI dongle and use DVI straight to DVI, the grey level problems become twice as bad, suggests otherwise. With the dongle and default settings I can see grey levels from 16-235 (bitmap displayed) and without the dongle and default settings I can see roughly grey levels from 30-220 (same bitmap).
Since the dongle does get rid of one level of expansion when compared to DVI direct, there must be some other explanation, other than the ATI card doesn't recognize a difference between HDMI dongle vs no HDMI dongle when an actual DVI port is used at least.
One explanation for example could be that my HDTV is simply out of calibration and can only display 30-220. From a brightness/contrast adjustment sense it definitely isn't out of calibration though (I can crank up the brightness and turn down the contrast until the display is totally washed out, yet it never shows any additional grey levels). I suppose it could somehow be out of calibration from strictly a digital perspective (able to only display 30-220), but this does not really make sense to me. How can the DVI port digitally be that far out? On top of that, why don't I see the same problem when an Nvidia card is used (it has many different problems, but desktop by default can visualize 16-235 grey levels without any adjustment to the Nvidia control panel when using DVI-DVI; as expected. Only ATI cards do this extra level of expansion on everything).
I think it is far more likely that the EDID is somehow confusing the ATI card into performing an extra level of expansion as a final step on output, but I don't know enough about the EDID to figure out if this is possible. Thus the reason I was asking the question.
HT Slider
06-03-08, 12:35 PM
Sorry HT, I missed that post.
I think Jong1 is correct. Your connection isn't end-to-end HDMI therefore the connection is treated as RGB/PC rather than YCbCr/video levels. I use my main system without the dongle and with my TV calibrated (as best I can) for RGB levels and using only the "UseBT601CSC=1" reg tweak, I'm fairly happy that all my levels (desktop, SD and HD) are consistent. At least they are acceptable to my eyes. All video looks good and photos are detailed enough for casual viewing (I don't use this system for editing anyway). Obviously, switching to a YCbCr input (such as my DVB-T tuner or Sky receiver) reveals a rubbish washed out, over bright image but this doesn't bother me as I use the HTPC for just about eveything anyway :)
Regarding your EDID report, as per other posts, it only seems to contain the first block (no extended data). Perhaps you could try RivaTuner as SugoE suggests and see if it's possible to capture the Version 1.3 extended block? It's likely that the 720p timing is in there as either an SVD or DTD.
Wo0zy
I didn't include the extended data because it was all 0's.
I do believe they really are 0's too because I've read the EDID using many different methods over the years with this particular HDTV and the extended EDID has always come up as 0's, even when reading from the HDTV directly.
jong1,
Try EDID Viewer
http://www.softpedia.com/get/System/...D-Viewer.shtml
It doesn't show SAD's (although it does show that my TV supports "Basic Audio") but is able to show EDID EXTENSION : EIA/CEA-861 video data including SVDs and DTD's.
It's interesting that this tool and Moninfo read the EDID from the registry. Perhaps clearing out the entries will force a redetection?
Wo0zy
I didn't include the extended data because it was all 0's.
I do believe they really are 0's too because I've read the EDID using many different methods over the years with this particular HDTV and the extended EDID has always come up as 0's, even when reading from the HDTV directly.
Yeah. that's more than possible. Toshiba used to be quite lazy with their EDIDs.
Judging from mine (a Tosh circ 2006) they finally sorted in out :).
Having said that, in Version 1.3 byte 126 specifies the number of extension blocks and yours suggests that there is one. As per my post to jong1, might be worth seeing what EDID Viewer comes up with just out of curiosity.
Wo0zy
While that might be a logical explanation, the fact that when I remove the HDMI dongle and use DVI straight to DVI, the grey level problems become twice as bad, suggests otherwise. With the dongle and default settings I can see grey levels from 16-235 (bitmap displayed) and without the dongle and default settings I can see roughly grey levels from 30-220 (same bitmap).
Since the dongle does get rid of one level of expansion when compared to DVI direct, there must be some other explanation, other than the ATI card doesn't recognize a difference between HDMI dongle vs no HDMI dongle when an actual DVI port is used at least.
One explanation for example could be that my HDTV is simply out of calibration and can only display 30-220. From a brightness/contrast adjustment sense it definitely isn't out of calibration though (I can crank up the brightness and turn down the contrast until the display is totally washed out, yet it never shows any additional grey levels). I suppose it could somehow be out of calibration from strictly a digital perspective (able to only display 30-220), but this does not really make sense to me. How can the DVI port digitally be that far out? On top of that, why don't I see the same problem when an Nvidia card is used (it has many different problems, but desktop by default can visualize 16-235 grey levels without any adjustment to the Nvidia control panel when using DVI-DVI; as expected. Only ATI cards do this extra level of expansion on everything).
I think it is far more likely that the EDID is somehow confusing the ATI card into performing an extra level of expansion as a final step on output, but I don't know enough about the EDID to figure out if this is possible. Thus the reason I was asking the question.Hmm. That does sound odd. When you say "bitmap" I assume you mean a photo - not a frozen video frame - with all 255 levels contained within it?
From what I have read over the last day I do not think the EDID should have any control over this. Not to say ATI could not have messed things up somehow though!
nicksan
06-03-08, 02:37 PM
Sorry to interrupt this highly technical conversation going on here...:)
Anyways, I did end up borrowing Ocean's 13 BR and it played back fine. Wasn't too impressed with the production of the movie, however that's another story! Again in the Information page I see it's using Dolby Digital 5.1, not MLP...so I guess it wasn't a great test. I need to get my hands on a title with TrueHD to see if it messes everything up again.
Also, once again, I was unable to get the main menu up at all. What's up with that?
Does PDVD figure out the audio streams on the BR DVD? What I mean by this is selecting the DD5.1 AC-3 track via the BR menu (which I can't reach at the moment...) the same as selecting the audio track in PDVD itself?
I am just concerned that b/c I can never seem to get at the menu for these BR titles to select the audio track, which for me will always be AC-3 DD5.1 or DTS since I have an older receiver, that if I encounter a title with TrueHD as the default track, then it will mess things up. That's what is happening with the Sony Sample BR disc. Stuttering like crazy...audio = MLP...and no way to choose DD5.1 b/c there isn't a section in the menu to select the audio track. Maybe it doesn't even have an AC-3 track?
Right now I can't remember if TrueHD uses MLP, but it could well be.
In addition to TrueHD there is often option to use another audio format, for example PCM or DD 5.1 (not downmixed) and those are lighter for processor. You can change audio streams using buttons in PDVD when there is no audio menu at hand.
HT Slider
06-03-08, 03:29 PM
Hmm. That does sound odd. When you say "bitmap" I assume you mean a photo - not a frozen video frame - with all 255 levels contained within it?
From what I have read over the last day I do not think the EDID should have any control over this. Not to say ATI could not have messed things up somehow though!
I'm using the calibration bitmap (bmp) that I posted about a page back with the links to various calibration sources. It contains a ramp of all 256 levels.
EDID does contain color information. I'm not certain what, but I wouldn't be surprised if it contains supported grey levels as well as RGB and/or YCbCr support.
Where did you read that EDID has no control over color?
Is there a separate DVI/HDMI handshaking protocol that negotiates color?
Something does....
Look honestly, I don't claim to be an expert on EDID at all. Just trying to help troubleshoot. But have a look at that HDMI document I posted a link to a little while back.
It says that YCbCr must be "limited range" (16-235) and RGB may be either. It is not negotiated, just stated. It also says that the colorspace to be used can be inferred by the resolution or explicitly stated in an optional AVI Infoframe broadcast along with a video stream, but even in that info frame there is no info about the levels that should be used.
Of course that is not the DVI spec and your TV is DVI, so things could be different there, but still I'd be surprised.
Look honestly, I don't claim to be an expert on EDID at all. Just trying to help troubleshoot. But have a look at that HDMI document I posted a link to a little while back.
It says that YCbCr must be "limited range" (16-235) and RGB may be either. It is not negotiated, just stated. It also says that the colorspace to be used can be inferred by the resolution or explicitly stated in an optional AVI Infoframe broadcast along with a video stream, but even in that info frame there is no info about the levels that should be used.
Of course that is not the DVI spec and your TV is DVI, so things could be different there, but still I'd be surprised.
You're one fast reader jong1 ;)
That's a truly mind-blowing document to digest!! Having said that, I have to agree. My sanitised understanding suggests that EDID is unlikely to be the source of HT's issue.
Firstly, he has no extension data available to comment on (despite the 1.1 data block claiming there is one). Secondly, the 1.1 data is almost identical in all important ways to mine and I don't experience the same problems. However, my display IS HDMI.
HT, page 115 onwards of the HDMI specification document is very interesting with regard to HDMI>DVI Compatibility. Well worth a read.
Wo0zy
bobbyrr
06-03-08, 07:06 PM
Anyone know why I might be getting unsmooth playback of a 1080p blue ray video on my hd2600xt?
Using powerdvd 8 v1662, vista 32bit sp1.
Possibly related, ticking and unticking use ati avivo hardware acceleration does not change the cpu usage of poweredvd, it stays around 38-45% on my core 2 duo 2.66ghz cpu.
catalyst 8.5s and latest AVIVO installed.
archibael
06-03-08, 08:56 PM
I wrote up a long message with footnotes and everything (it was impressive), then my crappy laptop crashed and I lost it. I'll summarize:
Per the HDMI 1.3a spec, an HDMI source is not allowed to send YCbCr to an HDMI sink at anything other than video levels (16/235), and can only send full range (0/255) on RGB if the resolution is 640x480.
HOWEVER, if a source doesn't see the second EDID block-- which I think it's clear from this thread that the ATI drivers do not-- it has to assume the sink is a DVI device, and it must send RGB. Unfortunately, I don't have a copy of the DVI 1.0 spec, so I can't tell what conditions (if any) require video levels. Presumably, the source would be sending full range by default.
The EDID does have some bits allocated for color space, chroma, and gamma info in the first block, but it is my experience those are ignored for the most part-- and in this case, I'm not seeing anything in there which would affect the situation.
HOWEVER, if a source doesn't see the second EDID block-- which I think it's clear from this thread that the ATI drivers do not
Incorrect. The ATI drivers do see the 2nd block, at least those drivers that I'm using (7.11 IIRC). I can adjust the EDID in my Lumagen Radiance video processor. So I can say that for sure.
@HT Slider
What are you using to display your .bmp file? Is there any chance this could be doing one level of expansion (16-235 -> 0-255)? With the driver at all default settings, have you tried taking a snapshot of your desktop with the .bmp displayed and using photoshop or other tool to check the levels?
I could imagine that if played back in some video player that it might erroneously do one level of expansion in the same way it would with video. This would be sufficient to get what you are observing above since your PC would then be losing the 0-15 data due to expansion and your TV would then be clipping the new 0-15 data (16-31 in the original) as you have explained it does.
Otherwise I would just repeat myself and suggest a complete clean uninstall and re-install of drivers and some new testing of HD & SD video and .bmp files at default, without dongle or registry tweaks using AVIVO only for video. I would using MPC-HC for video just to provide a solid easy to compare baseline.
archibael
06-04-08, 11:59 AM
Incorrect. The ATI drivers do see the 2nd block, at least those drivers that I'm using (7.11 IIRC). I can adjust the EDID in my Lumagen Radiance video processor. So I can say that for sure.
I stand corrected. I haven't been following the thread until it was pointed out to me by Wo0zy and I had gotten the impression nobody's EDID-read tools were working.
I stand corrected. I haven't been following the thread until it was pointed out to me by Wo0zy and I had gotten the impression nobody's EDID-read tools were working.
It seems some are and some aren't. HT's EDID report suggests there's an extension block but it shows as all zeros. jong1's definitely should have an extension block but that shows as all zeros as well. I can see my extension block using just about any tool.
Weird situation :confused:
You confirmed that it's unlikely to be EDID that's causing HT's expansion problems which ticks another possibility off the list.
Thanks for taking a look.
Wo0zy
It seems clear that even where the extended block is not visible it must have been read, as my driver is definitely in HDMI mode (says so in CCC). Technically I guess the dongle could be forcing HDMI mode regardless of whether the HDMI vendor specific block had been read from the sink, but then I am sure ATI would have problems with using the HDMI logo and stating HDMI compliance on their products.
It is a little confusing to say the least. I am going to take a look directly in the registry later and see if I can find anything.
HT Slider
06-04-08, 01:20 PM
You confirmed that it's unlikely to be EDID that's causing HT's expansion problems which ticks another possibility off the list.
I'm not sure anything archibael explained confirms that EDID isn't involved - at least I didn't read it that way.
Keep in mind:
The ATI card does display within CCC that my HDTV is hooked up using DVI, regardless of if the HDMI dongle is used or not (but I get double expansion if no dongle is used). The double expansion could be a bug in the driver perhaps and single expansion when the HDMI dongle is used, could be because the card ultimately does realize that DVI is the connection at the HDTV somehow. The question I have here is how does the ATI card know if the HDVT's final connection is HDMI or DVI? Is this in the EDID or somewhere else? Based on what I've read, my understanding is that the end connection being HDMI vs DVI shouldn't have any effect. The only thing the video card knows is that an HDMI cable is attached to it and a specific EDID report is received from the HDTV.
My HDTV does provides, within the EDID, a specific color space, chroma and gamma details. Most devices don't seem to contain this information. Moninfo reports the following from the EDID (as posted earlier):
Color characteristics
Display gamma............... 1.00
Red chromaticity............ Rx 0.640 - Ry 0.330
Green chromaticity.......... Gx 0.300 - Gy 0.600
Blue chromaticity........... Bx 0.150 - By 0.060
White point (default)....... Wx 0.313 - Wy 0.329
All I know for certain right now:
There appears to be an extra level of expansion going on somewhere.
The EDID contains some specific information regarding colorspace, chroma and gamma (that I don't know how to interpret).
For some reason, even though the HDMI dongle is used, CCC reports the connection as DVI (noting that the actual, final connection to the HDTV is indeed DVI).
When the HDMI dongle is NOT used, my HDTV experiences a double expansion (~30-220 visible).
To answer jong1's question earlier, I am using simply the desktop or Vista picture viewer to view the color ramp. Taking a screen capture of the entire desktop and viewing it shows exactly the same color ramp proving that the same 0-255 is retained in a screen capture. Again, by default, using the HDMI dongle I see 16-235 on my HDTV and without the dongle I see roughly 30-220. Note that a "typical" display seems to experience one less level of expansion (typically no HDMI dongle 16-235 visible; HDMI dongle used 0-255 visible)
When the HDMI dongle is used, I can get correct grey levels by cranking up the brightness in CCC to +31 and turning down the contrast to 73%. While this does work, I expect it reduces the number of grey levels available for image processing.
At this point I still don't know enough, but I do know that people have successfully re-written the EDID with my HDTV. There are also methods (at least there used to be) to overwrite the EDID within the registry and change the way the video driver interprets the display. I'd like to find a way to make my ATI card drive my HDTV in a "normal" fashion (an Nvidia card can, so why can't an ATI card).
I might be looking in totally the wrong direction, but I still suspect the specific EDID my HDTV provides might be involved.
archibael
06-04-08, 01:21 PM
Other possibility is that they are taking a shortcut and merely looking to see if the first block states the existence of the second (extension) block, and then just assuming from there that it's HDMI. The second-to-last byte in the first block will be a "01" if there is a second block, or a "00" if not.
HT Slider
06-04-08, 01:28 PM
Reading through the HDMI specification (thanks jong1 and woozy) I found the following:
8.3.3 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
It appears that the EDID is what tells the video card if an HDMI or DVI device is in use. My current guess is this most likely is where ATI video cards are getting confused and sending regular RGB, even when the HDMI dongle is used. The bizarre feature is how they do a double expansion when the HDMI dongle is not used.
archibael
06-04-08, 01:37 PM
I'm not sure anything archibael explained confirms that EDID isn't involved - at least I didn't read it that way.
The question I have here is how does the ATI card know if the HDVT's final connection is HDMI or DVI? Is this in the EDID or somewhere else?
EDID second block has an HDMI vendor specific block in it if the final connection is HDMI, otherwise it's DVI. That's how it's supposed to make the choice.
Based on what I've read, my understanding is that the end connection being HDMI vs DVI shouldn't have any effect. The only thing the video card knows is that an HDMI cable is attached to it and a specific EDID report is received from the HDTV.
It knows the dongle is attached, which may put the drivers into some interesting mode (actually, it definitely puts the drivers into a different mode, as the dongle activates audio), but it has no clue about the cabling. You can have a chain of various DVI and HDMI cables with converters and the card (nor the end device) will not be able to tell the difference.
My HDTV does provides, within the EDID, a specific color space, chroma and gamma details. Most devices don't seem to contain this information.
Really? It's been my experience that most reported EDIDs do have that information.
There appears to be an extra level of expansion going on somewhere.
The EDID contains some specific information regarding colorspace, chroma and gamma (that I don't know how to interpret).
For some reason, even though the HDMI dongle is used, CCC reports the connection as DVI (noting that the actual, final connection to the HDTV is indeed DVI).
When the HDMI dongle is NOT used, my HDTV experiences a double expansion (~30-220 visible).
Yeah, this sounds like a bug.
New Driver Guy: "Hey, what's this about "fix expansion for DVI color quantization?"
Old Driver Guy: "Oh, yeah, when you have DVI you need to modify the quantization by like 15 on both ends.
New Driver Guy: "Okay. Let's see... [typing] If DVI, then RGB(max) = RGB(max) - 15; if DVI, then RGB (min) = RGB (min) + 15. There, that ought to do it."
I might be looking in totally the wrong direction, but I still suspect the specific EDID my HDTV provides might be involved.
It is possible, I'm just not seeing anything which jumps out at me. It's also possible that there's something in the registry for that particular monitor which got hosed and it lingers beyond dongle removal. When you first hooked it up, was it with the dongle or not? Perhaps finding all the registry entries for that monitor, deleting, them, and then trying again without the dongle would fix things. [shrug] Sorry, I don't have too many useful suggestions on this one.
HT Slider
06-04-08, 01:39 PM
Other possibility is that they are taking a shortcut and merely looking to see if the first block states the existence of the second (extension) block, and then just assuming from there that it's HDMI. The second-to-last byte in the first block will be a "01" if there is a second block, or a "00" if not.
Here is the full EDID extracted with Moninfo:
Raw EDID base
00: 00 FF FF FF FF FF FF 00 52 62 00 01 62 91 02 00
10: 1F 0D 01 03 80 00 00 00 0E EE 91 A3 54 4C 99 26
20: 0F 50 54 20 00 00 01 01 01 01 01 01 01 01 01 01
30: 01 01 01 01 01 01 01 1D 80 18 71 1C 16 20 58 2C
40: 25 00 00 00 00 00 00 9E 8C 0A D0 8A 20 E0 2D 10
50: 10 3E 96 00 00 00 00 00 00 18 00 00 00 FC 00 54
60: 6F 73 68 69 62 61 20 54 56 0A 20 20 00 00 00 FD
70: 00 3B 3D 1F 2E 08 00 0A 20 20 20 20 20 20 01 E2
Raw EDID extension (reserved)
00: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
10: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
20: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
30: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
40: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
50: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
60: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
70: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
The second to last block is 01, but the extension is all zeros.
When the specification states "HDMI Source shall check the E-EDID for the
presence of an HDMI Vendor Specific Data Block within the first CEA EDID Timing Extension", what does that really mean?
My HDTV's E-EDID is all zero's, so I assume this means the EDID is telling the source device that is is talking to a DVI device (which also aligns with what CCC reports).
archibael
06-04-08, 01:45 PM
From the EDID Wiki:
A Vendor Specific Data Block (if any) contains as its first three bytes the vendor's IEEE 24-bit registration number, LSB first. It is usually followed by a three byte source physical address, LSB first. The source physical address provides the CEC physical address for upstream CEC devices. The remainder of the Vendor Specific Data Block is the "data payload",which can be anything the vendor considers worthy of inclusion in this EDID extension block.
Basically, in the second EDID block there is an optional sub-block of data called the VSDB. If included, it can have anything the "vendor" wants to put in there. The HDMI spec says that if you want to be recognized as a valid HDMI sink device, you must have some specific HDMI "vendor" data in the second block.
HT Slider
06-04-08, 02:00 PM
I missed your last post when I posted mine...
Really? It's been my experience that most reported EDIDs do have that information.
Perhaps I shouldn't have said don't contain it. My numbers just look different from what I've seen elsewhere. Do these numbers look typical in your (much greater) experience?
Yeah, this sounds like a bug.
New Driver Guy: "Hey, what's this about "fix expansion for DVI color quantization?"
Old Driver Guy: "Oh, yeah, when you have DVI you need to modify the quantization by like 15 on both ends.
New Driver Guy: "Okay. Let's see... [typing] If DVI, then RGB(max) = RGB(max) - 15; if DVI, then RGB (min) = RGB (min) + 15. There, that ought to do it."
I think you are bang on with this statement.
Most likely there is a bug in the ATI driver where it knows it is outputting to an HDTV and the HDTV reports a DVI connection (this would be relatively rare I would expect). Depending on which "output mode" the video card is in (by using or not using the ATI HDMI dongle), it either gets the grey levels partially confused (dongle in use) or totally confused (dongle not in use).
It is possible, I'm just not seeing anything which jumps out at me. It's also possible that there's something in the registry for that particular monitor which got hosed and it lingers beyond dongle removal. When you first hooked it up, was it with the dongle or not? Perhaps finding all the registry entries for that monitor, deleting, them, and then trying again without the dongle would fix things. [shrug] Sorry, I don't have too many useful suggestions on this one.
I guess my thinking wasn't that is EDID is actually wrong, but that the EDID is what is triggering the issue with ATI drivers. Since it is the EDID that tells the video card DVI is the final connection to an HDTV, it is most likely the EDID that triggers the bug.
If my guess is correct and there are very few individuals with HDTVs using actual DVI connections, this is likely one of those bugs that will never get fixed.
One work around for me would probably be to figure out how to reflash a new EDID to my HDTV. If I can make it appear to be using HDMI, this issue will likely go away.
Now the question becomes how easily can I reflash a new EDID (while still being able to go back to the original one later).
Maybe I should open a support request with ATI also about this bug, but it is a very complex issue and I'm still not 100% certain how it all works.
New thoughts...
We know that when the HDMI dongle is used, I end up with 16-235 visible (calibration ramp picture).
We know that when the HDMI dongle is not used, I end up with 30-220 visible (same ramp picture).
We know that my HDTV does report correctly that DVI is used (at the HDTV itself).
We also know that when an Nvidia card is used, without any HDMI dongle, that 16-235 is visible.
Essentially this means that as long as I use the ATI dongle and an HDMI to DVI cable with an ATI card that it behaves the same as an Nvidia card.
New questions:
Does this mean that there is a bug in both ATI and Nvidia drivers or is my HDTV somehow telling the video card through its EDID that it can actually display RGB (0-255) when it can only display YCbCr (16-235)?
Does the "sink" telling the "source" that DVI is used, automatically mean the sink is expected to output using RGB (0-255), or is there something else in the EDID that is supposed to tell the sink to send YcBcR (16-235)?
EDIT: I missed your very last post prior to submitting this again...
HT Slider
06-04-08, 02:08 PM
Basically, in the second EDID block there is an optional sub-block of data called the VSDB. If included, it can have anything the "vendor" wants to put in there. The HDMI spec says that if you want to be recognized as a valid HDMI sink device, you must have some specific HDMI "vendor" data in the second block.
This part makes sense. My HDTV is correctly using the EDID to tell the source that it has a DVI connection. It does have DVI connection, even though it can only display YCbCr (16-235). It cannot handle RGB (0-255).
Does the fact that it reports as DVI automatically tell the source to send it RGB or is there something else in the EDID that defines if it wants YCbCr or RGB?
Perhaps it is whatever within the EDID that tells the video card to send YCbCr that the video card is ignoring (or the EDID is missing)?
The double expansion issue does indeed seem to be a bug in the ATI driver, but as long as an HDMI dongle is used it seems to go away.
I looked in the registry and found the EDID for my monitor. The full extended EDID is there. I have no idea why all the tools I have tried seem not to display it.
HT I suggest you have a look in HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Enum\DISPLAY
There will probably be a lot there but look for the one with the right "DeviceDesc". Mine was @ HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Enum\DISPLAY\DON 0004\5&96259de&3&11335588&06&00
Export the "Device Parameters" key as text. Then open the text file and tidy it up (a bit fiddly).
In case anyone is interested mine in the end was:
00 ff ff ff ff ff ff 00 11 ee 04 00 01 01 01 01
00 00 01 03 80 46 28 78 0a ee 91 a3 54 4c 99 26
0f 50 54 00 00 00 01 01 01 01 01 01 01 01 01 01
01 01 01 01 01 01 02 3a 80 d0 72 38 2d 40 10 2c
45 80 df a4 21 00 00 1e 8c 0a d0 8a 20 e0 2d 10
10 3e 96 00 df a4 21 00 00 18 00 00 00 fc 00 44
45 4e 4f 4e 2d 41 56 41 4d 50 0a 20 00 00 00 fd
00 30 3e 0e 46 0f 00 0a 20 20 20 20 20 20 01 3f
02 03 31 76 51 14 03 04 12 13 05 01 07 16 9f 10
15 11 06 0f 1d 1e 23 0f 1f 07 23 09 7f 07 23 3d
1f c0 23 15 1f 51 83 5f 00 00 66 03 0c 00 21 00
80 01 1d 00 72 51 d0 1e 20 6e 28 55 00 df a4 21
00 00 1e 8c 0a d0 90 20 40 31 20 0c 40 55 00 df
a4 21 00 00 18 01 1d 00 bc 52 d0 1e 20 b8 28 55
40 df a4 21 00 00 1e 01 1d 80 18 71 1c 16 20 58
2c 25 00 df a4 21 00 00 9e 00 00 00 00 00 00 1d
Interpreted by DTD Calculator as:
No Established Timings Found
Block 1 DTD Found (See List)
Block 1 DTD Found (See List)
Display does not support underscan
Display supports basic audio
Display supports YCbCr 4:4:4
Display supports YCbCr 4:2:2
6 native formats in DTDs
Short Video Descriptor:
Resolution Supported - 1920x1080i @ 50Hz 16:9 Non-Native
Short Video Descriptor:
Resolution Supported - 720x480p @ 59.94/60Hz 16:9 Non-Native
Short Video Descriptor:
Resolution Supported - 1280x720p @ 59.94/60Hz 16:9 Non-Native
Short Video Descriptor:
Resolution Supported - 720x576p @ 50Hz 16:9 Non-Native
Short Video Descriptor:
Resolution Supported - 1280x720p @ 50Hz 16:9 Non-Native
Short Video Descriptor:
Resolution Supported - 1920x1080i @ 59.94/60Hz 16:9 Non-Native
Short Video Descriptor:
Resolution Supported - 640x480p @ 59.94/60Hz 4:3 Non-Native
Short Video Descriptor:
Resolution Supported - 720x480i @ 59.94/60Hz 16:9 Non-Native
Short Video Descriptor:
Resolution Supported - 720x576i @ 50Hz 16:9 Non-Native
Short Video Descriptor:
Resolution Supported - 1920x1080p @ 50Hz 16:9 Native
Short Video Descriptor:
Resolution Supported - 1920x1080p @ 59.94/60Hz 16:9 Non-Native
Short Video Descriptor:
Resolution Supported - 720x576i @ 50Hz 4:3 Non-Native
Short Video Descriptor:
Resolution Supported - 720x576p @ 50Hz 4:3 Non-Native
Short Video Descriptor:
Resolution Supported - 720x480i @ 59.94/60Hz 4:3 Non-Native
Short Video Descriptor:
Resolution Supported - 1440x480p @ 59.94/60Hz 16:9 Non-Native
Short Video Descriptor:
Resolution Supported - 1440x576p @ 50Hz 4:3 Non-Native
Short Video Descriptor:
Resolution Supported - 1440x576 @ 50Hz 16:9 Non-Native
Short Audio Descriptor:
Audio Format - LPCM
8 Channel Sound
Sampling Frequencies Supported :96kHz, 88kHz, 48kHz, 44kHz, 32kHz
Bit Depth :24 bit, 20 bit, 16 bit
Short Audio Descriptor:
Audio Format - LPCM
2 Channel Sound
Sampling Frequencies Supported :192kHz, 176kHz, 96kHz, 88kHz, 48kHz, 44kHz, 32kHz
Bit Depth :24 bit, 20 bit, 16 bit
Short Audio Descriptor:
Audio Format - DTS
6 Channel Sound
Sampling Frequencies Supported :96kHz, 88kHz, 48kHz, 44kHz, 32kHz
Maximum supported bitrate :1536
Short Audio Descriptor:
Audio Format - AC-3
6 Channel Sound
Sampling Frequencies Supported :96kHz, 88kHz, 48kHz, 44kHz, 32kHz
Maximum supported bitrate :648
Speaker Allocation Block :
Speakers Present :
Rear Left Centre / Rear Right Centre
Block 2 DTD Found (See List)
Block 2 DTD Found (See List)
Block 2 DTD Found (See List)
Block 2 DTD Found (See List)
It seems to have both the TV info and additional, mainly audio, info added by my AVR.
Jon
This part makes sense. My HDTV is correctly using the EDID to tell the source that it has a DVI connection. It does have DVI connection, even though it can only display YCbCr (16-235). It cannot handle RGB (0-255).
Does the fact that it reports as DVI automatically tell the source to send it RGB or is there something else in the EDID that defines if it wants YCbCr or RGB?
Perhaps it is whatever within the EDID that tells the video card to send YCbCr that the video card is ignoring (or the EDID is missing)?
The double expansion issue does indeed seem to be a bug in the ATI driver, but as long as an HDMI dongle is used it seems to go away.The HDMI spec seems clear - if DVI send RGB .
When communicating with a DVI sink device, an HDMI Source shall operate in a mode
compatible with that device. This requires that the Source operate under the following limitations:
• Video pixel encoding shall be RGB.
• No Video Guard Bands shall be used.
• No Data Islands shall be transmitted.
An HDMI Source may transmit Video Data Periods without Guard Bands only when
communicating to a DVI sink device or during the process of determining if the sink device is
HDMI capable.
An HDMI Source, upon power-up, reset or detection of a new sink device, shall assume that the
sink device operates under DVI 1.0 limitations. An HDMI Source shall determine if the sink device
is an HDMI Sink by following the rule(s) described in Section 8.3.3. Upon detection of an HDMI
Sink, the HDMI Source shall follow all of the HDMI Source-related requirements specified in this
document.
All electrical and physical specifications in Section 4 shall be followed by the HDMI Source even
when communicating with a DVI sink device
If your TV is really a DVI device that only accepts YCbCr this could certainly be leading to confusion!
I think though you are probably saying your TV clips to video levels (16-235) rather than that it must be YCbCr encoded. Are you?
I looked in the registry and found the EDID for my monitor. The full extended EDID is there. I have no idea why all the tools I have tried seem not to display it.
HT I suggest you have a look in HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Enum\DISPLAY
There will probably be a lot there but look for the one with the right "DeviceDesc". Mine was @ HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Enum\DISPLAY\DON 0004\5&96259de&3&11335588&06&00
Export the "Device Parameters" key as text. Then open the text file and tidy it up (a bit fiddly).
In case anyone is interested mine in the end was:
00 ff ff ff ff ff ff 00 11 ee 04 00 01 01 01 01
00 00 01 03 80 46 28 78 0a ee 91 a3 54 4c 99 26
0f 50 54 00 00 00 01 01 01 01 01 01 01 01 01 01
01 01 01 01 01 01 02 3a 80 d0 72 38 2d 40 10 2c
45 80 df a4 21 00 00 1e 8c 0a d0 8a 20 e0 2d 10
10 3e 96 00 df a4 21 00 00 18 00 00 00 fc 00 44
45 4e 4f 4e 2d 41 56 41 4d 50 0a 20 00 00 00 fd
00 30 3e 0e 46 0f 00 0a 20 20 20 20 20 20 01 3f
02 03 31 76 51 14 03 04 12 13 05 01 07 16 9f 10
15 11 06 0f 1d 1e 23 0f 1f 07 23 09 7f 07 23 3d
1f c0 23 15 1f 51 83 5f 00 00 66 03 0c 00 21 00
80 01 1d 00 72 51 d0 1e 20 6e 28 55 00 df a4 21
00 00 1e 8c 0a d0 90 20 40 31 20 0c 40 55 00 df
a4 21 00 00 18 01 1d 00 bc 52 d0 1e 20 b8 28 55
40 df a4 21 00 00 1e 01 1d 80 18 71 1c 16 20 58
2c 25 00 df a4 21 00 00 9e 00 00 00 00 00 00 1d
Interpreted by DTD Calculator as:
No Established Timings Found
Block 1 DTD Found (See List)
Block 1 DTD Found (See List)
Display does not support underscan
Display supports basic audio
Display supports YCbCr 4:4:4
Display supports YCbCr 4:2:2
6 native formats in DTDs
Short Video Descriptor:
Resolution Supported - 1920x1080i @ 50Hz 16:9 Non-Native
Short Video Descriptor:
Resolution Supported - 720x480p @ 59.94/60Hz 16:9 Non-Native
Short Video Descriptor:
Resolution Supported - 1280x720p @ 59.94/60Hz 16:9 Non-Native
Short Video Descriptor:
Resolution Supported - 720x576p @ 50Hz 16:9 Non-Native
Short Video Descriptor:
Resolution Supported - 1280x720p @ 50Hz 16:9 Non-Native
Short Video Descriptor:
Resolution Supported - 1920x1080i @ 59.94/60Hz 16:9 Non-Native
Short Video Descriptor:
Resolution Supported - 640x480p @ 59.94/60Hz 4:3 Non-Native
Short Video Descriptor:
Resolution Supported - 720x480i @ 59.94/60Hz 16:9 Non-Native
Short Video Descriptor:
Resolution Supported - 720x576i @ 50Hz 16:9 Non-Native
Short Video Descriptor:
Resolution Supported - 1920x1080p @ 50Hz 16:9 Native
Short Video Descriptor:
Resolution Supported - 1920x1080p @ 59.94/60Hz 16:9 Non-Native
Short Video Descriptor:
Resolution Supported - 720x576i @ 50Hz 4:3 Non-Native
Short Video Descriptor:
Resolution Supported - 720x576p @ 50Hz 4:3 Non-Native
Short Video Descriptor:
Resolution Supported - 720x480i @ 59.94/60Hz 4:3 Non-Native
Short Video Descriptor:
Resolution Supported - 1440x480p @ 59.94/60Hz 16:9 Non-Native
Short Video Descriptor:
Resolution Supported - 1440x576p @ 50Hz 4:3 Non-Native
Short Video Descriptor:
Resolution Supported - 1440x576 @ 50Hz 16:9 Non-Native
Short Audio Descriptor:
Audio Format - LPCM
8 Channel Sound
Sampling Frequencies Supported :96kHz, 88kHz, 48kHz, 44kHz, 32kHz
Bit Depth :24 bit, 20 bit, 16 bit
Short Audio Descriptor:
Audio Format - LPCM
2 Channel Sound
Sampling Frequencies Supported :192kHz, 176kHz, 96kHz, 88kHz, 48kHz, 44kHz, 32kHz
Bit Depth :24 bit, 20 bit, 16 bit
Short Audio Descriptor:
Audio Format - DTS
6 Channel Sound
Sampling Frequencies Supported :96kHz, 88kHz, 48kHz, 44kHz, 32kHz
Maximum supported bitrate :1536
Short Audio Descriptor:
Audio Format - AC-3
6 Channel Sound
Sampling Frequencies Supported :96kHz, 88kHz, 48kHz, 44kHz, 32kHz
Maximum supported bitrate :648
Speaker Allocation Block :
Speakers Present :
Rear Left Centre / Rear Right Centre
Block 2 DTD Found (See List)
Block 2 DTD Found (See List)
Block 2 DTD Found (See List)
Block 2 DTD Found (See List)
It seems to have both the TV info and additional, mainly audio, info added by my AVR.
Jon
Sweet :)
DTDCalc doesn't interpret ALL supported audio formats yet ( we're still working on the "codes" for the compressed HD formats) but it's still useful information.
Double-clicking on the DTD's extracted to the RH box will give you all the timing data. It's designed to convert to "modeline" values but you should still be able to work out the important information from the "calculation" tab.
It would be very interesting to see HT's entry for his Tosh display.
@Archibael,
Thanks again for taking the time to get involved.
Wo0zy
archibael
06-04-08, 02:40 PM
This part makes sense. My HDTV is correctly using the EDID to tell the source that it has a DVI connection. It does have DVI connection, even though it can only display YCbCr (16-235). It cannot handle RGB (0-255).
Does the fact that it reports as DVI automatically tell the source to send it RGB or is there something else in the EDID that defines if it wants YCbCr or RGB?
Perhaps it is whatever within the EDID that tells the video card to send YCbCr that the video card is ignoring (or the EDID is missing)?
The double expansion issue does indeed seem to be a bug in the ATI driver, but as long as an HDMI dongle is used it seems to go away.
I don't think there's anything in the EDID which tells the source to send RGB vs YCbCr other than the fact of whether it's DVI or not. If DVI, send RGB; if HDMI send either (at the discretion of the source), but must be 16-255. Though if this is something with your TV and no one else is encountering it, obviously it starts to point more at the EDID again.
As for colorspace parameters in the EDID, I'll give you some random examples of what I've seen:
Monitor
Windows description......... HannStar Monitor
Manufacturer description.... HW191A
Color characteristics
Display gamma............... 2.20
Red chromaticity............ Rx 0.643 - Ry 0.325
Green chromaticity.......... Gx 0.295 - Gy 0.616
Blue chromaticity........... Bx 0.143 - By 0.081
White point (default)....... Wx 0.310 - Wy 0.330
Monitor
Windows description......... Generic PnP Monitor
Manufacturer description.... 202M
Color characteristics
Display gamma............... 2.20
Red chromaticity............ Rx 0.640 - Ry 0.340
Green chromaticity.......... Gx 0.295 - Gy 0.610
Blue chromaticity........... Bx 0.145 - By 0.070
White point (default)....... Wx 0.313 - Wy 0.329
Monitor
Windows description......... ViewSonic Monitor
Manufacturer description.... VX2235wm-3
Color characteristics
Display gamma............... 2.20
Red chromaticity............ Rx 0.646 - Ry 0.339
Green chromaticity.......... Gx 0.290 - Gy 0.603
Blue chromaticity........... Bx 0.145 - By 0.065
White point (default)....... Wx 0.313 - Wy 0.329
Monitor
Windows description......... ViewSonic Monitor
Manufacturer description.... VX2235wm
Color characteristics
Display gamma............... 2.20
Red chromaticity............ Rx 0.640 - Ry 0.333
Green chromaticity.......... Gx 0.285 - Gy 0.602
Blue chromaticity........... Bx 0.152 - By 0.074
White point (default)....... Wx 0.313 - Wy 0.329
Monitor
Manufacturer description.... SAMSUNG
Manufacturer................ Samsung
Color characteristics
Display gamma............... 2,40
Red chromaticity............ Rx 0,632 - Ry 0,357
Green chromaticity.......... Gx 0,289 - Gy 0,596
Blue chromaticity........... Bx 0,143 - By 0,085
White point (default)....... Wx 0,280 - Wy 0,290
Monitor
Windows description......... Sony Monitor
Manufacturer description.... SONY TV
Color characteristics
Display gamma............... 2.20
Red chromaticity............ Rx 0.637 - Ry 0.321
Green chromaticity.......... Gx 0.289 - Gy 0.601
Blue chromaticity........... Bx 0.146 - By 0.081
White point (default)....... Wx 0.285 - Wy 0.294
Monitor Windows description......... Plug and Play Monitor
Manufacturer description.... DELL 2007WFP
Color characteristics
Display gamma............... 2.20
Red chromaticity............ Rx 0.640 - Ry 0.330
Green chromaticity.......... Gx 0.300 - Gy 0.600
Blue chromaticity........... Bx 0.150 - By 0.060
White point (default)....... Wx 0.313 - Wy 0.329
Monitor
Windows description......... SyncMaster 225MW(Digital)
Manufacturer description.... SyncMaster
Color characteristics
Display gamma............... 2,20
Red chromaticity............ Rx 0,639 - Ry 0,333
Green chromaticity.......... Gx 0,289 - Gy 0,597
Blue chromaticity........... Bx 0,153 - By 0,082
White point (default)....... Wx 0,313 - Wy 0,329
Monitor
Windows description......... Plug und Play-Monitor
Manufacturer description.... SANYO UK4
Color characteristics
Display gamma............... 2,30
Red chromaticity............ Rx 0,640 - Ry 0,330
Green chromaticity.......... Gx 0,290 - Gy 0,600
Blue chromaticity........... Bx 0,150 - By 0,060
White point (default)....... Wx 0,280 - Wy 0,290
archibael
06-04-08, 02:42 PM
Sweet :)
DTDCalc doesn't interpret ALL supported audio formats yet ( we're still working on the "codes" for the compressed HD formats) but it's still useful information.
I figured out what each is and updated the Wiki, but as for the details... without the 861D spec, I have no idea whether the new HD formats follow the same conventions as the older ones.
@Archibael,
Thanks again for taking the time to get involved.
Wo0zy
No problem. Sorry if I'm not much help.
Sweet :)
DTDCalc doesn't interpret ALL supported audio formats yet ( we're still working on the "codes" for the compressed HD formats) but it's still useful information.
Double-clicking on the DTD's extracted to the RH box will give you all the timing data. It's designed to convert to "modeline" values but you should still be able to work out the important information from the "calculation" tab.That's OK. My amp support 8-channel LPCM, but not HD bitstreams.
I did look at the DTDs. they still confuse me a bit.
- There are not DTDs for all the short descriptions by any means.
- There is a DTD for 1080i which show a very odd estimated refresh rate of something like 60.04Hz (not sure, but over 60hz anyway).
- There is no 1080p/60 DTD, but it certainly accepts it and it works perfectly.
- I don't understand the whole "native"/"non-native" thing. I thought it was just about the resolution of the panel but not according to the short descriptions.
Any thoughts?
I can only think the reason why none of the tools manage to extract the extended data is because they are getting confused by all the additional AVR data in there and dropping everything as a consequence.
archibael
06-04-08, 03:23 PM
That's OK. My amp support 8-channel LPCM, but not HD bitstreams.
I did look at the DTDs. they still confuse me a bit.
- There are not DTDs for all the short descriptions by any means.
That's correct. It is assumed that a mode with a SVD (Short Video Descriptor) uses the generic timings for that mode. DTDs would be redundant, as the source should know how to transmit general 720x480p, for instance.
- There is a DTD for 1080i which show a very odd estimated refresh rate of something like 60.04Hz (not sure, but over 60hz anyway).
Yep. The DTD notation, being integer based (except for the pixel clock) will often result in refresh rates which are slightly off. Generally speaking, the display should be able to cope with this (they're required by the EIA/CEA spec to allow for a certain percentage of pixel clock flex), but there can sometimes be problems.
- There is no 1080p/60 DTD, but it certainly accepts it and it works perfectly.
Yep. It's supported in SVD, so the source should know how to send it.
- I don't understand the whole "native"/"non-native" thing. I thought it was just about the resolution of the panel but not according to the short descriptions.
"Native" is supposed to be the resolution your TV/monitor is best able to display. Whether it is the actual resolution of the TV or whether it is just what the TV manufacturers have decided is the best scaling algorithm on their input chips, it's what a source device should nominally choose for you by default.
HT Slider
06-04-08, 03:45 PM
So in PowerDVD I checked the "Information" tab and compare between this and the Casino Royale BR DVD that works well and they look pretty much the same EXCEPT...the audio encoding.
Casino Royale = Dolby Digital
Sony BD-ROM Sample BR DVD = MLP
I know MLP is more "advanced" in that it is lossless but could this actually affect the HA and and cause stuttering? In PowerDVD since I send the audio to spdif, and because the receiver doesn't do MLP, I choose to mix it down to either DD or DTS. This has no effect. It stutters like crazy!
I borrowed a BR DVD from my co-worker, Oceans 13, and will see how that plays, however I am concerned about all this. The last thing I want to do is to upgrade my machine to a Core2Duo...my wife would kill me!:eek:
So gathering from what I have described, is there anything I am doing wrong? Again, Casino Royale BR DVD plays fine. So it's not a complete wash. There is something on that Sony BD-Rom sample BR DVD that is causing the CPU to spike up and cause stutter and the only difference I can tell is, again, the MLP audio vs. regular DD 5.1 on Casino Royale.
Any help would be greatly appreciated.
I don't think you are doing anything wrong, other than possibly expecting AGP cards to truly work as advertised and drop the CPU utilization to ultra low levels, especially when playing modern, high end, 7.1 channel audio tracks, especially if you want to transcode them into DD5.1 or DTS (using the CPU).
The reality is AGP cards are notorious for producing either no hardware acceleration (not your situation) or in general higher CPU utilization than PCIe cards (of course PCIe cards tend to also live in PCs with more CPU power).
The other (more significant in your case) reality is PowerDVD Ultra uses a lot of CPU power to decode higher end audio tracks.
More than likely you'll need to rely on selecting slightly lower end audio tracks, ideally DTS or DD5.1 and configure PowerDVD to send them directly to your audio hardware for decoding there. Although I don't know the actual number, I would guess that at least 95% of all Blu-ray movies available today contain either DTS or DD5.1 sound tracks. PowerDVD Ultra tends to automatically select high level