Home Theatre PC Motherboard Shootout

Very useful review imo, great to see bit tech covering all aspects of computers.

It is also good too see that companies are now creating a wide range of components that are competely focussed towards the media centre market.

Originally Posted by Andy Mc
Just wondering why no VIA solution was tested alongside these?

Its already been done. The VIA board was rubbish

http://www.bit-tech.net/hardware/2007/07/27/epia_ex_vs_amd_690g/1

Nice review, a good read :)

Originally Posted by Bindi
Yes, sorry, copy and paste fingers.

Don't mean to be a pain but you've changed the chart to read 2048x1536 but ISTR Tim hooking the 690G up to a Dell 30incher and ran it at it's native res namely 2560x1600.

Originally Posted by Bindi
The HDMI can only output at 1.2 standard though = 1080p (according to AMD).

ISTR there being some 24" displays with HDMI ports. Do they only accept 1080p over HDMI and not 1920x1200 which I assume is their native resolution. Essentially would the 690G (and the other chipsets given that they are all 1080p in the chart on page 4 for HDMI) not be able to do 1-1 pixel mapping with those displays over HDMI?

Originally Posted by Bindi
In actual fact it's MORE a test of the ADC than DAC since the ADCs are often of lower quality than DACs. However, this is the only comparative I (and a lot of other sites) have, but it does provide some general area of quality information, even if it's not exactly correct. I mean, it doesn't test every output either - only stereo, and I couldn't test HDCP LPCM output because we don't have a receiver

I agree it's the best you've got but just thought I'd remind people to keep it in mind. Perhaps some subjective listening tests would be appropriate to get a sense of real world differences?

Originally Posted by Bindi
Good point - it was enabled in PowerDVD because I simply ran out of time to do every iteration (was here til 12am two nights in a row), but next time when I do the AMD RS780 I will do.

It's obviously most important to have acceleration for the most demanding situations which is currently optical media but I see no reason why they can't accelerate any individual files you have in those codecs even if it's not quite as necessary. I eagerly await the review of the RS780 as it may just be close to the holy grail.

Originally Posted by Bindi
Well the output from the DVD was 480i and the screen was set to 1080p - the entire thing was scaled which makes seeing "errors" like jaggies far more apparent. This method (according to HQV) tests the processing quality of the system, not the screen.

I don't quite follow. Do you mean you sent 1080p from the IGP? If so then indeed all the processing (Deinterlacing & scaling) is being done by the system and the screen is merely outputting what it receives. What connection and screen were you using?

Originally Posted by Bindi
Cheers. I'll take the points on board for the next one. If there's any other tests you think are relevant, let me know (without requiring a sound room and £10k worth of kit )

Can't think of any more tests than the ones I've already mentioned but if I think of any I'll be sure to let you know :D

Looking forward to seeing reviews of the upcoming boards.
Looking to update the bird's HTPC. :)

It just came in my hands the GA-MA78GM-S2H which is some 20$ more than the one tested here (of course, is for AMD, but so is the MSI board). Why didn't you test this one? It has 4 DIMM slots, 5 SATA, 12 USB 2.0 and the rest almost the same with the one presented. AFAIK, the HD3200 onboard is the best onboard graphics so far. Or at least one of the best...

We have the board in-house, but we're waiting for drivers to catch up (they're very early at the moment) and I'm pretty sure that it is not for sale yet because AMD hasn't officially announced the 780G chipset yet.

Thanks for your fast answer.I think it will be an interesting solution for a HTPC. Maybe you can update this test in the future with the new board, just to see the differences? I am not sure I can reproduce your testing platform...

Originally Posted by dfrangu
Thanks for your fast answer.I think it will be an interesting solution for a HTPC. Maybe you can update this test in the future with the new board, just to see the differences? I am not sure I can reproduce your testing platform...

We'll be reviewing the board in due course and I would imagine that we will use pretty much the same testing methodology as was used here. ;)

Originally Posted by Renoir
Don't mean to be a pain but you've changed the chart to read 2048x1536 but ISTR Tim hooking the 690G up to a Dell 30incher and ran it at it's native res namely 2560x1600.

ISTR there being some 24" displays with HDMI ports. Do they only accept 1080p over HDMI and not 1920x1200 which I assume is their native resolution. Essentially would the 690G (and the other chipsets given that they are all 1080p in the chart on page 4 for HDMI) not be able to do 1-1 pixel mapping with those displays over HDMI?

http://www.amd.com/us-en/Processors/ProductInformation/0,,30_118_14603_14764,00.html

AMD claims:

2D Acceleration Features
# Highly optimized 128-bit engine, capable or processing multiple pixels per clock
# Supports a maximum resolution of 2048x1536 @ 32bpp
# Game acceleration including support for Microsoft's DirectDraw

But no confirmation of specific HDMI maximum output. We have previously used 2560x1600 on the Dell 30" a long, long time ago (I found YOUR post on Anandtech forums linking to OUR content telling me this :D:D)

Intel G35 supports 1920x1200 via HDMI but I can't find anything specific about 690G or GF7100 with respect to HDMI - whether it's less if it's passing through audio as well or simply acting as a video connection? On the 690G front - I'd expect at least 1920x1200 because the same internal ramdac is used for DVI as is HDMI so the limitation is absolute HDMI bandwidth. The problem is, most people only rate HDMI to TV standards: 1080p. I've just talked to TR who do more monitors and they've never come across at 24" with HDMI never doing native res, but they weren't specifically testing with onboard video. I'd expect the output design to be pretty much the same across discrete and onboard to save the cost of having to design another output design by AMD or Nvidia.

I changed the graph - I expect them to do the full size for a monitor, even if TVs are smaller.

I agree it's the best you've got but just thought I'd remind people to keep it in mind. Perhaps some subjective listening tests would be appropriate to get a sense of real world differences?

Subjective differences? That's like trying to describe what colour is blue, how well a game plays or what sound a your favourite band makes. At the very least, I'd have to be trained as an audio engineer to know the differences, otherwise I wouldn't be that confident. I have done it previously with soundcard reviews but was never that pleased with the outcome but the difference between some onboard sound and discrete cards is easily night and day. For the time it takes to test, it isn't worth it - this isn't audiophile review ;)

It's obviously most important to have acceleration for the most demanding situations which is currently optical media but I see no reason why they can't accelerate any individual files you have in those codecs even if it's not quite as necessary. I eagerly await the review of the RS780 as it may just be close to the holy grail.

Well remember that VC1 or h264 have to conform to a set standard, so accelerating them is easy. Media from file can be x264/aac/mkv wrapper or anything in better at any resolution and any bitrate. I don't have confirmation, but I'd expect that it's quite difficult to efficiently accelerate all these variables. Also, you have to remember that these companies DON'T want to be seen condoning "piracy" - they could easily lose their HDCP license if they don't jump through every-single-hoop thrown at them.

I don't quite follow. Do you mean you sent 1080p from the IGP? If so then indeed all the processing (Deinterlacing & scaling) is being done by the system and the screen is merely outputting what it receives. What connection and screen were you using?

Yep, the IGP does all the upscaling and processing (with some degree of CPU). It was HDMI from mobo to monitor. The screen is indeed merely outputting what it receives therefore this is removed from the processing equation.



The AMD RS780 review will arrive when the drivers are ready - it makes no sense to tell you what to expect if there's massive driver (and BIOS) improvements between now and the date of release, it would be unrepresentative. Also, we want to still remain friends with AMD ;) - only then can we know and report the full details of the chipset. We have already shown you what it looks like though: http://www.bit-tech.net/news/2008/02/01/amd_s_new_rs780_igp_and_sb700/1

We have previously used 2560x1600 on the Dell 30" a long, long time ago (I found YOUR post on Anandtech forums linking to OUR content telling me this)

The discrepancy at that time which I highlighted was the difference in your testing setups namely their use of vista as opposed to your use of xp which was later resolved with a vista driver update. Over the years I've noticed so many discrepancies between what a manufacturer lists on their website and the reality of their product that I can't help but follow up and try to resolve them. Just my nature I guess :D

Intel G35 supports 1920x1200 via HDMI but I can't find anything specific about 690G or GF7100 with respect to HDMI - whether it's less if it's passing through audio as well or simply acting as a video connection? On the 690G front - I'd expect at least 1920x1200 because the same internal ramdac is used for DVI as is HDMI so the limitation is absolute HDMI bandwidth. The problem is, most people only rate HDMI to TV standards: 1080p. I've just talked to TR who do more monitors and they've never come across at 24" with HDMI never doing native res, but they weren't specifically testing with onboard video. I'd expect the output design to be pretty much the same across discrete and onboard to save the cost of having to design another output design by AMD or Nvidia.

I'm very much inclined to agree in that it's merely a case of companies stating max res in TV standards rather than ACTUAL max res due to the current immaturity of HDMI on the PC platform as opposed to the CE market. Annoying for those of us technically minded but there you go.

I have done it previously with soundcard reviews but was never that pleased with the outcome but the difference between some onboard sound and discrete cards is easily night and day. For the time it takes to test, it isn't worth it - this isn't audiophile review

Your having done it previously is why I suggested it but I fully appreciate your concerns and upon reflection I feel you're probably right to leave it out.

Well remember that VC1 or h264 have to conform to a set standard, so accelerating them is easy. Media from file can be x264/aac/mkv wrapper or anything in better at any resolution and any bitrate. I don't have confirmation, but I'd expect that it's quite difficult to efficiently accelerate all these variables.

I'm certainly not a coding expert so you could very well be right and I guess there's no way for us to know without a confirmation from them.

Also, you have to remember that these companies DON'T want to be seen condoning "piracy" - they could easily lose their HDCP license if they don't jump through every-single-hoop thrown at them.

Interesting, I hadn't thought about that angle but it would be a crying shame if that was indeed the case. Damn the MPAA!!!

Yep, the IGP does all the upscaling and processing (with some degree of CPU). It was HDMI from mobo to monitor. The screen is indeed merely outputting what it receives therefore this is removed from the processing equation.

Cool. I agree that the upscaling would magnify the "jaggies" as you say but it's still not a true test of the quality of the upscaling itself. Having said that I'm not actually aware of a good test that is but thought I'd mention it given the amount of time and effort given to improving upscaling of video by the HTPC community e.g. FFDShow etc.

The AMD RS780 review will arrive when the drivers are ready - it makes no sense to tell you what to expect if there's massive driver (and BIOS) improvements between now and the date of release, it would be unrepresentative. Also, we want to still remain friends with AMD - only then can we know and report the full details of the chipset. We have already shown you what it looks like though: http://www.bit-tech.net/news/2008/02...gp_and_sb700/1

Agreed and I certainly don't mind waiting for a proper review.

I like to complain A LOT here.
The MSI Moard is great, we got one here. It plays Blu-Ray. But it can't do Extended Monitor onto TV with the onboard X1250.......

The Asus P5E-V board, can't play Blu-Ray on Windows XP with PowerDVD. So it's complete trash. Unless you decide to play the files ripped of the Blu-Rays alone. It requires the Windows Vista Driver (AACS/HDCP). And which sane person wants to install that ?
Also. It crashed a lot playing DVD's. Intel might have fixed some of that with latest driver. But it still happens the machine freeze up during play. And one can only Cold Reboot it.

We need some newer boards. :)

Welcome!

With PowerDVD? PowerDVD Ultra? Do you have a HDCP link? Yes, XP requires additional codecs because it's a pre-HD OS.

Extended monitor, it depends on the TV and the input. Try Theatre mode if you want video to play on the TV exclusively.

I never found the Intel board crashed - like all Intel chipsets it worked perfectly. There's something wrong with your setup I suspect :(