Too bad it will probably cost ~ $200 in AU

Yeah, really sucks how everything has to cost more in Australia.

$200 in America can buy a very good graphics card.

Ok. There is a big flaw, all those tops have in common, it's not just X-bit fault. Why do you have to test all those video cards with super expensive processor?!? Just to brag about you having the coolest and fastest CPU???? I mean this proc. is only possesed by 0.5% of the population! Most of the people are still using dual cores, not everybody has money for extreme quad cores and stuff. Besides, using such fast processor, the real life performance are still unknown, and the decoding thing? I mean, common!!! Who cares the proc. is 5% or 35% on an I7 extreme proc???? Wouldn't be more interesting to use a dual core like E8400, to see exactly how much it will make the difference using hardware decoding?? Just think a little to those aspects.

The four offerings of this product on newegg all have a different cooler, which is considerably worse looking, and probably worse preforming. This might change the conclusion of the review considerably. It also might be how they "fixed" the issue with random speed ups of the fan, although i hope this is not the case , as i would take random speed ups any day over the new cooler.

Considering it only consumes a smidge more power than the HD4670, I also think it could do without the PCIe connector even though ATI wants it on there.

sweet review ! keep up the good work Guys o/

Loved the review.But there are some inconvenient numbers in some games like HAWX and Far Cry 2.Maybe testing both ATI and Nvidia cards using MS-AA as someone mentioned previously, would fix those numbers.Otherwise, the article is well made and the charts are very obvious and easy check.
All in all, it's an excellent article but some modifications should be made to the anti-aliasing modes used for the tests.

As usual Xbit is totally biased towards ATI. I would really like to see a non biased review where the settings for the cards are not set against Nvidia's favor. Seroiusly setting up the nvidia cards for physx and multi-monitor performance when you are only running tests with one monitor. The physx is why many of the games have odd frame rates for Nvidia cards. There are a ton of games that support the physx used by nvidia cards and with that setting enabled, you will never get a real good comparison between ATI and Nvidia cards. Usually you can go to many review sites and find that the fps and numbers are generally the same give or take some for the system setup the benchers are running, but xbit labs takes the cake for having the lowest nvidia scores out there, even when reviewers are using dual core stock speed setups.

What i find funny is that you say this 4770 has good oc potential, when it is so obvious that its 128bit memory bus screws its OC gains. When you oc a chip to more than 100mhz above core and 200mhz for mem, you should expect to see a 8-10 fps increase not the 1-2 if even that in some games.

The 4770 is a nice budget card, perfect for shoving into the Dell's and HP's out there that come with onboard vid. Aside from that it looks like AMD has shown that it has a new process tech that does really well, and im sure will shine better on a 256bit mem bus and scale much better.

Why is there no way of hacking CCC so that its overdrive allows for higher overclocks.
I'd much prefer to overclock at a driver level as opposed to Riva Tuners hardware level since it also overclocks in 2d-mode.

Have you discussed the HQV video quality results with ATI?

It is surprising the card does less well than the 4670 with DVDs. Sounds likely to be an issue with early drivers. Have they said anything? Are you planning on re-testing?