Isn't this backwards? Reading the spec (well, the wikipedia page) for DVI, and making some educated guesses, it seems to me the only way to know that a DVi->HDMI adapter is connected is if it provides EDID saying it is an HDMI device.
What is EDID I hear you ask? Well, it's extended display identification data. Physically, it's provided by an EEPROM on the I2C bus of the DVI port at address 0x50.
Could it be as simple as only AMD providing adapters that accurately portray themselves as HDMI devices?
EDIT TO ENTIRELY REVERSE IMPLICATION OF ABOVE QUESTION: No, it sounds like said EEPROM is necessary to make the video part of such an adapter work at all.
It seems AMD additionally identify if it is a blessed adapter and only send audio packets if that is the case. Definitely annoying, but fair given that there is no actual standard. People would definitely moan more if AMD cards blew up non-AMD adapters/monitors connected via non-AMD adapters.
EDID tells you whether you're connected to an HDMI device.
AMD added a special chip to their adapters that adds extra data to the EDID data in-line to say "yes i am a real ati adapter." This was totally unnecessary.
> People would definitely moan more if AMD cards blew up non-AMD adapters/monitors connected via non-AMD adapters.
Oh please—Nothing is going to "blow up". Sending audio packets in the stream is not going to affect adapters (adapters are passive--no electronics!) and any reasonable monitor is only going to pay attention to packets it understands.
And that's not even mentioning that you can use EDID to determine if a monitor is capable of receiving audio packets or not.
This is entirely sensationalist. There is no safe way to just blast audio data to any old DVI monitor. Something bad might happen, and AMD didn't want to take that risk. So they (wisely) including configuration to disable a non-standard feature when they didn't think it would be supported. The only thing I could see them doing better is allowing vendors to add themselves to the whitelist, but that amounts to starting a standards group for audio over DVI.
EDID tells you whether you're connected to an HDMI device.
AMD added a special chip to their adapters that adds extra data to the EDID data in-line to say "yes i am a real ati adapter." This was totally unnecessary.
Actually, they intentionally crippled their DVI ports. Their HDMI adapters "uncripple" said ports. FTA:
Included with many graphics cards are DVI-to-HDMI adapters for running an HDMI monitor off a DVI port on modern graphics cards. However, for whatever reason, if you want to use HDMI audio, AMD only wants you using the adapters included with the graphics card itself.
Is audio even included in DVI? They're probably including it in some non-standard way that differs from other, non-standard ways that other video card manufacturers use to adapt DVI to HDMI with audio.
It might not be standard, but it's "close enough" that other adapters apparently works once the check is disabled.
But perhaps that wasn't the case when this solution was first implemented and they were just covering their ass in case and nobody have thought to change it.
This article is sensationalist and inaccurate, and the site has full screen splash ads without any obvious way to dismiss them. This article should be removed.
its not like they are crushing nvidia and can do whatever they want with a captive market.
also not documenting it will prevent people from giving them money for adapters. ..maybe it is riaa enforcing protections against the analog loopback on drm content? that is the only explanation. they are getting money to do that to users. so you can be the product even when paying?
I am always perplexed by stories like this. These ports are on all sorts of machines from all sorts of vendors - deliberately crippling your self-branded hardware that can easily be swapped out for a working version (potentially with a refund in some jurisdictions) seems very shortsighted.
What is EDID I hear you ask? Well, it's extended display identification data. Physically, it's provided by an EEPROM on the I2C bus of the DVI port at address 0x50.
Could it be as simple as only AMD providing adapters that accurately portray themselves as HDMI devices?
EDIT TO ENTIRELY REVERSE IMPLICATION OF ABOVE QUESTION: No, it sounds like said EEPROM is necessary to make the video part of such an adapter work at all.
It seems AMD additionally identify if it is a blessed adapter and only send audio packets if that is the case. Definitely annoying, but fair given that there is no actual standard. People would definitely moan more if AMD cards blew up non-AMD adapters/monitors connected via non-AMD adapters.