DTS-HD/Dolby TrueHD audio delay on CII or RadianceXD

[mux_martin]

I’m fairly tempted to add a scaler to my setup and after reading up on the subject I think the Crystalio II or the RadianceXD would do nicely.

However, as it would be an investment to live with for a few years to come I’m obviously concerned about future proofing it. So my question is this: How can either of these scalers handle the audio delay of the new uncompressed audio formats from Dolby and DTS?

Cheers, Martin

[VirusKiller]

My gut feeling is that if you want future-proofing, it has to be the Radiance (which is of course only just going into beta).

[mux_martin]

Maybe I should clarify my question a little. Obviously HDMI 1.3 would be required to handle the new formats internally in the scalers. And since it seems to take some time still for 1.3 chips to become readily available I’m wondering if there is any other way around the problem.

Can for instance either CII or RadianceXD output the video delay via e.g. RS-232 to be handled by some external audio delay box?

[Gordon @ Convergent AV]

I'll ask Jim about how this will or if it will be handled. You would require to buffer and delay 6 PCM channels of info for these sources I believe. Radiance on board processing has very low latency compared the Vision range, which is already faster than the competitions I believe. So it may well be a moot point. Audio delay is in Radiance primarily because market dictates it is a feature that is required, not because it is actually required in this product (if it had been HQV chip they continued with then I think it's latencey is higher and it would have been a necessity).

Gordon

[Liam @ Prog AV]

There would be quite a high possiblity that if you have a HDMI 1.3 amp or surround processor capable of doing TrueHD/DTS-HD it might be capable of delaying the audio itself. Unfortunately looking into the future is still quite a tricky thing so it may be a case of having to wait until these things are more mainstream.

[mux_martin]

Yes, but the trick is to know how much delay to add as the video delay through the scaler could vary. But then again maybe it isn't much of a problem if the latency is greatly reduced on the new generation scalers, as Gordon pointed out.

Thanks for the replies anyway!

[Liam @ Prog AV]

I imagine it would be no different to how I set up AV Sync delay currently where I install Lumagen Vision models which do not have incorporated audio delay. I run Sky at about 70-80ms and DVD at closer to 50-60ms on a typical installation.

[ailean]

Doesn't HDMI have an inbuilt audio delay feature? Can't remember which version it's meant to be in but I'm sure I read that video processing devices (scalers, tvs, pjs) can send a delay factor over the HDMI data path for an audio device to pick up and automatically apply that delay to it's sound output.

Of course as with so many HDMI 'features' the chances of it actually being implemented in all the devices in your chain and working as it should are fairly slim.

So long as everything (source, vp, amp) is HDMI 1.3 then you do have the option of passing the lossless compressed HD audio streams, which would require less memory to delay inside a VP then the uncompressed PCM multi-channel format you have to use with HDMI 1.1->1.2.

But there's not a lot of 1.3 around so I'd presume the Rad will be designed to coupe with full PCM and even if it is really fast there's got to be a delay of at least 3-4 fields surely to get accurate de-int and frame lock not to mention good (no detail loss) noise reduction?

[Gordon @ Convergent AV]

Quote taken from Jim Peterson of Lumagen on AVSforum.

"The RadianceXD, using the Gennum, will have approximately 1.25 fields of delay for video and film, when in genlock mode. This is about 20 mS in the US and 25mS in PAL countries. By comparison, current products using the SiI504 and other current products, such as DVDO, have between 55 and 80 mS in US and 70 and 100 in PAL countries, of video delay. The Silicon Optix is about 150 mS.

Most digital displays add one or two more fields of delay.

Generally, from feedback from many people I have talked to, most people do not notice a lip-sync error of 50 mS but do notice the lip-sync at 75 mS.

So, even without running audio through the RaidanceXD, and using a display with two fields additional delay (the most I have encountered driving at the displays native resolution as we do), the RadianceXD will not have visible lip-sync error. For all other processors mentioned, lip-sync can be noticed unless the audio is delayed to match (which can be done by a good quality receiver or amp).
"

[ailean]

Does sound impressive but just how accurate can it be with no look ahead, is this like a good source lock mode with an option for a slower mode for less reliable sources?