Originally Posted by darinf
In any case, Headphone X sure looks like the same thing as the Realiser, but in a more generic fashion rather than doing personalized ear/HRTF measurements.
A key difference is that if you want to hear your recordings as if played in the Concertgebouw with Realiser, you have to take your ears to the Concertgebouw to get them measured there. With Headphone:X that room model comes with the recording.
DTS could be using a binaural dummy head/ears to do their measurements to get a more generic HRTF, but even if you don't use a generalized head/ear shape, the effect works well, from my experience.
Whether it is a dummy head or based on real humans (dummy heads are derived from humans, too), it is still an approximation to any given listener’s personal HRTF. The trick is to find a good one, but moreover, to render it effectively. Typical headphone processors, to reduce computations, generalize what may have been a great HRTF into a "low res" version that degrades spatial/timbral accuracy. Realiser works not only because it starts with your own HRTFs, it preserves their details. That's not easy when computation capacity is limited.
The biggest problem with Headphone X is having to re-encode all content AND implement new codecs in all the playback devices.
Oh yes, and of course, Headphone X will never work with Dolby encoded audio.
Yes, the optimal solution to hearing the experience as in the original venue is to use soundtracks which carry the embedded room information. But the headphone rendering, with respect to HRTFs and room modeling, will work just as well with any existing content. I suspect they will offer a set of generic room models from which the user may choose. I may end up listening to a movie in a different dubbing stage, but do I really care?
Originally Posted by David Susilo
1. DTS is not using binaural dummy, it uses room modelling
These are two different things. Room modeling is not a substitute for HRTFs, it is in addition.
Originally Posted by darinf
2) On the DTS website when they say "requires surround content to be encoded as a DTS-HD bitstream, with the room information embedded in the stream" and then say "Finally, the reference data is decoded ". That may not be a codec, but it's the same idea. Or are you saying that HeadphoneX uses pre-existing codecs? Or are you saying there is no encoding or decoding? Is it just a signal processing algorithm? If so, then why do they mention encoding and decoding?
The room information is “decoded” from the bitstream and applied to the post-processor. The audio codec itself is not changed.
3) I am just going by what it says on their website, "requires surround content to be encoded as a DTS-HD bitstream, with the room information embedded in the stream". You may have more information than what's on their website. But to me the key word is "requires".
Yes, if you want to hear the audio within the venue where it was produced, you need the room information for that recording. Otherwise, you will be in some alternative (but presumably good sounding) room.