Virtual Stage - Page 2

Here's a very recent review of the Realiser: http://www.positive-feedback.com/Issue63/smyth_realiser.htm.

It's $4k including a Stax headphone set. It's "only" $3k without the headphones.

The concept does sound very intriguing, but looks like you need to go to a lot of effort to capture the sound stage you want - or exchange "sound stages" with other owners.

As far as I understand it: the head-tracker ensures that the sound does not move with your head movement, so the soundstage stays stable just like with loudspeakers. In other words, sound coming from the front will come from the right if you rotate your head 90° to the left.

As for localization, the ear has a different transfer function for different positions of a sound source. This means different time delays and level differences between left and right ear and also different frequency response. Since this is highly individual they ship in-ear mics to record your individual HRTFs.

There is probably not one, but several direction dependent HRTFs (different FR amplitudes and phases as a function of direction.) Therefore time delays (and amplitudes) maybe frequency and position dependent. The sound will "see" a different side (shape and absorption) of the ear as a function of direction, which among other things may be responsible for these different HRTFs.

Check this out (binaural recordings have the HRTFs built in because they are recorded with a dummy head) - hearing from the perspective of the dummy head (min 3:50 is quite interesting):

http://www.youtube.com/watch?feature=player_embedded&v=ecOrBqQAuXg#!

This one gives sound from rear channel (starting from min 0:43) - close eyes and locate the sound source:

http://www.youtube.com/watch?feature=player_embedded&v=ZWg2M4-pR_I

(From LFF post: http://www.head-fi.org/t/511850/awesome-binaural-albums#post_6918133)

It's not going to be an exact match because our head and ears may not exactly match the dummy head. The Realiser supposedly calibrates and personalizes the responses to improve the sound stage.

Note that a poor headphone could mess up everything though ('cuz it messes up the FR.) The Realizer dudes suggest an open full size. In particular the Stax SRS-2170 package which is one of the more affordable open full size electrostats.

Quote:

Originally Posted by bigshot 

I scanned that article and it raised more questions than it answered. In order to hear sound all around you, do you have to keep moving your head around? With my 5:1 setup, I can detect whether something is in front or behind even if I am standing still. I don't know enough about how the ear works to know how directionality works, but it does it somehow.
Also, the review says that the unit took a song that was multi-miked and mixed with absolutely no attempt at coherent soundstage, and when it was processed, the instruments were all laid out separated in a realistic spread. That one sounds like a reviewer who is treading into hyperbole. I don't see how it could take apart a mix and rearrange the position of the instruments. I've heard synthetic stereo systems that attempt that, but inevitably, the pluck of the bass gets separated from the fundamental, and the harmonics of an oboe gets spread all over the place.
If it's reworking the stereo spread itself, not just placing it within a virtual space, it can make stereo recordings out of mono ones. Does it know to put violins on the left?
It sounds like most people use it like the DSP settings on an AV amp for home theater. I can totally see that. I use a DSP setting for converting stereo to 5:1 on my Yamaha receiver, but that's a subtle thing, and it doesn't change the front soundstage imaging built into the recording. I don't know if I want a DSP to monkey around with that.
It's been my experience that I don't want multiple synthetic environments. I just want the one that works right in my particular room to fill it with a natural balance. I definitely want the layout of the channels that the engineer mixing the film creates in 5:1 movies. I wouldn't want to add reverberation simulating a big cavernous theater or anything like that. And in stereo and mono music, I want to maintain the forward spread the way it is in the recording.

I guess all this is answered in the Realiser threads, but let me chime in so there will be no confusion for anyone stumbling here.

Realiser does not do any "DSP magic". It does not take apart, rearrange or synthesize anything.

Everything will sound just like it would, with you sitting in that room with those speakers playing. There is no "virtual space", it's _that_ space that you recorded, including the ambiance and any acoustic properties.

If you want to upconvert stereo to 5.1 or something else "DSP magic" like, that has to be done by you in the source. Realiser does not mess with the sound (but it does have some EQ/LFE etc tuning capabilities).

There is no need to use the head tracker if you don't want to. You don't need to move your head around or stay still like a statue. No matter what, the imaging is uncanny, speaker locations are always there. Of course to best fool your brains, you should sit still and/or use the tracker, but regardless the result is remarkable. Anyways if you are listening some other room that you are not currently sitting in, it's hard to completely fool brains. I'm currently listening to a 7 x Genelec 8260A setup, but how am I supposed to "believe it" when I don't see the actual speakers? Sound is coming from my walls. Yet the speaker positions are dead accurate, imaging is spectacular and it's nothing like Dolby Headphone and other simulation crapola.

Quote:

Originally Posted by hekeli 

 

I guess all this is answered in the Realiser threads, but let me chime in so there will be no confusion for anyone stumbling here.

 

Realiser does not do any "DSP magic". It does not take apart, rearrange or synthesize anything.

 

Everything will sound just like it would, with you sitting in that room with those speakers playing. There is no "virtual space", it's _that_ space that you recorded, including the ambiance and any acoustic properties.

 

If you want to upconvert stereo to 5.1 or something else "DSP magic" like, that has to be done by you in the source. Realiser does not mess with the sound (but it does have some EQ/LFE etc tuning capabilities).

 

There is no need to use the head tracker if you don't want to. You don't need to move your head around or stay still like a statue. No matter what, the imaging is uncanny, speaker locations are always there. Of course to best fool your brains, you should sit still and/or use the tracker, but regardless the result is remarkable. Anyways if you are listening some other room that you are not currently sitting in, it's hard to completely fool brains. I'm currently listening to a 7 x Genelec 8260A setup, but how am I supposed to "believe it" when I don't see the actual speakers? Sound is coming from my walls. Yet the speaker positions are dead accurate, imaging is spectacular and it's nothing like Dolby Headphone and other simulation crapola.

I think I understand what you are saying: The Realizer does not "upconvert stereo to 5.1."

But it does seem do some "DSP magic" though the application of the personalized HRTFs to the individual channels...

physiologically the eardrum is small relative to audio wavelengths and samples the sound pressure arriving at its surface, the vibrations are coupled to the Basilar membrane by essentially 1 degree of freedom mechanical links in the stirrup/staples - so everything we "get" from sound comes from essentially 2  single channels of time, 1-D motion each

the directionality of our hearing is believed due to frequency dependent shading, diffraction, of the arriving 3-D sound field interaction with our body, head, pinnae and by extensive "modeling" built in to our low level neural processing which also correlates "features" between R, L ears, body pose, visual clues...

we (our brain's low level neural wiring) learn to interpret clues from our environment's and head/body pose shaping of the sound - stereo and other sound reproduction schemes hide their limitations by exploiting our internally wired model's expectations - using sound perceptual illusions rather than encoding, reproducing a realisitic 3-D sound field of the live event

the SVS Realizer head tracking is only good for ~ +/- 30 degrees - but is critical for maintaining the outside of your head sound stage - you may not be conscious of your small head movements while "sitting still" music listening or movie watching but your brain integrates all of its information including knowing these small body position shifts - try listening with your head really clamped and front/back discrimination is greatly diminished - correct for hrtf frequency response differences and it goes away altogether

Quote:

Originally Posted by jcx 

the SVS Realizer head tracking is only good for ~ +/- 30 degrees

What does this mean?

the SVS Realizer calibration process involves your sitting in a real room with a real loudspeaker system and measuring the sound at your ears, you are instructed to turn your head in small increments over ~ +/- 30 degrees in the horizontal view plane with a frequency chirp measurement made by the system for each angular increment

during listening the Relaizer head tracker senses your head angle relative to the unit and interpolates between the measured angular responses

to keep things tractable the practical compromise is to only cover a restricted front facing range of angles in the horizonal plane only - current stereo and surround systems all assume only a single source plane, no up/down information is explicitly encoded

Quote:

Originally Posted by ultrabike 

I think I understand what you are saying: The Realizer does not "upconvert stereo to 5.1."

 

But it does seem do some "DSP magic" though the application of the personalized HRTFs to the individual channels...

Sure, you can call the basic HRTF magic anything you want. 

Just saying that it does not additionally modify the signal or channels etc in any way. It plays what you feed it, in the channel configuration you feed it.

edit: To clarify, yes of course the basic HRTF functionality involves all sort of DSP/filter/etc magic. What I mean with channel configuration are the "virtual speakers" you would hear.

Quote:

Originally Posted by hekeli 

 

Sure, you can call the basic HRTF magic anything you want. 

 

Just saying that it does not additionally modify the signal or channels etc in any way. It plays what you feed it, in the channel configuration you feed it.

The HRTFs are a set of filters, i.e. DSP magic. I don't know exactly what the Smyth Realizer does, but headphones are two channel. To map 5.1 to two channel one would likely need to pass the front and back-right and back-left through the appropriate filters to map them to the left and right channels of the headphone. There are also cross-feed filters. Then all contributions would have to be added to the right and left channels that go into the headphone. Dunno if they do more stuff to the signals... That seems to be quite a bit of signal massaging.

That said, trying to get 5.1 out of 2 channels seems very difficult... that might need some real magic

Quote:

Originally Posted by jcx 

the SVS Realizer head tracking is only good for ~ +/- 30 degrees - but is critical for maintaining the outside of your head sound stage - you may not be conscious of your small head movements while "sitting still" music listening or movie watching but your brain integrates all of its information including knowing these small body position shifts - try listening with your head really clamped and front/back discrimination is greatly diminished - correct for hrtf frequency response differences and it goes away altogether

I actually dislike the tracker, sometimes the changing sound field sounds unnatural. My brain has probably already learned to function fine without it, I really don't find any major diminishing in the effect if any.

BTW guys, here is a link to an example of 3D audio allegedly using HRTF banks:

http://www.youtube.com/watch?v=P-87i9rLwZE

Hope you guys like it

This one is really nice (the matchbox was very convincing)

http://www.youtube.com/watch?v=oPTa4_HrPhs&feature=related

Classical music (Beethoven):

http://www.youtube.com/watch?v=uDiTTXlchiY

Classical music (Bach/Kempff)

http://www.youtube.com/watch?v=E55CsPXChtA

Quote:

Originally Posted by hekeli 

 imaging is spectacular and it's nothing like Dolby Headphone and other simulation crapola.

In all fairness, if someone is close to the generic HRTF that Dolby uses DH can offer a relatively impressive option for the price.  With DH it sounds like my headphones are actually 5.1 speakers themselves.  It's not of Realizer quality in that it is not a perfect match for my HRTF, can't mimic listening rooms/systems, and can't track my head.  However, for gaming and movie purposes with my HRTF, it's a stretch calling it "crapola".  Of course, I think the hardware based solutions are better than the software DSP versions I've tried.

For the cost it can be an interesting solution, as long as one realizes their mileage may vary - for those that twitch at $3K it may be at least worth trying.