BACCH4Mac Audiophile Edition 3D Audio Playback System

Availability: Direct
What's Included: Theoretica’s BACCH-dSP software, BACCH-BM binaural microphone, RME’s Babyface Pro
Price: $4,980.00 (Audiophile Edition)
Website: www.theoretica.us

"Ground-Breaking Technology"

I said that back in 2015 after my first experience listening to music processed through a BACCH 3D Sound processor (the BACCH-SP) at CES. Now that I've experienced BACCH in-barn, I would say that the BACCH4MAC 3D Audio Playback System is ground-breaking technology.

BACCH 3D breaks more ground than any digital technology I've encountered. That includes hi-rez (yawn), MQA, Ethernet cables, USB tweaks, etc. etc. While my Hi-Fi Hierarchy doesn't need to be entirely re-written, it certainly needs an update. This update would place the BACCH4MAC 3D Audio Playback System in the #5 position, directly after the DAC, moving the current #5, Plumbing, into the #6 position.

What Is The BACCH4MAC 3D Audio Playback System

While I've already addressed this question, let me give give an overview; The BACCH4MAC Audiophile Edition1 3D Audio Playback System eliminates crosstalk when listening to stereo and binaural recordings.

That's really it. That's all it does. The BACCH-dSP software does not mess with anything else in the digital signal, according to the good people at Theoretica Applied Physics. According to my ears, this a true statement.

When you buy the BACCH4MAC 3D Audio Playback System, you get what's listed up top; Theoretica’s BACCH-dSP software (for MAC only), BACCH-BM binaural microphones, and the RME’s Babyface Pro. The package can include an optional webcam (for an extra $80), or you can use any webcam including your Mac's built-in camera. The in-ear microphones go in your ears and are connected to the Babyface Pro, which in turn is connected to your MAC via USB, for the initial measurements. These measurements, which are stored in the BACCH-dSP software, coupled with the Webcam, which is connected to the MAC, allow the software to track your movements, while seated, from left to right so the 3D image remains fixed even if you don't.

"You lookin' at me? I'm the only one in here. You lookin' at me?

The measurement regime is fully automated and takes but a minute or so to perform. Easy peasy. There are 3 "bins" which allow you to save up to 3 different user profiles. Everyone's listening apparatus, pinna, head, etc, are different, so each profile will vary to some extent but when friends come over and listen through your profile, they'll be delighted with 3D sound.

the Advanced screen in BACCH-dSP software

Once these measurements are complete and stored, you may have no further use for the BACCH-BM binaural microphones unless you plan on doing some binaural recordings. The Babyface Pro can remain in the system and be used as a DAC, it has a pair of balanced outputs, or as a pass-through taking the USB output from your MAC and converting it to Toslink to send to your DAC. I will cover these options in a follow-up review. You can also connect your phono stage to the unit's line level TRS inputs so that your vinyl will also enjoy the benefits of 3D audio.

the Regular screen in BACCH-dSP software

The purists among you who might raise en eyebrow at the use of a pro audio (albeit well-built and well-regarded) USB interface such as RME's BabyFace Pro, should note, before typing disapproving comments, that you can completely bypass the DAC of the RME and use it as just a pass-through to your DAC, as I have done for this review. You can also take the RME out all together from your playback chain (it is only used when you need the input from its mic preamps to make the initial BACCH measurements) and route the audio directly from the Mac to your DAC via USB. In theory, doing that is a purer path but in practice it may be more convenient to leave the RME in so you can use it to make the measurements whenever you want without having to worry about switching device clocks. As you'll soon read, listening showed that it was doing no harm passing data through to my DAC.

the Minimal screen in BACCH-dSP software

In terms of interfacing with the BACCH-dSP software, once you complete the setup, which includes telling the software which connected device to output through and, if you like, what fixed sample rate to send it (the horror), there's really no reason to mess with it (Theoretica informed me that the latest version of BACCH-dSP, v 3.3, which was released after I did my listening tests, allows for adaptive sampling rate). The BACCH-dSP software can be a set it and forget it kinda deal (except when you want to impress friends and family members with the seeming complexity of it all, fooling them into thinking you know what every last thing on-screen means and does).

For this review, I used a borrowed-from-Theoretica MAC mini loaded with the BACCH-dSP software and the rest of the kit that every buyer gets. My totaldac D1-six, Ayre AX-5 Twenty, and DeVore gibbon Xs handled the analog bits. Tellurium Q Black cables are used throughout, with the exception of the Toslink link from the Babyface to the totaldac. Here I used a generic cable supplied by Theoretica and an AudioQuest Forest because I get nervous too. For playback control, I am very pleased to say that Roon was put to use running on my iPad. To so do, all one need do is point Roon at the desired output device.

Mind-Boggling Technology

Back at CES 2015, I was telling Joel A. about my BACCH 3D experience and he asked a very interesting question (Joel has a habit of doing that), "If someone told you you could have it for free as long as you used it all of the time, what would you say?" At that time I said I didn't know and only long-term listening would allow me have an answer. Well Joel, my answer is I would say, "Thank you. Thankyouverymuch." (like Elvis).

During my weeks of listening to my music in 3D, there was not one recording that was hurt by BACCH and most everything was helped, some more than others. The most extreme 3D listening experience was when sampling David Chesky's binaural recordings because binaural recordings and 3D audio are like peas and carrots: The binaural recording captures the location of every player in space and BACCH 3D reproduces what was recorded with extreme accuracy. Trombone player 15' directly to the left? Trombone player is reproduced 15' to your left. And so on. With pinpoint accuracy.

I know what some of you are thinking, who cares? Who listens to binaural recordings? And how many of them are there? True enough. But, and here you'll want to lean in, stereo recordings are reproduced with whatever spatial information exists in the recording with mind-boggling accuracy. It's like listening to old, familiar recordings for the first time. Yea, I went there.

To boggle means, among other things, to be "astonished" and I've chosen my word advisedly. Listening to Jimi Hendrix's "Rainy Day, Dream Away" I was astonished with the uncanny unfolding (wink) of the space of the recording because, little had I known, lo these past 40+ years, Jimi's voice during the intro comes at you from hard right, a few feet to the right of the speaker and about a foot in front, while the rest of the band is located around the room some farther back and centered, some hard left, etc. The sound image exists in three dimensions as a solid, stable, reproduction. The same level of "wow" occurred when listening to Einstürzende Neubauten's "Ring My Bell" from Tabula Rasa where sounds were coming from all over the place, completely divorced form the speakers and room.

I also queued up some tracks that are already spatially wow-inducing including every single one from Fritz Hauser's stunning Solodrumming and the Ensemble of Irreproducible Outcomes' Intonazione/The Foggy Dew and with BACCH's help, they became even more alive in-barn, completely obliterating any sense of stereo speakers sitting in front of me. Remember, this is not something that's been added to the digital signal, I'm hearing a more accurate reproduction of the recording's spatial information. Yea, I know—Wow.

Once I got over the "Wow" (kinda), I noticed that listening through BACCH as a normal every day thing made me more relaxed, more immersed in the music. This is an interesting observation because in my experience, listening to digital can make me uptight. Uncomfortable, edgy, and feeling like I'm swimming upstream in tar, more or less. My totaldac wiped all that away or so I thought. Listening through BACCH 3D adds another layer of ease in that it expands the sound image dimensionally—more akin to the real thing.

Again, some recordings are effected more than others but none are hurt by BACCH. For me, the majority of the music I listened to, which consisted of the music I would otherwise listen to, benefited from BACCH. If you move outside of the head-tracking zone, you hear regular old stereo unless you stay in-line with the listening seat and sit behind it. In the barn, the BACCH 3D sweet spot extended some 8' behind the red Eames LCW.

If Wishes Were Fishes

I'd love to see the BACCH software have a "network control" feature, like HQPlayer, which allows Roon to send its data to the BACCH software and then have BACCH send its output to a Roon Ready device. For example, I'd love to be able to use my dCS Network Bridge with the BACCH software.

I would also like to explore other options beyond the RME Babyface even though it sounded very, very good as a pass-through device. I am most interested in trying the Tascam UH-7000 ($400) because it has XLR out which I prefer over Toslink. At least in theory ;-) While the RME Babyface Pro is the only device that Theoretica supports specifically (see the dynamic chart on BACCH4Mac’s website that shows the various ways the device can be used with various possible playback setups), I am told by the company that some of its customers have used other mic-preamp-equipped USB interfaces with no problem.

I suppose a better way to get at my wish list is to simply say I'd love to see BACCH4Mac integrate into my existing system, dCS Network Bridge and all. My guess is there are many of us with microRendus, SOtMs, and various and sundry other network devices that we'd be hard-pressed to give up for the RME's Toslink out. [/wishlist]

Space Is The Place

Unlike some things in hi-fi, BACCH 3D is in no way controversial. It's not even a new concept, 3D audio, and it's rooted in real science (look at its lineage, born out of a Princeton lab, and check the credentials of its maker). The only real question is—does it work with enough music to justify the cost?

My answer is ab-so-lutely! All one need do is look at what we are willing to spend on incremental improvements from cables, file formats, tweaks, etc. and I'd go as far as to suggest BACCH4Mac is a bargain considering how much it improves reproduction.

To put it another way, if I did not do this for a living, where I need to have a system that will accommodate all manner of stuff, I would be thinking long and hard about purchasing BACCH4Mac. Truth be told, I'm thinking long and hard anyway.


1. Here's a pic from Theoretica of the other BACCH4Mac options:

Associated Equipment

COMMENTS
PeterMusic's picture

Awesome! Thanks, Michael. Just as many of us have hoped/expected. A couple of practical questions--probably for the manufacturer directly, but I'm sure many readers will want to know. First, are there plans for a national rollout? From the website it looks like only two dealers in the US, so a bit difficult for a person who lives in Boston, just for example, to hear for themselves before buying. Second, what is the return policy?

Michael Lavorgna's picture
Yes, these are questions best answered by Theoretica. I believe they will be checking in on the comments here.
bgardineer4's picture

Thank you for your interest in BACCH4Mac. Currently the BACCH4Mac package is sold only directly from Theoretica in the US. At present our US dealers can give demos of the BACCH-SP processor (which is the standalone hardware version of the BACCH4Mac product). We plan to do demos at upcoming audio shows around the country and will post announcements about this on our website. As to the next event: Professor Choueiri, President of Theoretica, will be giving a 1-hour tutorial talk on BACCH 3D Sound using the BACCH-dSP software at the 143rd AES Convention in NYC this October and there will be demos of BACCH4Mac following his talk.

The return policy has been posted at the bottom of the BACCH4Mac webpage (https://www.theoretica.us/bacch4mac/).

Buddy Gardineer
Development Engineer, Theoretica Applied Physics

CarterB's picture

I'm not sure if I can have a permanent webcam set up in my system. Is the webcam just for set up or part of the day-to-day playback?

bgardineer4's picture

The webcam is used for head tracking during listening, and it is also needed during the set up measurements (calibration) for head tracking to work during listening. You can forego the webcam (and therefore head tracking) and listen without head tracking. In that case the 3D image would be perfectly fine as long as you do not move your head too far to the right or the left (roughly +/- 3 inches). Without head tracking, we have found that people can adjust their head instinctively to get the optimal 3D image then relax. With head tracking you can be anywhere between the extreme left and right limits you chose when you did the calibration and the 3D image will stay rock solid. It is very effective.

You can place/hide the camera practically anywhere as long as it has a clear view of your head and as long as the perspective is not too lopsided for the head tracking algorithms to recognize a human head.

Buddy Gardineer
Development Engineer, Theoretica Applied Physics

PAR's picture

I have the impression that the system can only cope with one listener at a time due to the need to link to their stored profile. Presumably also only one person's head movements at a time can be tracked. Is that correct?

I also presume that as the processing is related to the particular listener's characteristics then whilst he or she will have an enhanced experience anyone else in the room will experience a degraded experience.

bgardineer4's picture

> "I have the impression that the system can only cope with one listener at a time due to the need to link to their stored profile."

While the stored profile is based on a measurement done with an individual head, what make two profiles sound different are the following factors (in decreasing order of importance):

1. The positions of the speakers and the listener
2. The characteristics of the speakers (frequency, impulse and phase responses, directivity pattern, etc.)
3. The level of sound reflections in the room
4. The Individual’s head (i.e. the so-called head-related transfer function, HRTF, which is a description of how one’s head and outer ears interact with sound waves).

In fact, the 4th factor (the Individual’s head) ranks quite low in importance as long as the speakers angular span (from the head of the listener) is kept below about +/- 45 degrees. Since most people listen with their speakers configured in the standard stereo configuration (a speakers span of +/- 30 degrees) or close to it, the difference between BACCH filters measured with two individual heads is often practically inaudible. (The technical reason for this is because most humans have similar HRTFs for frontal sound sources located in the azimuthal plane and confined to about +/- 45 degrees).

The bottom line is that if two BACCH filters (i.e. a set-up calibration measurements) from two individuals are stored in two presets (called "bins" in the software) of the BACCH-dSP software, a listener who switches between these two bins (which can be done with a click of the mouse) might perceive (if anything) a small difference in the location of the reproduced sound sources. That difference is subtle and, to many, inaudible. This is the case, of course, only if the the first 3 factors in the above list are the same for the two presets.

> "Presumably also only one person's head movements at a time can be tracked. Is that correct?"

That is right. The BACCH-dSP has a setting that allows zooming on the intended listener's head during calibration. During that calibration process the user can decide to make the sweet spot to be as wide as desired. The head of a listener (and only one listener) sitting in that sweet spot will then be tracked by the software during listening (a green cursor will appear anchored, almost too eerily, between the eyes of the tracked listener's head displayed in real time video on the screen).

Anyone whose head is not far behind that of the listener in the sweet spot will also perceive an excellent 3D spatial image. Since only the head of the person in the sweet spot is being tracked, the person behind that listener must move his or her head to always be behind the sweet spot listener to perceive the 3D image.

The sweet spot is narrow laterally (i.e. side to side), only +/- 3 inches, but can extend a long distance (longitudinally) behind the main listener. How far depends on the radiation pattern of the speakers. In the case of Mr. Lavorgna’s set-up the sweet spot extended more than 8 feet, as he reported in the review.

One can imagine the amusing sight (often witnessed at our demos at hi-fi shows ) of a row of 5-7 adult listeners sitting in a row behind a more privileged listener who makes them all move their heads to follow her head movement in order to stay in the 3D sweet spot.

> "I also presume that as the processing is related to the particular listener's characteristics then whilst he or she will have an enhanced experience anyone else in the room will experience a degraded experience."

The only perceptible difference between the sound in the sweet spot and that outside of it is that in the sweet spot the sound is 3D and outside of it is simply regular stereo (the same sound you would have without BACCH processing). In fact we often ask people to stand outside the sweet spot and switch on and off (i.e. bypass) the BACCH processing to see if they can perceive a difference. Virtually no one can. The lack of any tonal distortion imparted by the BACCH filter is the essence of the central patent behind the technology (one of the patents owned by Princeton University and licensed to Theoretica).

Unlike many audio technologies where bypassing (or instantaneous A/B’ing) is not encouraged, the bypass button on the BACCH-dSP software (or the BACCH-SP hardware processor) is one of its most used features during demos. Our customers, however, rarely turn the bypass button on again, after the initial few comparisons.

Buddy Gardineer
Development Engineer, Theoretica

FHC's picture

Dear Michael,

Thanks for bringing this ground breaking tech into the light. We couldn't be more pleased (and surprised) by what we heard. It's a new frontier. Looking forward to your introductions to other products in Bacch's stable. Phenomenal.

Fred Crane
Audio Prana

Michael Lavorgna's picture
Nice to hear from you. I hope all is well with you and yours.

Cheers.

insertusernamehere's picture

... in an earlier exchange you said that one couldn't use the head tracking w/o the RME Babyface. But the review seems to say otherwise: "[y]ou can also take the RME out all together from your playback chain (it is only used when you need the input from its mic preamps to make the initial BACCH measurements) and route the audio directly from the Mac to your DAC via USB;" the the dynamic chart on BACCH4Mac’s website also shows the webcam (... but NOT the in-ear mics) hooking up to the computer (vs. the RME Babyface).

So, just to be clear (and, perhaps, a little obtuse), one can remove the RME Babyface completely, obviously once the HRTF measurements are taken, and still use head-tracking? Thanks!

bgardineer4's picture

Yes, to use your own words and remove any ambiguity: One can remove the RME Babyface [Pro] completely, obviously once the HRTF measurements are taken, and still use head-tracking.

Again, the RME Babyface Pro is only strictly needed to get the stereo mic signals to the BACCH-dSP application running on the Mac for the HRTF measurements. (Even for that task it can be replaced by any USB interface that has a good pair of well-matched mic pres.) Once the measurements are done, the RME is not strictly needed but can be used, if one wishes, as a pass through to get the BACCH-processed audio out of the Mac (like in Mr. Lavorgna's review) to an outboard DAC, or as a DAC (for people who do not have a DAC and desire a turnkey BACCH system). Alternatively, one can get the BACCH-processed audio from the MAC to a DAC with a USB input through a USB connection, without hooking up or using the RME.

We recommend the RME mostly for its mic pres, which have outstanding stability, flat response, and are very well-matched (within a small fraction of a dB). We have also found, as supported by Mr. Lavorgna's listening tests, that as a passthrough, it is very transparent, but as I said, you can take it out altogether and get the BACCH-processed audio from the Mac to your DAC any way you want.

Buddy Gardineer
Development Engineer, Theoretica

gorkuz's picture

Interesting if complex tech. Reminiscent of Ralph Glasgal's "Ambiophonics", which over years progressed from a crude (and annoying) mattress placed immediately in front of a listener's head to physically stop interaural crosstalk to an algorithm processed method of stopping the crosstalk - apparently the same effect this is after (plus head position tracking).

Any connection between these two?

seatrope's picture

Hi Buddy, thanks for taking the time to answer questions. I had emailed the info@ email yesterday but have not heard back.

1. I understand that sample rates are now allowed to vary up to 192kHz? What are allowable input sample rates to the BACCH filter/program?

2. I have quite a bit of DSD native music in my library. Are there any plans to incorporate DSD into the BACCH filter system? I'm guessing not since it seems to be quite complicate to perform operations on DSD.

3. I am a staunch PC/Linux user. Will there be any option for a PC version of the BACCH program?

4. Finally, is there an option where the output of the BACCH program can be streamed via Ethernet to a small low noise playback computer such as a microRendu or a SOTM SMS-200? I own a microRendu and would like to keep using it – it takes a Ethernet input and can function as a DLNA or MPD receiver. Alternatively, if the BACCH-SP program can pass audio within the Mac to HQPlayer, the microrendu can serve as a HQplayer endpoint as well.

bgardineer4's picture

Ambiphonics is another method for crosstalk cancellation which relies on simple time-domain manipulation of the signal (the so-called RACE algorithm). Unlike BACCH 3D Sound, it is not based on individualized HRTF measurements that take into account the (frequency and phase) response of the speakers and the listener’s head and ears, and therefore cannot achieve the high levels of crosstalk cancellation that BACCH can. This leads to a significantly lesser ability to reproduce the depth of a 3D sound field and the accurate proximity of sources to the listener.

Moreover, one of the main features of BACCH filters (as explained in the main BACCH patent (https://www.google.com/patents/WO2012036912A1?cl=en) and in more detail in chapter 5 of the new book “Immersive Sound” (https://www.crcpress.com/Immersive-Sound-The-Art-and-Science-of-Binaural...) is that they are optimized to have completely (ruler) flat frequency response at the speakers, which RACE/Amphionics cannot do. BACCH filters are therefore completely free from tonal coloration and have zero dynamics range loss, which makes them ideally suited for the exacting standards of high-end audio.

Buddy Gardineer
Development Engineer, Theoretica

bgardineer4's picture

I apologize for the delay as we are in the midst of a production run of the new BACCH-SP processor. I was just about to respond to your to email but I will do so here.

Q1: "I understand that sample rates are now allowed to vary up to 192kHz? What are allowable input sample rates to the BACCH filter/program?”

A1: The BACCH-dSP program of BACCH4Mac can be set to process at the native sampling rate of the audio, up to 192 kHz.

Q2: "I have quite a bit of DSD native music in my library. Are there any plans to incorporate DSD into the BACCH filter system? I'm guessing not since it seems to be quite complicate to perform operations on DSD.”

A2: As you guessed, it is indeed extremely difficulty to carry out computations on 1-bit audio streams, and therefore DSD is not well suited for sophisticated processing such as BACCH 3D Sound.

Q3: "I am a staunch PC/Linux user. Will there be any option for a PC version of the BACCH program?”

A3: The BACCH-dSP program is the fruit of a Mac-based university research laboratory. Porting the program to work stably on another operating system, while doable, is a daunting task that we have not prioritized. That is why the commercial product is called BACCH4Mac. We are sorry that we have no plans at present to produce a PC version.

Q4: "Finally, is there an option where the output of the BACCH program can be streamed via Ethernet to a small low noise playback computer such as a microRendu or a SOTM SMS-200? I own a microRendu and would like to keep using it – it takes a Ethernet input and can function as a DLNA or MPD receiver. Alternatively, if the BACCH-SP program can pass audio within the Mac to HQPlayer, the microrendu can serve as a HQplayer endpoint as well.

A4: The output of BACCH-dSP can be streamed to any audio device that appears as an output audio device in the Mac’s list of devices in Mac’s Audio Midi Setup application. However, since microrendu does not have a USB input (only a USB output) it cannot serve this purpose. Furthermore, since the audio is coming from a Mac that is running BACCH-dSP, it does not make sense to send the audio to another computer that would send it to a DAC via its USB output. Instead the audio should be sent from the Mac via USB directly to the DAC.

I hope that I answered your questions.

Buddy Gardineer
Development Engineer, Theoretica

X