A article on BACCH4Mac that answers a lot of questions.

[mieswall]

Hmm, to be clear, this is how you should diagnose the problem. The signal chain is:

Phono --> AN1/2 input --> B4M --> AN1/2 output

Questions you should ask yourself:

1. Is the AN1/2 input receiving input? Your metering dancing bars clearly says yes. So it is.
2. Is B4M receiving input from AN1/2? I don't know if B4M has input metering or any way to show signs of activity. If there is no sign of activity, your chain is broken at B4M. Check your B4M settings to make sure that the inputs and outputs have been set correctly.
3. Is the AN1/2 output receiving input from B4M? If it is, you should see metering. If there is no sound, the signal chain is broken downstream from your RME. Check amplifiers, volume, cables, etc.

View attachment 499475

What this tells me is:

AN1/2 hardware output (DAC) is receiving input from AN1/2 hardware input, AN1/2 software input, and CH5/6 software input.
AN3/4 hardware output (DAC) is receiving input from AN3/4 software inputs ONLY.

If you need further help, can you please state:

1. Which inputs have you plugged your phono stage into?
2. What are the input and output settings in B4M?
3. What outputs have you used to connect your Babyface Pro to your amps? AN1/2 or AN3/4?

Thanks Keith,

1. Yes, It is clearly receiving, even clipping with the signal unless the gain is almost 0.
2. No: B4M is NOT receiving. The bars of the bins show no activity when fed by AN1/2 signals, but they do when other sources are used.
3. No, in the B4M implementation babyface does not connect directly to amps. What B4M does is to build a "virtual dac" using the "Aggregate device" feature of the midi setup of macOS. I understand the signal is routed from B4M to your standard DAC (part of the aggregate device), its analog outputs connected to the preamp.(*)

(*) I recall Edgard once installed me an option to use the babyface also as the lone DAC, without a second one. But that needs to install a freeware for routing signals to B4M, then back to the babyface. B4M interacts in the middle of the path of the digital signal.

I think the problem is that the routing config of babyface that Egard implements specifically avoids the AN1/2 being routed to outputs, as those inputs are used by the calibration mic.

 

[mieswall]

Hmm, to be clear, this is how you should diagnose the problem. The signal chain is:

Phono --> AN1/2 input --> B4M --> AN1/2 output

Questions you should ask yourself:

1. Is the AN1/2 input receiving input? Your metering dancing bars clearly says yes. So it is.
2. Is B4M receiving input from AN1/2? I don't know if B4M has input metering or any way to show signs of activity. If there is no sign of activity, your chain is broken at B4M. Check your B4M settings to make sure that the inputs and outputs have been set correctly.
3. Is the AN1/2 output receiving input from B4M? If it is, you should see metering. If there is no sound, the signal chain is broken downstream from your RME. Check amplifiers, volume, cables, etc.

View attachment 499475

What this tells me is:

AN1/2 hardware output (DAC) is receiving input from AN1/2 hardware input, AN1/2 software input, and CH5/6 software input.
AN3/4 hardware output (DAC) is receiving input from AN3/4 software inputs ONLY.

If you need further help, can you please state:

1. Which inputs have you plugged your phono stage into?
2. What are the input and output settings in B4M?
3. What outputs have you used to connect your Babyface Pro to your amps? AN1/2 or AN3/4?

1. I initially I had phono stage plugged to inputs 3/4 (line inputs) of babyface, to works, but give me occasional ground loop noise problems, so now I connected to the XLR inputs 1/2, with the problem discussed here: those signals not being routed to B4M.
2. I really don't know in detail, but see the issue of the aggregate divide in previous post. I think B4M uses babyface as digital signal source (analog signals going through ADC in the same babyface); then dsp'ed by B4m and passed through to you standard DAC. To do so, Edgard creates an aggregate device in the midi setup of macOS, as described above.
3. It is the standard DAC and not babyface outputs the ones connected to the amp.

Probably all this problem would be avoided if using the optional RME firebase UCX II interfase that Theoretica offers, that has more input channels, thus avoiding using the same ports used by the calibration mic as analog XLR inputs.

 

[Dialectic]

I bought the BACCH-sp as the b4m is a PITA if you ever try to change system settings. The SP is eye wateringly expensive but is an awesome unit. I can also switch between sanders ESL’s with the sp as the system controller and my main system with just BACCH processor into an optical jack on my avr.

May I DM you about the SP?

 

[Keith_W]

2. No: B4M is NOT receiving. The bars of the bins show no activity when fed by AN1/2 signals, but they do when other sources are used.

This is where your signal chain is broken, at B4M. You need to set it to receive input from AN1/2.

Alternatively, you can route AN1/2 (phono) into an existing channel that you know works with B4M. Your RME is able to route any channel into any other channel. You did not state what inputs your "other sources" are plugged into. So with this example, I am going to use OUT5/6:

In your matrix, click on the two boxes I indicated in red. This routes hardware input AN1/2 to hardware outputs 5/6. Now you need to route the hardware OUTPUT 5/6 to hardware INPUT 5/6. To do this, go to the mixer view. In the bottom row ("Hardware Output"), find OUT5/6. Click on the spanner to expand it, then click on "Loopback". It will glow orange to indicate that loopback is active. This sends output from OUT5/6 to IN5/6.

Pay attention to the terms I am using, because if you don't, you will get confused:

"Hardware input": denotes a physical port where external hardware physically connected to your RME for input. Examples: AN1/2 (mic input), ADAT3/4 (digital input) for phono, Wiim.
"Hardware output": denotes a physical port where your RME sends output to. Examples: AN1/2, AN3/4 (DAC channels), ADAT3/4 (digital out)
"Software playback": denotes signal coming from playback software. This is automatically routed from software input to the corresponding hardware output, e.g. AN1/2 to hardware output AN1/2.

I don't know if all this will work for you or not, but it's worth a try.

FWIW, as stated previously, I don't own B4M, let alone a Mac. This is what I would try if it was on Windows. I don't even have my RME interface with me right now so I can't experiment to confirm if it will work. So sorry, I would rather give you a guaranteed solution but at the moment this is the best I can do.

Merry Christmas!

 

[mieswall]

This is where your signal chain is broken, at B4M. You need to set it to receive input from AN1/2.

Alternatively, you can route AN1/2 (phono) into an existing channel that you know works with B4M. Your RME is able to route any channel into any other channel. You did not state what inputs your "other sources" are plugged into. So with this example, I am going to use OUT5/6:

View attachment 499639

In your matrix, click on the two boxes I indicated in red. This routes hardware input AN1/2 to hardware outputs 5/6. Now you need to route the hardware OUTPUT 5/6 to hardware INPUT 5/6. To do this, go to the mixer view. In the bottom row ("Hardware Output"), find OUT5/6. Click on the spanner to expand it, then click on "Loopback". It will glow orange to indicate that loopback is active. This sends output from OUT5/6 to IN5/6.

Pay attention to the terms I am using, because if you don't, you will get confused:

"Hardware input": denotes a physical port where external hardware physically connected to your RME for input. Examples: AN1/2 (mic input), ADAT3/4 (digital input) for phono, Wiim.
"Hardware output": denotes a physical port where your RME sends output to. Examples: AN1/2, AN3/4 (DAC channels), ADAT3/4 (digital out)
"Software playback": denotes signal coming from playback software. This is automatically routed from software input to the corresponding hardware output, e.g. AN1/2 to hardware output AN1/2.

I don't know if all this will work for you or not, but it's worth a try.

FWIW, as stated previously, I don't own B4M, let alone a Mac. This is what I would try if it was on Windows. I don't even have my RME interface with me right now so I can't experiment to confirm if it will work. So sorry, I would rather give you a guaranteed solution but at the moment this is the best I can do.

Merry Christmas!

THANK YOU KEITH... !
It worked!!
That was a true Christmas present you gave me. I'm really, really grateful.
One note for others: as this alters the settings needed by Edgard to support your systems, it is advisable to store these settings in another snapshot, so to recover Theoretica's settings in the snapshot Edgard originally made. Here stored as the 4th snapshot ("mc boost"). Then click in the third (at least in my case) if you need to restore settings made by Edgard.

 

[mieswall]

Having corrected the vinyl input in the mixer, it is now easy to compare the sound against other sources.
Even with a decent capsule I'm using (AT OC9XML), there are noticeable better XTC results with digital sources. I did comparisons with the record "Round Again" of Joshua Redman & friends, and the scene clearly opens up more using the streamed source (Tidal).
Perhaps this should be obvious: with only 25 -30 db of channel separation (27db in the specs of mine), there is less content to feed the XTC algorithm, that I understand deals with the differences between both channels to work its magic. Best results are achieved with 25-30db of XTC, and that is the limit of channel separation of the whole content in vinyl.
I'm curious this has not been commented previously anywhere, afaik.

 

[Gwreck]

Having corrected the vinyl input in the mixer, it is now easy to compare the sound against other sources.
Even with a decent capsule I'm using (AT OC9XML), there are noticeable better XTC results with digital sources. I did comparisons with the record "Round Again" of Joshua Redman & friends, and the scene clearly opens up more using the streamed source (Tidal).
Perhaps this should be obvious: with only 25 -30 db of channel separation (27db in the specs of mine), there is less content to feed the XTC algorithm, that I understand deals with the differences between both channels to work its magic. Best results are achieved with 25-30db of XTC, and that is the limit of channel separation of the whole content in vinyl.
I'm curious this has not been commented previously anywhere, afaik.

Really good results with BACCH can be had with 5-15 db xtc so I don’t think LP’s should be a limitation.

 

[mieswall]

Really good results with BACCH can be had with 5-15 db xtc so I don’t think LP’s should be a limitation.

I'm not sure both measures are equivalent: if the 27db of channel separation of vynil is comparable to, say, 15 db xtc (in fact, I have just 12 db here). I mean, if the channel separation received is the "source material" for xtc and being able to separate 12 db with that signal, those theoretical 27db of the capsule are good enough.
If so, it must be another explanation. In this nonesuch record, which I have pristine in vynil:

With volume levels previously leveled between sources, I obtain obviously greater differences in XTC when playing the streamed version in Tidal and Roon (only version available, 96/24).
My analog chain is: Rega p3, Audio Technica OC9xML (Mc), Music Fidelity MXVynl. DAC/Babyface operating at 96 Khz. Btw, I'm pretty sure the capsule is well aligned.

If the separation provided by the capsule is enough, there would be no benefit of feeding the XTC algorithm with a signal with a much greater channel separation, like the one probably most digital files provide.

I'll test other records to see what happens.

 

[mieswall]

In this album by Cyrille, Smith & Frissell (Lebroba, ECM 2018), with not so much effect on xtc other than make in it more tangibly real (if that's not enough...),

It also shows xtc enhancing more the digital file (Tidal 88/24) than the analog one. Both when comparing a/b the xtc data between both sources, and listening how much it changes with or without xtc on the same source. Specially noticeable on percussion by Cyrille, and in articular with cymbals.
I must say that my OC9xML is slightly less resolving in highs than the digital file, but this is more obvious when applying xtc than with plain unprocessed files.

Btw, great album...

 

[Gwreck]

I'm not sure both measures are equivalent: if the 27db of channel separation of vynil is comparable to, say, 15 db xtc (in fact, I have just 12 db here). I mean, if the channel separation received is the "source material" for xtc and being able to separate 12 db with that signal, those theoretical 27db of the capsule are good enough.
If so, it must be another explanation. In this nonesuch record, which I have pristine in vynil:
View attachment 501302
With volume levels previously leveled between sources, I obtain obviously greater differences in XTC when playing the streamed version in Tidal and Roon (only version available, 96/24).
My analog chain is: Rega p3, Audio Technica OC9xML (Mc), Music Fidelity MXVynl. DAC/Babyface operating at 96 Khz. Btw, I'm pretty sure the capsule is well aligned.

If the separation provided by the capsule is enough, there would be no benefit of feeding the XTC algorithm with a signal with a much greater channel separation, like the one probably most digital files provide.

I'll test other records to see what happens.

Are you saying that if the phonograph capsule has a certain amount of XTC between channels then BACCH would be unnecessary(or not beneficial)?

 

[mieswall]

Are you saying that if the phonograph capsule has a certain amount of XTC between channels then BACCH would be unnecessary(or not beneficial)?

No, I'm suggesting basically the opposite. That the 25db of channel separation of a capsule is not a measure (or limit) of how well XTC will perform with that material. If that were the case, there would be no benefit in the higher channel separation of some source material, and I think it´s clear the cross cancellation improves the better the source is. Even if a full cancellation is perceived with 25 db.
Or in other words, 25 db of channel separation is not equivalent to 25 db of cross cancellation , which probably needs a separation higher than that.
As (good) digital files provide that higher separation, XTC performs better with that material.

 

[Gwreck]

No, I'm suggesting basically the opposite. That the 25db of channel separation of a capsule is not a measure (or limit) of how well XTC will perform with that material. If that were the case, there would be no benefit in the higher channel separation of some source material, and I think it´s clear the cross cancellation improves the better the source is. Even if a full cancellation is perceived with 25 db.
Or in other words, 25 db of channel separation is not equivalent to 25 db of cross cancellation , which probably needs a separation higher than that.
As (good) digital files provide that higher separation, XTC performs better with that material.

Definitely mono material is not helped by BACCH and early studio recordings with hard panned instruments will have the instruments behind one’s head.

 

[Sir Sanders Zingmore]

early studio recordings with hard panned instruments will have the instruments behind one’s head.

Do you have any examples of this? It certainly is not my experience

 

[goat76]

The problem will occur with every audio mix that contains hard-panned instruments; they all will sound somewhat strangely isolated to one ear, as these hard-panned sound objects are actually expected to be heard by both ears and therefore require crosstalk to sound as intended.

 

[Gwreck]

Do you have any examples of this? It certainly is not my experience

I don’t have any specific examples but almost and early stereo studio recording has a mono and a stereo mix that one can compare.

 

[mieswall]

Do you have any examples of this? It certainly is not my experience

Take any album where you here an instrument is locked in one speaker (30 degrees of you centre axle, mono at that center). If you have b4m’ xtc of some 10 dB or higher, you’ll listen that instrument very close or straight at 90 degrees.
With some rare material where the mix plays with phases between channels, and depending on imaging capabilities of your speakers (say, ls50) and relative freedom of early reflections in your listening room (… those were several ifs…), you may hear sounds coming at 40-45 degrees in normal stereo. In those cases, b4m will Render them beyond your head, as far a 150 degrees in some cases.
If that rendering is too much (yes, it is unnatural in those early stereos), you may reduce the amount of xtc with a slider provided in the software interface.
But in (most) other albums , i guess that is exactly what the recording Engineer and musicians and would love to hear fron their mix,

Al in alll, i’ve found that in the long term those fireworks are not the most rewarding aspect of baach . It is the retrieval of the ambience of the recording venue. To really perceive the true reverberances stored in the recording, in a way that i believe no normal stereo can, no matter how good or expensive that stereo is.

If that ambience retrieval is part of what you value most of your listening experience (and it should if you dream about the most realistic reproduction, the one closest to the real thing), then baach will be the best music reproduction you have heard, even playing with fairly modest gear. But you need to set así de your cell phone and pay attention in your listening session. The more you focus in what you are hearing, the more you will value the experience.

I mention it because some people that i’ve showed my system seem to expect just instant reward. You end up showing them the chopers of The Door’s “The End” of the Apocalipsis mix, or the Bohemian Rhapsody of Queen (if they can wait long enough for the choruses …)… instead of recreating for them the experience of being in the Village Vanguard with Bill Evans at the reach of your hand, or in a big church listening Tocatta and Fugue played in a huge pipe organ.

 

[STC]

Definitely mono material is not helped by BACCH and early studio recordings with hard panned instruments will have the instruments behind one’s head.

BACCH is fundamentally a crosstalk cancellation (XTC) system, just like other XTC solutions out there. All of them work primarily through phase manipulation to cancel the unwanted signal that crosses over from one speaker to the opposite ear.

The thing is, you can push phase manipulation so far that the stereo image gets shifted dramatically—even out to ±90° or beyond. At that point, though, you’re no longer doing proper interaural crosstalk cancellation; you’re basically creating an artificial widening effect instead.

This extreme image shift often happens when the cancellation parameters (delay, attenuation, filter curve) aren’t dialed in correctly for the specific listening setup. It’s a very common reason why people try XTC, don’t hear the expected natural improvement, and conclude it “doesn’t work” or sounds weird.

In a well-calibrated XTC system, hard-panned sounds (full left or full right in the recording) should still appear only a few degrees outside the physical speaker positions—not floating way out wide or behind you. When they do stretch unnaturally far, it’s usually a sign the settings are off.

And just to be clear, this isn’t unique to BACCH. Pretty much every XTC implementation can run into the same issue if the cancellation isn’t properly matched.

 

[Keith_W]

This extreme image shift often happens when the cancellation parameters (delay, attenuation, filter curve) aren’t dialed in correctly for the specific listening setup. It’s a very common reason why people try XTC, don’t hear the expected natural improvement, and conclude it “doesn’t work” or sounds weird.

In a well-calibrated XTC system, hard-panned sounds (full left or full right in the recording) should still appear only a few degrees outside the physical speaker positions—not floating way out wide or behind you. When they do stretch unnaturally far, it’s usually a sign the settings are off.

This particular point requires special emphasis. I was under the impression from reading all those subjective reviews about BACCH - that it created a soundstage widening effect through 2 speakers. After talking to you, you said that the point is to obtain increased clarity, and the soundstage widening effect is an artefact of improper tuning. That was a bit of a lightbulb moment for me, because up till then I could not understand HOW crosstalk cancellation could possibly widen the soundstage, if it was not already in the recording.

So - a well tuned XTC should widen the soundstage IF (1) the widened soundstage is in the recording, and (2) your system would benefit from XTC.

 

[STC]

After talking to you, you said that the point is to obtain increased clarity, and the soundstage widening effect is an artefact of improper tuning.

Small clarifcation: that extreme soundstage widening happens when the cancellation gets pushed too far. Instead of just cleanly removing the interaural crosstalk, it starts adding extra artificial spatial effects that aren't actually in the original recording. It can be really exciting the first time you hear it—everything feels huge and immersive—but pretty soon you'll run into problems with certain recordings. The over-processing creates errors that are more noticeable and distracting than regular stereo without any crosstalk cancellation, and a lot of people end up losing interest because of it.

a well tuned XTC should widen the soundstage IF (1) the widened soundstage is in the recording, and (2) your system would benefit from XTC.

Another clarification here too: point (1) isn't quite accurate. In conventional stereo without XTC, crosstalk mixes signals between the ears, which compresses the soundstage and pulls images inward—typically limiting things to roughly the angle between the speakers. When you properly cancel interaural crosstalk, you let each ear receive much cleaner, more isolated channel information, closer to how headphones deliver it. That naturally shifts localization outward for many sources, because the brain relies on those pure interaural differences (time, level, and spectral cues) to place sounds further apart or more precisely in space.

If someone were born with only one functioning ear (monaural hearing), they'd learn to localize using other cues like head movements and visual references, so their brain adapts. But for most of us who grew up with normal binaural hearing, suddenly getting near-monaural separation per channel through XTC feels like the direction has shifted further away—because the conflicting crosstalk that was anchoring everything inward is gone. It's not artificial widening; it's just removing the distortion that was narrowing the stage all along.

So a well-tuned XTC should indeed widen the soundstage (and deepen it, improve clarity, etc.) compared to plain stereo, but only by revealing what's already in the recording—never more than that. Push it further with over-processing, and you get those artificial effects we talked about earlier.

 

[Gwreck]

BACCH is fundamentally a crosstalk cancellation (XTC) system, just like other XTC solutions out there. All of them work primarily through phase manipulation to cancel the unwanted signal that crosses over from one speaker to the opposite ear.

The thing is, you can push phase manipulation so far that the stereo image gets shifted dramatically—even out to ±90° or beyond. At that point, though, you’re no longer doing proper interaural crosstalk cancellation; you’re basically creating an artificial widening effect instead.

This extreme image shift often happens when the cancellation parameters (delay, attenuation, filter curve) aren’t dialed in correctly for the specific listening setup. It’s a very common reason why people try XTC, don’t hear the expected natural improvement, and conclude it “doesn’t work” or sounds weird.

In a well-calibrated XTC system, hard-panned sounds (full left or full right in the recording) should still appear only a few degrees outside the physical speaker positions—not floating way out wide or behind you. When they do stretch unnaturally far, it’s usually a sign the settings are off.

And just to be clear, this isn’t unique to BACCH. Pretty much every XTC implementation can run into the same issue if the cancellation isn’t properly matched.

My system is well optimized and my statement was about of an exaggeration of the difference between recording types with BACCH. Early stereo recording with hard panned instruments will often place to instruments at 3 o’clock and 9 o’clock or even a bit further back. If one does not like the xtc they can dial down the XTC or shut it off altogether. I am getting over 20db of XTC with my BACCH-sp which is an indicator of optimal settings and speaker type and placement.