BACCH4Mac "Absolute Sounds Product of the Year 2024"

[jimbill]

I have Sonus Faber Olympica Nova III. Thinking about ORC, but I have Anthem ARC already and it isn't cheep.

View attachment 368005

Regarding XTC, I'm not sure what you're expecting so:
View attachment 368006

My problem with finding a track is that usually I'm just playing a shuffled playlist of 2800 tracks. I went through some tracks I know to have the aforementioned behavior. Couldn't hear the issue in Led Zeppelin - Whole lotta love. Maybe I'm imagining to myself that I can hear this a little bit in East17 - Deep (No need to hear the whole song just the first second). Maybe it is just my imagination. Maybe when I have time i'll try to create a test track for this. Anyone knows a nice free app for that?
If we are talking about memory. I was sure I had a noticeable reflection in the Impulse Response.

You had mentioned that your speakers weren't overly directional, so that is why I wondered about your XTC readings. Those are really good numbers, so I don't think you're missing anything.

I also have Anthem that came with my Martin Logan speakers. The ORC gives me a much flatter FR than you are getting. I don't know if it's enough to justify the $1,000 investment. I have an open area house, so a very large listening space and no room treatment. I've considered putting something behind my speakers, but Edgar didn't think it would make much of a difference. And the WAF has been stretched pretty much to the limit.

 

[STC]

Maybe I'm imagining to myself that I can hear this a little bit in East17 - Deep (No need to hear the whole song just the first second). Maybe it is just my imagination.

You may not be imaging at all. During my Stereo days, I used QSound tracks and also experienced that the stage seemed inconsistent. 3D effects as opposed to crosstalk cancellation tend to do that.With XTC, you shouldn’t be having the illusion collapsing after a brief moment. Will try out the tracks you suggested.

….but it is so hard to be objected. After years with XTC, I am also wondering whether I am beginning not to see errors as expectations could be clouding my judgment. : ))

 

[Dialectic]

I have Sonus Faber Olympica Nova III. Thinking about ORC, but I have Anthem ARC already and it isn't cheep.

View attachment 368005

Regarding XTC, I'm not sure what you're expecting so:
View attachment 368006

My problem with finding a track is that usually I'm just playing a shuffled playlist of 2800 tracks. I went through some tracks I know to have the aforementioned behavior. Couldn't hear the issue in Led Zeppelin - Whole lotta love. Maybe I'm imagining to myself that I can hear this a little bit in East17 - Deep (No need to hear the whole song just the first second). Maybe it is just my imagination. Maybe when I have time i'll try to create a test track for this. Anyone knows a nice free app for that?
If we are talking about memory. I was sure I had a noticeable reflection in the Impulse Response.

Are we sure those simulated measurements are correct? I've experimented a lot with BACCH filter creation, and I have never been able to get XTC that high, especially with conventional speakers and in rooms with a window on one side (even when covered by drapes). Those measurements look very similar to those with one of the filters that comes pre-installed with BACCH.

If the measurements are correct, congratulations, and we need to get some tips from you on making BACCH filters...

 

[tree]

Are we sure those simulated measurements are correct? I've experimented a lot with BACCH filter creation, and I have never been able to get XTC that high, especially with conventional speakers and in rooms with a window on one side (even when covered by drapes). Those measurements look very similar to those with one of the filters that comes pre-installed with BACCH.

If the measurements are correct, congratulations, and we need to get some tips from you on making BACCH filters...

I'll re-measure in the near future. I remember having a reflection point in the IR filter (added a carpet since then). I also need to re-measure because of the camera position going out of sync anyway. Maybe it was reset somehow when I updated the software. The avg XTC should be lower or Higher for better result?
If I remember correctly Edgar said he liked my correlation in measurement between the speakers but the reflection will be determent to getting better result.

You had mentioned that your speakers weren't overly directional, so that is why I wondered about your XTC readings. Those are really good numbers, so I don't think you're missing anything.

I also have Anthem that came with my Martin Logan speakers. The ORC gives me a much flatter FR than you are getting. I don't know if it's enough to justify the $1,000 investment. I have an open area house, so a very large listening space and no room treatment. I've considered putting something behind my speakers, but Edgar didn't think it would make much of a difference. And the WAF has been stretched pretty much to the limit.

I limit the ARC to 2000Hz as far as I remember.

 

[Smaestro]

Wow! I didn't know Toole had already tried to explain this back in 2018, if I had known about this old thread I would have linked to it instead of trying to explain it myself in recent forum threads.

Toole is of course right, most audio productions are not made with crosstalk canceling in mind so in most cases it will only act as an applied "sound effect" that sometimes can sound convincingly impressive, but more often than not, it will just be heard as a pretty obvious post-production sound effect. It's completely okay if some people like what BACCH does to the sound, but it will only be a "more correct" reproduction for audio productions specially made for crosstalk cancellation.



Exactly as Toole explains it in this text:

"You describe this as: "Therefore what people are reporting about when they listen to an acoustic recording played through a BACCH filter has nothing to do with phantom speakers or location of the actual speakers, but everything to do with the more correct reproduction of the sources whose spatial cues were captured in the recording." Because we are not talking about an encode-decode process, one cannot say that the reproduction is "more correct". It is what it is, and what it is may be very attractive to many listeners, but it cannot be "more correct". Stereo mixing for the bulk of music is multitrack, using isolation booths, or at the very least voices and instruments individually miked. These components are amplitude panned to various locations across the soundstage as the mixer/sound designer chooses. Spatial effects are often electronically generated - artificial - having nothing to do with real acoustical spaces. Classical recordings are a very mixed bag of methods, ranging from a coincident Blumlein pair to a pair of widely spaced omnis, to several mics placed over and within the orchestra, and others farther out in the hall. Again, these are combined in the sound design. All of these are created to sound as desired by the mixer/producer/mastering engineer while listening to a pair of loudspeakers in a small room. So, because the mix did not anticipate crosstalk-cancelled reproduction, what one hears through such a system is not "more correct" but instead the result of spatial post processing of a particular kind."

Edit: This post assumes wrongly how the BACCH works and can be ignored.
Apologies if this has been answered already, I am still plowing throught this thread.

But has anyone already described if crosstalk cancellation is possible?

How can BACCH determine what MIC crosstalk possibly is and isn't?

In anything not recorded with two mics, then mixed hard panned, by definition has no MIC crosstalk, and this no crosstalk cancellation possible. That is 99% of all recordings. Common for small choir recordings maybe?

The way I see BACCH can work, is by removing delayed copies of stereo channel information.

But the common Decca tree mixing setup for classical has 5 mics.. What is cross talk (delayed stereo info) and what is a delayed signal from the mics further away?

So if BACCH works as described in article it also removes early reflections. In that case BACCH is not a decrossfeeder but a De-reverber (for only early reflections)

 

[Smaestro]

Edit2: the content of this post is irrelevant to the BACCH.

Also, I will add, that if the recording with 2 mics was done with a Blumlein pair (with proper figure 8 mics), or XY, or Mid-Side technique, then there will be zero or near zero crosstalk. Even with a NOS or ORTF setup crosstalk will be minimal.

So what is this crosstalk cancellation about? The cases where music was performed and recorded in front of an AB pair of cardiode or omni mics?

I have no problem with tools making sound pleasurable, so I don't mind if it sounds nice. I just can't wrap my head around its intended way of operation. What does it do other than:

1. Cutting mids (mid channel not mid tonality) for that removal of Right sound in the Left ear
and
2. Removing early reflections in the oppositie channel (basically inverse Haas delay), to (doubtful imo) remove MIC crosstalk (the LmodM)?

I'm sure im missing something because if it is really as good as claimed, more is probably happening. (Although a good dereverber is impressive).

Dr. Edgar explains the workings a bit in this post, and it is indeed about the first form of cross talk (between the source at the mics) ILD and ITD in two mic setups.
And they also add extra reverb after removing 'crosstalk' reflections.
He doesn't describe the second form of cross talk (between the speakers and ears), which I thought would be the more interesting part.

https://audiosciencereview.com/foru...acch4mac-pro-edition-a-report.2373/post-75471

I think this is fine. Nothing wrong with good widening DSP. It's just that the cases where it can actually reduce crosstalk is pretty limited, in other cases it synthesizes wideness.

 

[Dialectic]

Also, I will add, that if the recording with 2 mics was done with a Blumlein pair (with proper figure 8 mics), or XY, or Mid-Side technique, then there will be zero or near zero crosstalk. Even with a NOS or ORTF setup crosstalk will be minimal.

So what is this crosstalk cancellation about? The cases where music was performed and recorded in front of an AB pair of cardiode or omni mics?

I have no problem with tools making sound pleasurable, so I don't mind if it sounds nice. I just can't wrap my head around its intended way of operation. What does it do other than:

1. Cutting mids (mid channel not mid tonality) for that removal of Right sound in the Left ear
and
2. Removing early reflections in the oppositie channel (basically inverse Haas delay), to (doubtful imo) remove MIC crosstalk (the LmodM)?

I'm sure im missing something because if it is really as good as claimed, more is probably happening. (Although a good dereverber is impressive).

Edit:
Dr. Edgar explains the workings a bit in this post, and it is indeed about the first form of cross talk (between the source at the mics) ILD and ITD in two mic setups.

And they also add extra reverb after removing 'crosstalk' reflections.

He doesn't describe the second form of cross talk (between the speakers and ears), which I thought would be the more interesting part.

https://audiosciencereview.com/foru...acch4mac-pro-edition-a-report.2373/post-75471

I think this is fine. Nothing wrong with good widening DSP. It's just that the cases where it can actually reduce crosstalk is pretty limited, in other cases it synthesizes wideness.

A lot of confusion here. BACCH filters are generated from sweeps played through loudspeakers. They are intended to remove crosstalk from stereo speakers to unmask recordings' subtler interaural timing and level cues, which are lost in conventional stereo playback.

 

[Smaestro]

A lot of confusion here. BACCH filters are generated from sweeps played through loudspeakers. They are intended to remove crosstalk from stereo speakers to unmask recordings' subtler interaural timing and level cues, which are lost in conventional stereo playback.

Yes... oh man... I misread something in the Absolute Sound explanation it sent me down the wrong rabbit hole...

I read that they wanted to remove the "R+modM" from the Left ear, and I assumed part of the way would do that is by figuring out what modM is and subtract that from the signal. Whereas the approach is it to remove the right channel sound from the left ear entirely.

Seems like it'll result in a combination of the good wideness and detail from headphones with the good of speakers.
I'll have to get a demo.

 

[Dialectic]

Firsthand technical information regarding BACCH and related technologies is available here:

Papers

3d3a.princeton.edu

 

[Smaestro]

Amazing, thanks!

 

[tree]

Are we sure those simulated measurements are correct? I've experimented a lot with BACCH filter creation, and I have never been able to get XTC that high, especially with conventional speakers and in rooms with a window on one side (even when covered by drapes). Those measurements look very similar to those with one of the filters that comes pre-installed with BACCH.

If the measurements are correct, congratulations, and we need to get some tips from you on making BACCH filters...

Didn't have time yet to re-measure, but I Imported existing configuration and now I have the old reflection points. I assumed that when I'm upgrading the software the data is migrated also and with fine changes to the algorithm the plot changes with this migration.

Does someone knows what XTC level means?

 

[Dialectic]

Didn't have time yet to re-measure, but I Imported existing configuration and now I have the old reflection points. I assumed that when I'm upgrading the software the data is migrated also and with fine changes to the algorithm the plot changes with this migration.
View attachment 368556
Does someone knows what XTC level means?

The very high XTC filter that you posted is a default one that comes preloaded in BACCH4Mac. XTC levels above 15 dB are consistently achievable only with highly directional speakers (i.e., electrostats and horns) in large or treated rooms.

The measurements above are more typical of what can be accomplished with conventional speakers in a typical domestic room. From experience, I know that it's a filter that could be made using the setup in an earlier photo you posted.

You don't want to use the default filter; the high predicted XTC levels won't be attained in your room.

I've had some earlier versions of BACCH4Mac occasionally revert to that default filter after upgrades. I always save my workspace before an upgrade, unless it's a major upgrade, in which case new filters need to be generated.