FYI I have just released a plugin to implement crosstalk reduction (as alternative to the u-Bacch solution)
You can find download link on my plugins page (X-talk Shaper)
This is terrific! Thanks.
I’ve just sent you my ID to obtain a key.
It differs from the uBACH filter in that it doesn’t exhibit the problems I’ve reported here which uBACCH has over UPnP. It just works over UPnP with Real-Time Processing turned off in Audirvana
Another difference is the angle setting appears to be the total angle measured at the apex where the listener sits to both speakers, whereas with uBACCH the angle is half that amount as measured from the listener to each speaker.
Apart from that it doesn’t have the ability to generate a ball of pink noise to determine the best angle to each speaker; but that’s easily measured with a tape measure or laser measure. Your plug-in also has a couple of features uBACCH lacks, namely, “Interaural distance”, “effect” (amount of), “min phase” and “volume” (to avoid clipping).
Your manual explains all these clearly. I haven’t had a chance to compare directly with uBACCH yet; but it sounds as though it’s doing its job, without any obvious (to me) adverse effects on tonality.
Good job!
1 Like
I’m totally agree with you.
I’ve been using the crosstalk cancellation solution with measurements in combination with active Room correction by Home Audio Fidelity for a few years now and it works very well.
I’ve made some changes in my listening room, and I don’t really need active room correction now.
I’ve tried X-talk Shapper with success and I’m agree with you, soundgals.
It works without problem with Audirvana.
For windows it not only offers VST3 plugin but VST plugin too (dll format) for players which don’t support VST3 plugins. There is a standalone solution too.
And the most important, it is cheaper than u-BACCH by far.
Compared to U-BACCH, I’ve just noticed a higher latency which prevent me to use it for video and introduce gapless on some players (JRiver Media Center or Audirvana are not concerned).
Is the application of X-shaper an effect or a solution to a problem?
The majority of stereo productions have been mixed and mastered using speakers as the monitor reference, so cross-talk is inherently imbued in the contextual left-channel + right-channel harmonic, dynamic and spatial relationships of the production, as an essential element of an artistic decision… Listening to a production that has been mixed and mastered using speakers on a set of headphones skews the contextual left-channel + right-channel harmonic, dynamic and spatial relationships of the production by nature of the isolation inherent in the use of headphones (without HRTF DSP)… Why would I want to alter this inherent cross-talk information on playback over a set of speakers, just for the sake of altering it, if not just for an entertaining presentation? (I use HRTF DSP on my headphone playback system)
Are you identifying the spectral information/energy presented in the listening scenario acoustics that are interfering with the contextually fundamental spectral information/energy imbued in the stereo production?
The way I see it, Stereo is an illusion that tries to convince us that we are there, or they are here at the time the music was made.
The Stereo illusion has its limitations. Whether it was mixed and mastered using studio monitors (likely very different to your own speakers), the idea of it all is to be able to reproduce the sound of live musicians in an acoustic space. So the mixing engineers should be trying to get as close to that as possible.
I don’t believe this is either an effect or a solution to a problem as such, other than just trying to get beyond the limitations of Stereo. If it creates a more convincing illusion of the musicians being in my room (or me being in their’s), I’m all for it. I don’t think it makes sense to be purist about something which is an illusion in the first place.
Obviously there are other attempts to get beyond the limitations of Stereo. This one happens to only require two speakers!
Thanks for raising the question and for your valid statement about mixing and production of multi tracks recordings. I see crosstalk reduction as a solution for following reasons
- there are still some recordings where the spatial cues are not coming from pan potting at the mixing desk but directly from a pair or mics
- mixing is usually done in nearfield conditions where ILD (interaural level difference) is higher than in usual consumer listening conditions
- the comb filtering produced by stereo for centered vocals (or instruments) is really a pain (cf. center speaker for home cinema) and any solution to alleviate it is welcome
A two-microphone recording or a Decca-tree configuration entails levels of cross-talk in the creative decisions that produce the stereo image. Binaural recordings depend on the cross-talk factor of head related shadowing (HRTF) in the alignment of the microphone pair or multi-microphone array, including torso resonance and diffraction.
Generally, stereo mastering (and some mixes, especially mixes done for movie/theater productions that are mixed and mastered in actual theater/studios) are done midfield/farfield, so to facilitate better translation into any given playback in the home playback, acoustic-space, scenario…
It seems to me that an analysis of the early-reflection spectra of the listening space is integral to maintaining the inherent contextual synergies of the left-channel + right-channel signal spectra imbued in the mix of any given recording, and would this not require a continuously dynamic feed-back mechanism in real-time, to remove/negate the intrusive reflection energies? It would seem that a continuous real-time comparison of the inherently recorded left-channel + right-channel phase-related spectra to the listening environment early reflection spectra is required… No?
That statement is a bit too obtuse… Would we call all art-forms ‘illusions’ in this way?
Stereo production and playback is a tangible product of creative decisions, that ploy on the neurophysiology of the hearing cognitive processes… Once the wavefronts from musicians, vocalists, etc, are translated into electrical impulses, these elements become the ‘paint’ from which the engineers, producers, musician(s) and mastering engineer(s) work into a tangible audio experience.
In the context of auditory experiences… a stereo production is no different from the experience of listening to the wind blow through the trees in a mountain forest… the tangible difference, is the real-time, real-world experience of the moment… Is this an illusion?
Is a Salvador Dali painting an illusion or is it a tangible representation of an illusion for effect… MC Escher drew illusions, but are the physical drawings illusions? Is the sound of clocks ticking, and the cacophony of the clock-chimes in Pink Floyd’s piece, “Time” an illusion or an element of contextual artistic design, that produces a tangible effect in the production of the album “Dark Side Of The Moon” ?
As far as the “playback” part is concerned that product will be profoundly altered by the playback chain, to the listeners liking, unlike the Salividor Dali or Escher art you mentioned. These filters are just another element in that playback chain. If they improve the experience for the listener he/she will want to use them.
I agree with this perspective… if it is subjectively entertaining, then many will find this effect enjoyable. Nothing wrong with this… 
However… We can apply the effect of the museum or gallery environment as influential in the interpretation of the Dali or Escher…
That is also true. In the gallery environment though, I believe we can have pretty broad agreement as to what degree the lighting and choice of where the art is hung, allows the viewer to appreciate the work.
Go to any audio show though, and the same piece of music played through different systems will sound quite different, and some of those systems will reveal more of what was happening at the original event than others.
To rule this out we’d all have to be listening to standard issue stereo systems, just like in the old USSR, and in standard rooms; but then how could we know those systems; etc. were the best possible?
The uBACCH, is now joined by the X-talk Shaper filter, and both have a fairly subtle effect on most recordings. On standard stereo recordings the sound stage seems to extend in width, and each of the various voices and instruments tend to have a more three dimensional portrayal, and project further into the room.
On some recordings you can get a wrap-around effect, as though you had rear speakers with discrete elements coming from them. This is more likely to happen on binaural recordings intended for headphone listening. That effect is indeed “entertaining” and I believe is probably what the artists intended.
So enhanced or “super stereo”, if you like, with standard stereo recordings, plus the added bonus of fully appreciating binaural recordings over speakers.
I liken the effect to having a giant set of ear-speakers with minimal cross-talk… Like plain vanilla headphone listening, some recordings will sound decent, even though the isolation of the L-ch and R-ch energy is subtracting some contextual information.
I believe we all probably draw a line somewhere when it comes to applying DSP filters. For some that may be none at all, for others, it depends.
Personally I try to consider what the original intentions of the artists were. Recently I posted about the Quark plug-in, which can decode QS encoded stereo files to full Quadraphonic. I like the idea of a kind of secret code hidden in stereo files that can be decoded to (unfold?) the audio to something beyond the plain stereo signal.
I’ve listened to some apparently plain stereo albums from Qobuz in full Quadraphonic this way. That was because I’m fairly sure they still contain the QS encoding. I have to judge by ear if I’m actually getting discrete signals from the rear channels.
I don’t like the idea of simulated surround sound. So I would never knowingly run audio through any kind of decoder to create pseudo surround.
For example, I’m tempted by the Binauralizer plug-in for further converting these decoded QS signals to Binaural stereo for listening on headphones. My thinking there is that the artists intended a spatial representation and that’s why they recorded in Quad. To listen to the Quad audio processed to spatial binaural listening might prove to be a satisfying experience when listening in Stereo on headphones…
… or even when listening on speakers by further processing with one of these cross-talk cancellation plug-ins, when only a stereo pair of speakers are available.
Quad SQ never really made it… It was surpassed by discrete 4-channel 8-track tape (Had this in my van) and later with Pro Logic and Dolby Digital and DTS, 5.1 SACD (I have them all)… Today we have psychoacoustic systems like the Yamaha Virtual Cinema Front that can produce perceived virtual acoustic spaces and reproduce 7.1 Dolby Atmos productions via a frontline array of five (5) speakers (L, LCF, C, RCF, R) + Subwoofer with great accuracy in a proper listening space, with reasonably reflective surfaces… Listening to 5.1 SACD and DVD-A is very compelling when the the array is properly aligned (I have them all)… However, today I am headphone-centric… The future of audio playback is in the virtual space for all but the most esoteric playback scenarios.
MEMS speaker technology under DSP control is where we are headed…
https://xmems.com
Thanks for the link, interesting stuff.
No disputing Quad (SQ, QS and CD-4) didn’t make it commercially. I still find it fun to play these old quad recordings, hidden in stereo streams, back with the convenience of streaming services such as Qobuz.
Is the problem originating from the plug-in (Bacch labs) or AudirvAna?
This is taking almost 3 months…
I wish I knew the answer to this. It could be a bit of both. I know the BACCH guys opened a ticket on it some time ago; but have not heard anything from them, and nothing here from the Audirvana side.
My suspicion is that it’s quite a demanding filter to run in real-time in a VST supporting player, over UPnP. I had a similar problem trying to run it over UPnP within J River.
Also, the situation is a little better since recent updates to Audirvana when running on my Mac Studio. I still get the silence at the beginning of the first track. After that all is fine for the rest of the tracks from the same album. When playing a playlist with mixed sample rates though, I get the silence on each new track that involves a sample rate change.
On a less powerful machine, such as my M1 Mac Air, I get distorted sound following the silence.
1 Like
Hi @Nvk,
The Bacch guys came back to me last week and wanted to know more on the way non-real time mode vs real time mode works on how those modes manage the audio in our app. We will come back to them today or tomorrow on this. For the time being, it is an issue on their side but we will see this with the answer we will give them.
2 Likes
It seems that music sample rate change cause issues in many systems… A quick google search is very eye opening… I imagine it is exaggerated across a UPnP network connection…
I’ve never had any problems with UPnP and sample rate changes before.
I regularly use about 5 different UPnP control points and one DLNA server (J River) and never had problems when playing playlists involving frequent sample rate changes.
I’ve only come across the problem with this particular plug-in.