Revisiting stereo shuffling

[fresponse]

Hi all,

 

I have revamped my Soundstage shaper plugin by extending the concept of stereo shuffling and making it applicable to headphones listening.
The objective is to better align all frequencies of sound sources sharing the same position in the virtual soundstage.
As we all have a different hearing, the plugin will come with a companion configuration tool to perform optional individualization.

Some more technical background is available here

 

I am now looking at some beta testers to evaluate the plugin&configuration tool combo.
The main requirements to apply are 
- use a player accepting external plugins (JRiver, Audirvana, Foobar for instance)
- being a regular headphones user

 

If interested, please contact me at [email protected]

 

Thanks,
Thierry 

[Bill Brown]

Interesting!  I have read a lot about this, including:

 

https://www.diyaudio.com/forums/multi-way/277519-fixing-stereo-phantom-center.html

 

I have played with it a bit on my system (not headphones) with mide-side EQ.  For hard-panned instruments it adds body (tenor sax), though I have decreased the high-frequency attention in the side channels a bit since this picture to restore a bit of "presence."  Don't note the slight increase in side level for increased stereo effect- not terribly accurate but enjoyingly euphonic:

 

 

[fresponse]

Yes, the idea is to go beyond this type of mid / side correction. I mentioned headphones because the calibration process is easier, but of course my tool is applicable to speakers as well.

[Bill Brown]

11 minutes ago, fresponse said:

Yes, the idea is to go beyond this type of mid / side correction. I mentioned headphones because the calibration process is easier, but of course my tool is applicable to speakers as well.

 

Definitely- I am sure your process is significantly more advanced and would go far beyond my little experiments.  I primarily wanted to point out that there are merits to the concept and I am excited to see you working on it.

 

Bill

[Confused]

On 3/25/2021 at 5:45 PM, Bill Brown said:

Interesting!  I have read a lot about this, including:

 

https://www.diyaudio.com/forums/multi-way/277519-fixing-stereo-phantom-center.html

 

I have played with it a bit on my system (not headphones) with mide-side EQ.  For hard-panned instruments it adds body (tenor sax), though I have decreased the high-frequency attention in the side channels a bit since this picture to restore a bit of "presence."  Don't note the slight increase in side level for increased stereo effect- not terribly accurate but enjoyingly euphonic:

 

 

 

There are some interesting comments in your diyaudio link Bill.

 

The OP reminded me of something I noticed a while ago, and that has baffled me since. 

 

With a Devialet amplifier you can adjust volume, tone controls etc. from the remote.  The remote has a big dial, which is normally used simply for volume control.  The remote has a "tone" button, this scrolls through a menu of things you can adjust, so it runs through balance, bass, treble, (all then adjustable with the volume dial), then on to other options.  

 

Here is the interesting point.  There was one time I was listening to a track, I pressed the "tone" button on the remote, which on first click takes you to the "balance" adjustment.  When the "balance" adjustment is selected, the amp has this trick that it switches to mono operation, just for the duration of the balance adjustment, I guess the idea is that the balance adjustment is more accurate or easier to do with a mono signal.  The interesting thing I noticed was that the very act of switching to mono, without any other adjustments, resulted in a much darker sound.  This is entirely repeatable with a wide range of music, hit the "tone" button, the amp switches to mono, and in many cases a much darker sound results.  This has always baffled me, and I had wondered if it was maybe some kind of psychoacoustic effect resulting from first reflection points or similar, but looking at the attached I really am not sure.  Also, this is not one of these subjective things that might be imagined, with some music the effect is quite pronounced.

 

Is there an examination for what causes this, and how exactly this effect manifests itself?

 

@fresponse - My apologies if this is a little off topic, but at least it gives the thread a bump!

[fresponse]

@Confused One explanation for what you observe is comb filtering due to crosstalk when you play the same signal with 2 loudspeakers. It is explained in my video on acoustic crosstalk at 4:30 here

[Bill Brown]

Thank you for posting @Confused, I have been nervous that my first post put a damper on this thread, when what I really intended to convey was my excitement and enthusiasm.  I haven't ventured into any room correction yet; I want to maximally optimize my acoustics first, then venture in. These days I am trying to avoid "audiophile nervosa" and simply enjoy the music, not be too "fiddly." When I do give it a shot I am going to look hard at what @fresponse is doing, he seems to have deep knowledge of these issues and seems cutting edge. 

 

Re. a balance control, I have always been plagued by a bit of "balance shift" (would like to blame never-fully symmetric acoustics, but it is always to the left.....this implies it is my hearing, but thankfully it remains good for a man in his 50s- early 50s! :) ) and was frustrated by a feeling of trying to mentally pull and "rebalance" the soundstage.  A tiny adjustment of balance in my DAC is nice.  The Devialet approach is cool- center the mono signal, done.

 

Re. mono, with recordings designed for stereo you are missing the "side signals" that produce the ambience you are used to and experiencing comb filtering (the latter one of the benefits of the shuffler re. tonal balance across the soundstage that this plugin (I think!) addresses).  At one of my crazy moments (not uncommon!) I pondered moving away from spatial effects and designing a single, dynamic speaker (mono of course) to get back to the simplicity and "tone" of music.  I tried doing some experiments in dsp with the side signals, but couldn't really get it right so abandoned the idea.

 

Bill

[Bill Brown]

Oh! @fresponseposted while I was typing!  Off to check out his video and hope I was correct :)

[Bill Brown]

Great video @fresponse, though I am now being drawn to the same rabbit hole I have ventured down many times :).

 

My first experiences with IAC reduction were way back with Carver's processor (can't remember the name) and Polk's SDA speakers in the 80s.  I actually heard their new speakers with this technology at the last RMAF.  Pretty impressive for a single listener.

 

Occasionally I have been limited to headphone listening.  That of course lead to many experiments with crossfeed.  There are tons of plugins for this now, some with simple crossfeed, others with additional measures to make the bass more like that perceived with speakers in a room.  Some say you can eliminate the "in the head" soundstage with headphones and move it out in front of you.  I never could achieve that.  As implied in the video, I think it requires the individual's true HRTF. I only occasionally use it now for older recordings with hard pan-potting (The Beatles a classic example).  With these recordings they are at the ear, with a resultant "pressure" phenomena that is improved with crossfeed.

 

http://recherche.ircam.fr/equipes/salles/listen/index.html

 

With naturally miced recordings I don't like it.  With those I can "get inside" the natural acoustic.  It can be wonderful.

 

Isn't it interesting that with headphones we add crossfeed and now with speakers try to decrease/eliminate it?  Both probably have merit.

 

You mentioned a center speaker.  Another of my rabbit holes!  I have researched ad infinitum how to add one/generate a center channel in dsp. I would use an identical center channel as my L/R.  I think it is a very, very worthwhile approach.  There are many plugins that are working on this (usually in the context of surround channels also, though with some there is the possibility of starting with 3.x).  Examples:

 

https://www.blueripplesound.com/products/o3a-upmixers

https://nugenaudio.com/haloupmix

https://www.pro-tools-expert.com/home-page/2018/6/1/shootout-results-the-best-upmix-plug-ins-which-is-your-favourite

 

Many of the towering figures from the history of audio engineering have worked on these issues- Blumlein, Gerzon, Blauert, Griesinger, Fosgate, etc.  Wanting to use legacy two-channel recordings I have researched lots of them- Triefield and Ambisonics from Gerzon, stereo shuffling, Logic 7, etc.  I haven't been impressed by Dolby Digital's approach in this regard.  BTW I am discussing ambience extraction rather than generation.  I am very intrigued by Trifield but it seems only available in Meridian's processors.  I sought out the original equations thinking of trying to do it in dsp- the simplified flow diagram is fairly straightforward, but the underlying math.....suspect it is in @fresponse's locker, but not mine.

 

 

 

Seems I am a habitual thread-hijacker @fresponse, but I am excited about this and ask your forgiveness (and I am going to reference this thread in the diyaudio thread).  And at least the addition of a center channel is an approach to the problems your plug-in is addressing.  If you could add a center channel based on Trifield (or your preferred approach) to your stable of impressive offerings (especially IAC cancellation and room correction) I would be the first in line! Pretty please! Alas, I suspect the appeal is limited; most in our hobby remain wedded to two-channel, feeling it sufficient and resisting the added complexity.

 

I know you are familiar with all of this @fresponse, but will add some links if anyone else wants to do some research:

 

http://www.audiosignal.co.uk/Gerzon archive.html

 

https://www.sfxmachine.com/docs/FrequencyDomainUpmix.pdf

 

https://www.researchgate.net/project/Stereo-to-multi-channel-upmixing

 

A nice background including the simplified flow diagram I referred to with good additional links:

 

https://www.stereophile.com/reference/407upward/

 

Bill

[fresponse]

Hi  @Bill Brown

 

It is not yet in my plans to work on center channel addition, as there are already alternatives to do that.

As intensity panned recordings are quite different from spaced mics recordings, it makes sense to sometimes prefer crossfeeding for headphones or crosstalk reduction for loudspeakers... and all this is subject to individual subjective preferences...

 

The first feedbacks on the new plugin are very positive, and I will soon close the testing phase : this is the last call ;-)

 

Thanks,

Thierry