Center channel extraction from stereo

[wonderfulaudio]

Hi,
My existing system consists of TV whose line level audio out L/R is connected to active-amplified bookshelves, while the system is fine, I do need more clarity of dialogues.
I want to build and add an active-amplified center channel. I found this interesting article http://elias.altervista.org/html/SingleSpeakerStereo.html
It claims success in extracting center channel and provides measurement. The essence is below

I would appreciate if somebody could convert this to op-amp based circuit that takes L/R inputs and provides L/C/R outputs with a pot for Rps.
Thanks and Warm Regards,
WonderfulAudio

 

[Heinrich]

Hi,
My existing system consists of TV whose line level audio out L/R is connected to active-amplified bookshelves, while the system is fine, I do need more clarity of dialogues.
I want to build and add an active-amplified center channel. I found this interesting article http://elias.altervista.org/html/SingleSpeakerStereo.html
It claims success in extracting center channel and provides measurement. The essence is below
View attachment 485649
I would appreciate if somebody could convert this to op-amp based circuit that takes L/R inputs and provides L/C/R outputs with a pot for Rps.
Thanks and Warm Regards,
WonderfulAudio

I once read a complicated article on that topic. It was driven by experimentation. Alas, can‘t remember its headline. My conclusion, or the take-away was, substract like 20% left from right and vice versa, add these two 20% to mono and send it to the center. No frequency dependency. Place the stereo speakers at 45 degrees to the side instead of the commonly suggested 30 degrees. When implementing such using opamps a fine grained level adjustment is easy anyway, giving room to endless optimization.
If you do something around those lines, please report back!

 

[DVDdoug]

Summing Amplifier

FYI- Mixers are built-around summing amplifiers. The inverting version is most common but since you don't want to invert (so it will be in-phase with left & right) you may want to put an inverting buffer inline with the output to invert it back.

Have you used op-amps before? You generally need bipolar plus and minus power supplies (+/-12 or +/-15V is common) and the power should have 0.1uf "bypass" capacitors on the power supplies near the chip. And you might want to ad "DC blocking" capacitors to the inputs and outputs (although there shouldn't be any DC in the signal anyway).

My conclusion, or the take-away was, subtract like 20% left from right and vice versa, add these two 20% to mono and send it to the center.

Its probably better if you make it adjustable. If your active speakers have volume/gain controls you can adjust the relative volumes that way.

Or you could give it a some gain (maybe 2X) and put a pot between the summing amp and the re-inverting stage. In a mixer that pot would be the master level control. (I don't recommend using pot in the feedback circuit for adjustable gain because it the pot gets noisy (or goes open) you can suddenly get very-high or infinite gain.)

It can also be done passively with 2 resistors (I'd recommend 10K resistors). Or you can buy a Passive Stereo to Mono Adapter.

These circuits assume an unbalanced signal (RCA connections).

Note that this is NOT "center extraction". It's simply mixing-summing stereo to mono. Center extraction can be done digitally with some "advanced processing" and it may create unwanted side-effects.

 

[Speedskater]

Two more not so east solutions:

a 5.1 channel processor/pre-amp/receiver (with line level outputs). Then send the Center channel to both speakers. But I don't know of any inexpensive units.

a Main & Sides circuit. Again, I don't know of any inexpensive units. But there are DIY boards.

 

[wonderfulaudio]

@DVDdoug , The original article I posted does something more than just summing, it provides a shelving high pass to L/R and shelving low pass to C and I think that makes a lot of difference to get a center channel and yet maintain the imaging of mains.

@Speedskater , 5.1 processors are overkill, they generate surround channels which I dont need, they cant be avoided because some of the 'energy' from mains is diverted to surrounds, if the processor is used then surround channels also must be used. Same with the sub channel which I dont need.

I am not looking for multichannel solution, if only somebody could convert the technique to op-amp active, it would help.

 

[wonderfulaudio]

This is my inspiration

 

[marin.weigel]

You might want to check out the pdf in the last post of this thread on forum.allaboutcircuits.com.
Haven't ever tried it, but it likely might get you what you're looking for.
Good luck!

 

[DVDdoug]

I posted does something more than just summing, it provides a shelving high pass to L/R and shelving low pass to C

Yeah, you can find op-amp filter circuits online. The classic Active Filer Cookbook is available online. But you might want to experiment with equalizers before building hardware. It could be VERY time consuming experimenting with different filter designs and it's more difficult to A/B different filters than moving a knob on an equalizer.

If I wanted to do this, I'd probably just use an equalizer, and another equalizer if you want to EQ all 3 channels.

And different recordings have different amounts and different "styles" of stereo panning so you'll get different results with different recordings.

and I think that makes a lot of difference to get a center channel and yet maintain the imaging of mains.

You can filter-out the highs or lows, etc. But if there's a guitar on the left it will still go to the center, just with different-altered tone.

It WILL alter the stereo image and you may like it or not. You aren't listening "as intended", but the important thing is that you enjoy it!

I am not looking for multichannel solution,

Well... It's not regular 2-channel stereo or true 3-channel either. (I've never heard of a commercial 3-channel recording but I know it's been done experimentally with good results.) Once multi-channel started becoming practical for home, use we got surround sound instead of just 3-channels... Actually there was quad (4-channel stereo in the 1970s before it was practical!

Of course, multi-channel recordings include a true-discrete center. And if you have a surround recording you can down-mix to 3-channels with an AVR.

What you are doing is an effect. Intentionally altering the sound like with EQ* or reverb, or up-mixing /down/mixing with an AVR.

5.1 processors are overkill, they generate surround channels which I dont need,

OK. But you you don't have to use all 5 (or more) channels, and it's easy to play with different settings on an AVR. Personally, with regular stereo music I like to use a "hall" or "theater" setting on my AVR for some delayed reverb in the rear and the "feel" of a larger space. This is hi-fi heresy since I'm not listening accurately as intended.

Same with the sub channel which I dont need.

With stereo you don't need a sub as long as you have full-range speakers that can reproduce bass.

But with surround you lose the "point one" LFE channel without a separate sub. The LFE isn't normally included in the down-mix so you only get the "regular bass".


* EQ can also be corrective to correct for "imperfect" speakers or rooms, or bad recordings.

P.S.
You might look into the miniDSP. I don't have one but it can do all kinds of filtering, EQ, delay, and I assume it can do summing and subtraction and other fun stuff.

 

[EERecordist]

It could be a fun project. I strongly believe it will not solve your problem.

Film/movie mixing is part of the post production project and a lot of money is spent on it. When they mix, they make a stereo mix with the dialog panned to centerish, where the actor is in the screen. And they make higher order mixes, such as 5.1.1 with a center channel in the 5. They put the dialog on the center channel of the 5 for exactly the audibility you seek, because of the psychoacoustics of our coupled visual system with hearing. You might be able to use AI demix plugins on a DAW to partially separate out the dialog from stereo, but imperfectly.

I think you may get better results in dialog audibility from headphones or spatial audio earbuds. Consider boosting some voice frequencies with equalization.

 

[JaapD]

Let’s for a moment turn back to the OP’s problem, being “while the system is fine, I do need more clarity of dialogues”. So it is about clarity.

I fully recognize this from my TV set as well, but fail to see why adding a center channel would solve this. In other words, you might be a bit disappointed after adding a center channel.

Shouldn’t we need to find the solution w.r.t. clarity in a different direction, i.e. in tone control, by raising the mids?

Update: I do think you'll need an equalizer to solve your problem. See following attachment ('borrowed' from the Interweb) for some quidance in the settings.

 

[Heinrich]

Let’s for a moment turn back to the OP’s problem, being “while the system is fine, I do need more clarity of dialogues”. So it is about clarity.

I fully recognize this from my TV set as well, but fail to see why adding a center channel would solve this. In other words, you might be a bit disappointed after adding a center channel.

Shouldn’t we need to find the solution w.r.t. clarity in a different direction, i.e. in tone control, by raising the mids?

The center phantom is notoriously delicate. To replace it by a true center might be an option. Only experimentation will tell. Trustworthy report is sparse, but must not be substituted by theorizing.

That said, for starters a software solution might be in order. Equalizer APO (windows) provides a mixer, that may be suitable. Or a plain resistor ladder network exploiting the symmetric outputs of a pro-usb interface.

 

[tifune]

@Speedskater , 5.1 processors are overkill, they generate surround channels which I dont need, they cant be avoided because some of the 'energy' from mains is diverted to surrounds, if the processor is used then surround channels also must be used. Same with the sub channel which I dont need.

I'm not sure where you're getting this idea. With any AVR, you just set the unused channels to off and no "energy" is diverted anywhere. Routing LFE to front is optional IIRC

If you're doing this for fun, by all means build to your heart's content. But the easiest/fastest answer here is an AVR. If vocal clarity is the issue, a used AVR with Neural:X is cheap and easy to find for only L/C/R usage.

 

[olds1959special]

The joy of my life

 

[wonderfulaudio]

With any AVR, you just set the unused channels to off and no "energy" is diverted anywhere.

My understanding is that:When TV is giving just 2 channel stereo then AVR cannot enable center channel, it needs 5.1 or 7.1 to be supplied by TV, encoded

 

[tifune]

My understanding is that:When TV is giving just 2 channel stereo then AVR cannot enable center channel, it needs 5.1 or 7.1 to be supplied by TV, encoded

Yes and no - an AVR can synthesize a center channel using an upmixer. If you do not want this feature, you just turn off the up mixer.

 

[RPoitou]

Understanding TV audio connection options

Hi. I am trying to buy a new TV (small 43") and I need some help in understanding the connectivity options. I am mainly interested in good quality 2-channel audio for music, and am hoping to reuse my current hifi amp/speakers to improve my TV sound. The TV sound is very much a secondary...

www.audiosciencereview.com

 

[Fred H]

Consider boosting some voice frequencies with equalization.

We have difficulties with unfamiliar accents. Following this link we boost 2KHz, Q=0.26 (4 octaves bandwidth) by 1 or 2 dB, generally 1dB is enough for intelligibility.

Excerpt from the above link:
The important frequencies in non-tonal (Western) languages are illustrated by the diagram below. Here, the frequency band around 2 kHz is the most important frequency range regarding perceived intelligibility. Most consonants are found in this frequency band.

(Ref: N.R. French & J.C. Steinberg: Factors governing the intelligibility of speech sounds. JASA vol. 19, No 1, 1947).

 

[moonlight rainbow dream]

The used market is flooded with old, super cheap AVRs that can do exactly what you need. You can use optical connection from your TV (possible fidelity upgrade), feed your active speakers with preamp outs, and make use of Dolby/DTS upmixer to get to 3.0. There's no need to reinvent the wheel.

Dialogue intelligibility is rarely single-factor issue though... so you should also consider the native performance of your speakers, their placement, and room acoustic treatments.

 

[olds1959special]

Center fill channel.