line level processing deriving a center channel from stereo channels

Thread Starter

PeteHL

Joined Dec 17, 2014
340
The project that I'm posting here today is an analog line level audio signal processor that derives a center channel signal from the left and right channel signals of conventional stereo recording. The derived center channel signal is to be reproduced by a loudspeaker system positioned in the front of a listening area in the middle between the conventional left and right channel loudspeaker systems.

This project is somewhat more complex than I might usually do and requiring nineteen dual op amps, TL072. For that reason, three of the figures that are attached to this post, Figures 2, 3 and 5, show sub-circuits of the processor. Attached Figure 6 shows interconnection of the sub-circuits.

Also attached is five pages of text explaining the design and the improvement provided compared to a "derived" center channel consisting of the summation of left and right channel signals.

Attached Figures 1 and 4 are included for explaining the methodology. A parts list and photo of the completed project constructed on socket board are also attached.

Regards,
Pete
 

Attachments

Thread Starter

PeteHL

Joined Dec 17, 2014
340
In listening tests with this processor, I found that the cross-feeding scheme of Fig. 5 substantially reduced separation of the reproduction of left and right -hand signals. To fix that, resistance of resistors R41 and R51 was reduced from 33k Ohm to 15k Ohm. This results in the "ideal" voltage division of the attenuating and subtracting circuit of Fig. 4.

Then so that attenuation of the common signal in the range of 400 Hz to 1600 Hz at least comes close to equal to -3 dB, I found it necessary to widen the frequency range of subtracting to 100 Hz to 6400 Hz. This required the following value changes in the circuit of Fig. 5.

R35 = 56k
R41 = 15k
R43 = 56k
C15 = 2n
C18 = 2n
R46 = 56k
R51 = 15k
R53 = 56k
C19 = 2n
C22 = 2n

Attached is a simulation with LTspice showing attenuation of the common signal VS frequency given the above changes to filtering cut-off frequencies and voltage division of the cross-fed signal.

Then finally I found that for some recordings that the center channel was (relative to the level of the left and right channels) too loud. Including a L-pad connected to the center channel loudspeaker allows for reducing the level of center reproduction when necessary.

Regards,
Pete
 

Attachments

I think it odd to 1/2-wave rectify the audio and then sum both halves in the "logic adder" as you call it.
It can't see any advantage to an ordinary single op-amp mixer performing L+R=C just as the ear receives with sound waves as air for a mixer.
 

Audioguru again

Joined Oct 21, 2019
2,055
I agree that a "center sound" has the same signal in both stereo channels. Simply add the channels together with two resistors and feed the center channel amplifier. but then a signal only on one channel will still be produced at -6dB in the center channel.
 

Thread Starter

PeteHL

Joined Dec 17, 2014
340
I think it odd to 1/2-wave rectify the audio and then sum both halves in the "logic adder" as you call it.
It can't see any advantage to an ordinary single op-amp mixer performing L+R=C just as the ear receives with sound waves as air for a mixer.
The advantage is the sum of the center signal is +6 dB with respect to the level of the summation of left and right signals. The ordinary summing (straight addition) of the left and right channels results in the center signal +3 dB with respect to the level of the summation of the left and right (uncorrelated) signals.

Admittedly the increase of only 3 dB is requiring a lot of circuitry. Except for my impression in listening to the 3 channel system with this processing, I would think that all of the processing would not be very much worthwhile.

Regards,
Pete
 

Thread Starter

PeteHL

Joined Dec 17, 2014
340
I agree that a "center sound" has the same signal in both stereo channels. Simply add the channels together with two resistors and feed the center channel amplifier. but then a signal only on one channel will still be produced at -6dB in the center channel.
What you are saying I think is that with respect to the level of one or the other of the left and right signals, then just adding the left and right channels results in the center sound at +6 dB. If you make the level of one or the other of the signals the basis, then the logic adder causes the center sound to be at +9 dB relative to the level of the single (left or right) signal

In a simulation with LTspice, output level of the logic adder followed by a low pass filter causes a decrease of level at output of the filter when the relative phase of two input sine waves is changed from in phase to 90 degrees out of phase. In phase is correlated, 90 deg. out of phase is uncorrelated. If you like I will attach the simulation circuit to a future post so that you can see for yourself.

Regards,
Pete
 

Audioguru again

Joined Oct 21, 2019
2,055
The center channel is usually used for vocals. If you cutoff the important high frequency consonant sounds then you cannot understand what is said since the remaining sounds are only vowels.
 

ci139

Joined Jul 11, 2016
1,696
MEDIUM = (LEFT + RIGHT)/2
STEREO DIFFERENCE = ( RIGHT - LEFT ) / 2 -- use "-()" if you want to use the RIGHT as the "REFERENCE"
basic simple , then
RIGHT = MEDIUM + STEREO DIFFERENCE
LEFT = MEDIUM - STEREO DIFFERENCE

but it (the MEDIUM) would have a "flat sound" or "distant sound" effect

however you might want simply to add delayed LEFT to not delayed RIGHT
while the frequencies - inverse proportional to that delay - get amplified . . .

so some combination of both the above may sound better
 

Thread Starter

PeteHL

Joined Dec 17, 2014
340
The center channel is usually used for vocals. If you cutoff the important high frequency consonant sounds then you cannot understand what is said since the remaining sounds are only vowels.
Attenuation by cross-feeding of the middle signal in the left and right channels drops out for frequency > 1600 Hz . So the left and right channel loudspeakers reproduce the middle signal frequencies above 1600 Hz without any attenuating by the processor.

Filtering of the processor is designed to operate something like the cross-over network of a three-way loudspeaker system. The left and right channel loudspeaker systems reproduce the middle signal without attenuation exclusively at frequencies below 400 Hz and above 1600 Hz. The middle signal fed to the center channel loudspeaker system is band-pass limited to 400 Hz to 1600 Hz.

-Pete
 

Thread Starter

PeteHL

Joined Dec 17, 2014
340
MEDIUM = (LEFT + RIGHT)/2
STEREO DIFFERENCE = ( RIGHT - LEFT ) / 2 -- use "-()" if you want to use the RIGHT as the "REFERENCE"
basic simple , then
RIGHT = MEDIUM + STEREO DIFFERENCE
LEFT = MEDIUM - STEREO DIFFERENCE

but it (the MEDIUM) would have a "flat sound" or "distant sound" effect

however you might want simply to add delayed LEFT to not delayed RIGHT
while the frequencies - inverse proportional to that delay - get amplified . . .

so some combination of both the above may sound better
Come again? Sorry, but I don't understand how the points that you are making apply to the problem that my processor seeks to solve.

The first part of your post seems to be making the point that you can derive the original stereo channels by either adding or subtracting the difference of the channels from the sum. How does this apply to separating out the medium signal from the left and right signals? By "MEDIUM", do you mean the center or middle signal?

What would be the purpose of adding one channel delayed to the other channel not delayed?

Some combination of both of the two manipulations may sound better in what way?

Regards,
Pete
 

ci139

Joined Jul 11, 2016
1,696
i guess i misinterpreted where you gonna use your "Center channel"
? why you need the center channel in a first place ?? like a boom box only higher cut ← if you move off the normal to your center spk the vectors wont merge correct

is there any source article that describes in detail the effect you are trying to gain ?

you don't like to hear it but making all your effects adjustable by depth - would provide a greater flexibility - but then doing it in PC programmatically would have a benefit over many pot's in circuit that may add noise (and making your already complex circuit even more cryptic) - https://www.google.com/search?q=sound+effect+programming+library+cpp

the number , of op amps , alone does not describe anything - you should only worry abut it when there might be a way producing your channels with the lesser one
 
Last edited:

crutschow

Joined Mar 14, 2008
25,645
It would seem all you need to do is sum the right and left channels with an op amp with a variable gain at its output to adjust the center channel to the desired volume.
Why do you need more than that?
 

Thread Starter

PeteHL

Joined Dec 17, 2014
340
It would seem all you need to do is sum the right and left channels with an op amp with a variable gain at its output to adjust the center channel to the desired volume.
Why do you need more than that?
Doing what you say does not attenuate the left and right signals in the center channel. Ideally the level of the left and right signals in the center channel would be very much attenuated to the point of not being audible. The upmixer of my processor (generating the center channel) results in the center signal (signal common to both stereo channels) at a level of +6 dB with respect to the level of the addition of the left and right signals. So the upmixer produces an increase of 3 dB of level difference compared to simply adding the left and right channel signals.
 

Thread Starter

PeteHL

Joined Dec 17, 2014
340
i guess i misinterpreted where you gonna use your "Center channel"
? why you need the center channel in a first place ?? like a boom box only higher cut ← if you move off the normal to your center spk the vectors wont merge correct

is there any source article that describes in detail the effect you are trying to gain ?

you don't like to hear it but making all your effects adjustable by depth - would provide a greater flexibility - but then doing it in PC programmatically would have a benefit over many pot's in circuit that may add noise (and making your already complex circuit even more cryptic) - https://www.google.com/search?q=sound+effect+programming+library+cpp

the number , of op amps , alone does not describe anything - you should only worry abut it when there might be a way producing your channels with the lesser one
Conventional stereo reproduction produces what is called a "phantom center image". In a stereo recording, what is meant for you to hear as located directly in front of you is in phase and at equal level in the left and right channels. So this gives an indistnct quality to the center sound as this is not the way that people usually determine where a sound is coming from. That is, the sound is coming from the left and right channel speakers obliquely to the sides of the listener whereas it should be arriving to the listener from a point directly in front.

Floyd Toole in his book Sound Reproduction discusses the phantom image. Searching "stereo phantom center image" on the web would produce many results.

My processing results in the center signal increased in level by 3 dB in the generated center channel while at the same time the center signal is attenuated by -3 dB in the modified left and right channels, both in the range of 400 Hz to 1600 Hz. This is sufficient to cause the listener to distinctly hear reproduction in the middle.

Regards,
Pete
 

Thread Starter

PeteHL

Joined Dec 17, 2014
340
Okay.
But a 3dB change in volume is just barely audible and doesn't seem worth the effort.
Taken at the output of each of the two sections of logic addition followed by a low-pass filter, the level of the summation of the middle signal in the two channels is +6 dB relative to the level of the summation of the left and right signals. Voltage gains and losses of the upmixer are configured so that the level of the middle signal at the output of the upmixer is +3 dB with respect to input level. Thus the level of the summation of the left and right signals becomes -3 dB. From this it follows that the level of the left signal in the middle channel is -6 dB with respect to the level of the left signal in the left channel. The same applies to the right signal.

Cross-feeding of the processor attenuates the middle signal in the left and right channels each by -3 dB. Thus the level of the middle signal in the generated middle channel is +6 dB relative to its level in the left or right channel. This is the same level difference that results when deriving a middle channel by addition of the left and right channels.

The above occurs over the frequency range of 400 Hz to 1600 Hz (two octaves). So the main advantage of the upmixer versus the method of simply adding the left and right channels is the attenuation of the left and right signals in the middle channel. Adding left and right channels results in no attenuation of the left and right signals in the derived middle channel.

Granted that even -6 dB is not a lot of attenuation but it seems to work.

-Pete
 
Top