Passive MS Encoder/Decoder with bypass switch

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

isophase

Well-known member
Joined
Jul 2, 2007
Messages
240
Location
Lutece
Hi there,
My friend asked me to build him a passive Mid-Side Encoder/Decoder with bypass switch.
We choose to use the lundahl LL1588 as it seemed to be a very good candidate for this application. The attached diagram is based on the BBC sum & difference schematic (only the original attenuator is replaced by XLR connectors).
Wanted to check with you guys if there was anything wrong with the switching and overall wiring i drew.
Thanks in advance :)
Jon
 

Attachments

  • passive MS encoder-decoder.pdf
    624.1 KB · Views: 227
Actually, i'm not sure if i should connect all pin1 together or if it would be better to leave the outputs floating?
Also, i'm not sure if i should make the decoder part as a mirror like on my drawing, or if i should reproduce two exact same "encoder" and use the second encoder as the decoder (since they do the same thing). My concern here is that i'm worried that the crossed wiring on the windings of the M/S input would mess with the output impedance of the inserted gear?
sorry if its not clear what i'm saying, i hope you understand.
 
here is the data sheet for the lundahl LL1588
www.lundahl.se/pdf/1588.pdf
and attached is the original BBC sum and difference schematic for reference.
 

Attachments

  • BBC MS.png
    BBC MS.png
    35.5 KB · Views: 228
oh, i forgot to mention about the switch. As shown on the diagram, i am thinking of using a latching 8 pole switch such as this one sold by audio maintenance. http://www.audiomaintenance.com/acatalog/IT-03-001_extended_info.html
thanks!
 
The mirror thing probably is for better separation and matching.

I guess Jensen are more transparent transformers, on the page is a decoder to M/S mic technique, is also has a posible solution to your grounding question.

http://www.jensentransformers.com/as/as065.pdf

JS
 
isophase said:
Actually, i'm not sure if i should connect all pin1 together or if it would be better to leave the outputs floating?
and attached is the original BBC sum and difference schematic for reference.

Connect each pin 1 to chassis at the XLR. You don't need to do anything else with them.

Cheers

Ian
 
ruffrecords said:
isophase said:
Actually, i'm not sure if i should connect all pin1 together or if it would be better to leave the outputs floating?
and attached is the original BBC sum and difference schematic for reference.

Connect each pin 1 to chassis at the XLR. You don't need to do anything else with them.

Cheers

Ian

If doing this do it as close to the conector as you can, at the same xlr screw for example.

JS
 
Actually, from looking at the schematic again i don't think the "cross wiring" on the M/S input is messing with impedance looking from the output of the inserted gear (i know transformers only reflects impedances) but babe more the crosstalk of the M and S channels could be problematic? i don't know?
Though for sure if i replicate the encoder wiring and use it also as the decoder, the ratio is changing from 1:2 to 2:1 (since the primary is wired in series) but i think is no big deal since the gain is controlled by the line amp of the processing gear anyway and can be compensated.
sorry if its not clear what i'm saying and if i'm saying bullshit. i'm no expert.
will give it a try as it is and see if it works well.
thanks for your advice.
 
Wait a second, let's go for basics...
L+R=M
L-R=S

M+S=L+R+L-R=2L
M-S=L+R-L+R=2R

So, if it would be made of 1:1 transformers level would be doubled at the end of the chain, if you use some pull down in one stage and the same pull up in the other also would be doubled.
So you need to lose 3dB in each stage, both the same, unless transformers loads are contemplated or don't pay atention to 6dB boost in the chain both shouldn't be mirrored and be pull down.

Maybe I'm not making sense and losing some fact, just an observation.

JS
 
joaquins said:
Wait a second, let's go for basics...
L+R=M
L-R=S

M+S=L+R+L-R=2L
M-S=L+R-L+R=2R

So, if it would be made of 1:1 transformers level would be doubled at the end of the chain, if you use some pull down in one stage and the same pull up in the other also would be doubled.
So you need to lose 3dB in each stage, both the same, unless transformers loads are contemplated or don't pay atention to 6dB boost in the chain both shouldn't be mirrored and be pull down.

Maybe I'm not making sense and losing some fact, just an observation.

JS
That's an almost centennial debate. Some advocate losing 3dB because signals combine quadratically (do they?), others advocate losing 6dB because the L & R signals are strongly correlated, so they combine almost arithmetically. No one has given a perfect universal answer. Using 1:1+1 runs the risk of overloading the M channel; what's more, the L & R input impedance becomes 300 ohms when both signals are correlated (the impedance varies from half-nominal to near-infinity in function of the correlation factor). Using 1:0.7+0.7 (-3dB) maintains proper impedance and level for all sorts of signals. IMO 1:0.5+0.5 is overkill. The only advantage is it is easier in terms of manufacturability (quad-filar winding is easy and accurate).
Now regarding the M-S to L-R side, there are disadvantages to the 1+1:1 approach. The L & R signals recovered from a nominal M are at -6dB, which may force the equipment inserted in the M chain to work a little hot.  Using 0.5+0.5:1 makes the reflected impedance too low, which the inserted chain may not like. Again, the impedance of the M & S inputs vary considerably against correlation. In typical cases, where the correlation factor is about 0.8, the impedance of the M input is about nominal and the S input is about 5 times larger.
As you can see, the (apparently) more trivial subjects hide some real complexities. No wonder all major manufacturers have gone active. Any transformer is much more expensive than a handful of IC's, which also offer much more flexibility.
 
abbey road d enfer said:
joaquins said:
Wait a second, let's go for basics...
L+R=M
L-R=S

M+S=L+R+L-R=2L
M-S=L+R-L+R=2R

So, if it would be made of 1:1 transformers level would be doubled at the end of the chain, if you use some pull down in one stage and the same pull up in the other also would be doubled.
So you need to lose 3dB in each stage, both the same, unless transformers loads are contemplated or don't pay atention to 6dB boost in the chain both shouldn't be mirrored and be pull down.

Maybe I'm not making sense and losing some fact, just an observation.

JS
That's an almost centennial debate. Some advocate losing 3dB because signals combine quadratically (do they?), others advocate losing 6dB because the L & R signals are strongly correlated, so they combine almost arithmetically. No one has given a perfect universal answer. Using 1:1+1 runs the risk of overloading the M channel; what's more, the L & R input impedance becomes 300 ohms when both signals are correlated (the impedance varies from half-nominal to near-infinity in function of the correlation factor). Using 1:0.7+0.7 (-3dB) maintains proper impedance and level for all sorts of signals. IMO 1:0.5+0.5 is overkill. The only advantage is it is easier in terms of manufacturability (quad-filar winding is easy and accurate).
Now regarding the M-S to L-R side, there are disadvantages to the 1+1:1 approach. The L & R signals recovered from a nominal M are at -6dB, which may force the equipment inserted in the M chain to work a little hot.  Using 0.5+0.5:1 makes the reflected impedance too low, which the inserted chain may not like. Again, the impedance of the M & S inputs vary considerably against correlation. In typical cases, where the correlation factor is about 0.8, the impedance of the M input is about nominal and the S input is about 5 times larger.
As you can see, the (apparently) more trivial subjects hide some real complexities. No wonder all major manufacturers have gone active. Any transformer is much more expensive than a handful of IC's, which also offer much more flexibility.

But if using 1:0.5+0.5 and then 0.5+0.5:1 we still have 6dB boost at the end of the chain... I'm just talking with nothing in the middle and forgetting about impedance and correlation. From what you said about this I think a good option is to use 1:0.5+0.5 for the encoder and 1+1:1 for decoder, this way we have best headroom for 'almost mono' signals and 6dB loss that make chain gain 1.

0.5L+0.5R=M
0.5L-0.5R=S
M+S=0.5L+0.5R+0.5L-0.5R=L
M-S=0.5L+0.5R-0.5L+0.5R=R

I work with low correlation for my mixes but the biggest signals remain centered like kick and snare what are usually which determine peaks of the track. So in the M/S processing we have this 6dB of headroom for those 6dB higher peaks at M and at the end of the chain we have 0dB gain.

JS
 
joaquins said:
But if using 1:0.5+0.5 and then 0.5+0.5:1 we still have 6dB boost at the end of the chain... I'm just talking with nothing in the middle and forgetting about impedance and correlation. From what you said about this I think a good option is to use 1:0.5+0.5 for the encoder and 1+1:1 for decoder, this way we have best headroom for 'almost mono' signals and 6dB loss that make chain gain 1.

0.5L+0.5R=M
0.5L-0.5R=S
M+S=0.5L+0.5R+0.5L-0.5R=L
M-S=0.5L+0.5R-0.5L+0.5R=R
You must redo your calculations. The 1:0.5+0.5 LR to MS matrix delivers M=1 and S = 0. The 1+1:1 MS to LR matrix submitted to these will output L=R=0.5 so you'll have 6dB loss; that's because the M signal is split between the two primaries. It's not a problem, can usually be dealt with quite easily.
I work with low correlation for my mixes
So do most of SE's doing pop/rock production.
but the biggest signals remain centered like kick and snare what are usually which determine peaks of the track.
Not to mention lead vocals.
 
abbey road d enfer said:
joaquins said:
But if using 1:0.5+0.5 and then 0.5+0.5:1 we still have 6dB boost at the end of the chain... I'm just talking with nothing in the middle and forgetting about impedance and correlation. From what you said about this I think a good option is to use 1:0.5+0.5 for the encoder and 1+1:1 for decoder, this way we have best headroom for 'almost mono' signals and 6dB loss that make chain gain 1.

0.5L+0.5R=M
0.5L-0.5R=S
M+S=0.5L+0.5R+0.5L-0.5R=L
M-S=0.5L+0.5R-0.5L+0.5R=R
You must redo your calculations. The 1:0.5+0.5 LR to MS matrix delivers M=1 and S = 0. The 1+1:1 MS to LR matrix submitted to these will output L=R=0.5 so you'll have 6dB loss; that's because the M signal is split between the two primaries. It's not a problem, can usually be dealt with quite easily.
I don't get what you said here... I want 6dB loss in the chain because encoding and decoding with 1:1+1 will make 6dB boost in L and R at the end of the chain.

abbey road d enfer said:
joaquins said:
but the biggest signals remain centered like kick and snare what are usually which determine peaks of the track.
Not to mention lead vocals.
I also use lead vocals in stereo often, I meant REALLY LOW correlation, lower than usual, 3 channel mix with bass, kick and snare at mid and toms in mid positions and the rest of the band L or R... Lead recorded with 2 mics, one for L other different for R. Same for guitars when just drum bass guitar.

JS
 
joaquins said:
I don't get what you said here... I want 6dB loss in the chain because encoding and decoding with 1:1+1 will make 6dB boost in L and R at the end of the chain.
You wrote: "But if using 1:0.5+0.5 and then 0.5+0.5:1 we still have 6dB boost at the end of the chain..." No. In that case, you end up with 0dB L & R outputs, no boost.  Encoding with 1:1+1 gives 6dB boost of M. Decoding with 1+1:1 gives 6dB attenuation of L & R (for M=0dB) 
I also use lead vocals in stereo often, I meant REALLY LOW correlation
How do you achieve that? Asymetrical position of mics or a lot of ambient sound? I would think the vocals would have a strange or blurred imaging.
 
Let's say that in a moment we have 1V in L and 1V in R, totaly correlated, instant voltage, no AC signal. In M I have L+R so 2V in S L-R so 0V. When I came back and decode I have M+S or M-S so I have 2V in both channels. I have 6dB boost when encoding decoding. I don't know if I have something wrong... If I use a summer with gain 1 then 1V+1V is 2V and 2V+0V is 2V. The same if I have gain of 0.5 in the first stage and gain of 2 in the second.

offtopic:

I use 2 different mics, usually a dynamic and a condenser. To archive a 'central' point I just listen it moving a fader and put it at the center, then I take this difference for the rest of the song... It may vary in some parts of the song because the singer had moved a little closer to one mic or the other, I usually put the dynamic closer, so if the singer goes away the dynamic will be lower than the condenser in comparison.

Those variations don't bother me, a slight change from side to side gives more life to the vocals, the decorrelation takes it closer to the listener in the mix... The only thing to have in mind is the mono version, the vocals will take about 3dB to the back than kick, snare and bass but stay with the level of guitars and such. I don't pay much attention to the mono listener because if it's listening in mono you have one of two cases,
The first, the radio or whatever isn't so good so no hi fidelity nor high nothing there. He is listening in a mono radio with a single 3" speaker to an AM or mono FM or too far away the station, anyway he won't hear a nice sound.

The second option is the listener it self doesn't bother in plug it stereo and is just listening in mono with a super fancy stereo, in this case he could be listening only L or R or in the best case Mono. If he doesn't work to listen it right I won't make that job. A friend in his music shop has a JBL510 with only L (or R, I don't know), what can I do!

All this consideration gives me more freedom to play with stereo, even with canceled things to have that effect or with stereo to vocals, I take care that don't cancell too much, I correct the delay by hand and phase inverting if needed, no more than that. It works for me and in the mastering studio I usually take my mixes like it.

JS
 
joaquins said:
Let's say that in a moment we have 1V in L and 1V in R, totaly correlated, instant voltage, no AC signal. In M I have L+R so 2V in S L-R so 0V. When I came back and decode I have M+S or M-S so I have 2V in both channels. I have 6dB boost when encoding decoding. I don't know if I have something wrong...
Yes, you have something wrong. When you inject 2V in the 1+1:1 MS to LR matrix, you get only 1V in each output. No 6dB boost. That's because the M voltage is split equally between both primaries.
offtopic:
I use 2 different mics, usually a dynamic and a condenser. To archive a 'central' point I just listen it moving a fader and put it at the center, then I take this difference for the rest of the song... It may vary in some parts of the song because the singer had moved a little closer to one mic or the other, I usually put the dynamic closer, so if the singer goes away the dynamic will be lower than the condenser in comparison.
  Are both microphones in front of the singer or is one to the left and the other to the right? Do you pan them L & R in the mix?
Those variations don't bother me, a slight change from side to side gives more life to the vocals,
Most of my customers would disagree...
the decorrelation takes it closer to the listener in the mix...
I don't get that... Decorrelation increases the ratio of "ambient" to direct sound, which usually results in more width and depth. Similar to adding reverb.
The only thing to have in mind is the mono version, the vocals will take about 3dB to the back than kick, snare and bass but stay with the level of guitars and such. I don't pay much attention to the mono listener because if it's listening in mono you have one of two cases,
The first, the radio or whatever isn't so good so no hi fidelity nor high nothing there. He is listening in a mono radio with a single 3" speaker to an AM or mono FM or too far away the station, anyway he won't hear a nice sound.

The second option is the listener it self doesn't bother in plug it stereo and is just listening in mono with a super fancy stereo, in this case he could be listening only L or R or in the best case Mono. If he doesn't work to listen it right I won't make that job. A friend in his music shop has a JBL510 with only L (or R, I don't know), what can I do!
Indeed, the constraints that were dictated by vinyl discs and mono radio are a thing of the past. I've seen a report saying that the most used listening environments are: first car, second earbuds. These are not really stereo, thet are binaural or dual-mono, very little or no acoustic combination in both cases. If you look at kids in the subway, they share music by sharing earbuds, one listening to L, the other to R, and both having a free ear for chattering and blabbering.
All this consideration gives me more freedom to play with stereo, even with canceled things to have that effect or with stereo to vocals, I take care that don't cancell too much, I correct the delay by hand and phase inverting if needed, no more than that. It works for me and in the mastering studio I usually take my mixes like it.
Well, I'm pleased that you found an unconventional and creative mixing paradigm.
 
abbey road d enfer said:
joaquins said:
Let's say that in a moment we have 1V in L and 1V in R, totaly correlated, instant voltage, no AC signal. In M I have L+R so 2V in S L-R so 0V. When I came back and decode I have M+S or M-S so I have 2V in both channels. I have 6dB boost when encoding decoding. I don't know if I have something wrong...
Yes, you have something wrong. When you inject 2V in the 1+1:1 MS to LR matrix, you get only 1V in each output. No 6dB boost. That's because the M voltage is split equally between both primaries.
Thanks... I always saw it wrong and something don't like me!
offtopic:
I use 2 different mics, usually a dynamic and a condenser. To archive a 'central' point I just listen it moving a fader and put it at the center, then I take this difference for the rest of the song... It may vary in some parts of the song because the singer had moved a little closer to one mic or the other, I usually put the dynamic closer, so if the singer goes away the dynamic will be lower than the condenser in comparison.
  Are both microphones in front of the singer or is one to the left and the other to the right? Do you pan them L & R in the mix?
Both in front. Usually dynamic closer, the other option is both toghether. Dynamic closer because of its properties. And yes, one to L and the other to R.
Those variations don't bother me, a slight change from side to side gives more life to the vocals,
Most of my customers would disagree...
That's another issue, give it a shot and see what they say... The other option is to make some dynamic processing to have always in the center... I can't tell now how but would be an option.
the decorrelation takes it closer to the listener in the mix...
I don't get that... Decorrelation increases the ratio of "ambient" to direct sound, which usually results in more width and depth. Similar to adding reverb.
As 3D movies or real life sight you see something is closer to you because you see 2 diferent images with both eyes and because you see it sharper.

When you listen to something the same thing is happening, if you have a person speaking close to you you will hear 2 diferent things in both ears, when he's away the same signal came to both ears, more bright will give you 'sharpener' image... The other thing to have in mind is if ITDG (initial time delay gap) or pre-delay of the reverb is longer the sound seems to be away from the walls and close to the listener. If ITDG is shorter the sound seems to be closer to walls and away to the listener.

Indeed, the constraints that were dictated by vinyl discs and mono radio are a thing of the past. I've seen a report saying that the most used listening environments are: first car, second earbuds. These are not really stereo, thet are binaural or dual-mono, very little or no acoustic combination in both cases. If you look at kids in the subway, they share music by sharing earbuds, one listening to L, the other to R, and both having a free ear for chattering and blabbering.
As dual mono this work fine, it's just a mic for each channel, both dynamic and condenser should be good mics and both have nice sound on its own.
As I said before if someone in on the subway listening with one earbud he won't be listening nice, so if he doesn't bother of what he is listening me neither. In the car not so much but it isn't the best listening condition either.

Mix will work fine if it's 2 diferent mono mixes, let's say one with dynamics for vocals, guitars, etc. plus the kick and snare (OH are stere and toms are panned nothing to do here...) and the other one the same but with condensers. Just an example, if I use lead vocals dynamic to left lead guitarr go with dynamic to right to compensate the mix colour. If I have 2 second vocals the same, etc...

I think that it's better let who want to listen it in the best condition enjoy the best than try that who don't mind listen the better he can.

Well, I'm pleased that you found an unconventional and creative mixing paradigm.
Thanks for that!

JS
 
Hi there,
Long time no post (i became a happy father in the meantime :)
Just to say, I did finish the passive mid-side encoder/decoder for my friend using x4 LL1588 lundahls and it works really good !
In the end I opted for a 1:0.5-0.5 ratio for LR to MS conversion and 0.5-0.5:1 ratio for MS to LR conversion. Doing the job just right AND sounding very good.
The lundahls are very transparent indeed, i have not done any measurements as i don't have the tools for that but i do hear a little something that is not at all unpleasant to the sound that is, listening to both the encoder and decoder at the same time with no gear inserted (straight wire).
thanks again to those who contributed and shared.
photo attached.
jon
 

Attachments

  • pmsm.jpg
    pmsm.jpg
    971.6 KB · Views: 164
Back
Top