Mid-side Preamp with DOA's

I have been scetching out a few possible designs for a DOA based preamp for mid-side tracking. I have read the other posts which all seem to have the 3 signals at equal levels and non-adjustable. My need came about after trying to track M-S guitar overdubs in my DAW(PT). The problem is that I need to flip the phase and bus route the signal and there is enough latency to be noticeable.

That said, I am trying to map out the best way to deal with the Side signals that I need to split. It seems that Y cable to 2 preamps would be OK for a makeshift solution, but I want to do this properly wtihout extra loading on XFMRS.

I am basing this on a simple 312 type circuit

my rough ideas are the following:

I need 2 outputs with one signal having opposite polarity.

1. XFMR to noninverting opamp and inverting opamp:
-This may be the best way but I am not sure if it would load the input XFMR too much(although the secondary will be terminated with 150K).



2. XFMR splitter XFMR secondary + to Left opamp and XFMR secondary - to Right opamp -both opamps are non-inverting.

-The problem here is that to use 75101APC type I yould need to have 2 transformers on the input. eg the 75101 feeding the 1:1:1

3. XFMR> Opamp split to L out and inverting opamp for R
-I would rather not have caps in the path or use a servo bias circuit.


These are the basic building blocks -a bunch of other compenents are not shown obviously.

Hi,

Just wondering why to combine the functions of pre-amping & matrixing? Making them separate blocks will make them more universally usable.
Like for instance a M/S-matrix without preamps has its own applications as well, especially when you have two.

But without doubt you've thought about that yourself already and decided you just wanted to have them combined, so then please skip my suggestion :wink:

You'll have seen the uneeda-circuits, right ?
http://www.uneeda-audio.com/#ms
http://www.uneeda-audio.com/mat03.pdf
The #3-circuit has it all, just put a few of your favourite pre-amps in front of it and you're done.

I use my mixers channel direct out at line level with a flipped insert cord which I bring back at line level in another channel. Not sure if you have a mixer but I think it accomplishes the end effect in good fashion.Just a thought :? K2

Just a thougt, there is really no plugin in protools that does ms-decoding?? I am on Nuendo (PC) and always use a freeware from Voxengo for this... If you had a plug, which you insert into a stereo track, then the possible very few samples delay shouldn't matter.
There are so many other - to me seemingly more important - things to build like comps, eqs and pres, spending time on an ms matrix seems a bit like a waste of time, considering that lifetime is limited - unless you absolutely want to have this analog matrix :wink:

Absolutely no offense intended, I am just surprised.

Michael

If you are going a 990/transformer route, why would not you use a splitted secondaries trafo?

One of the big advantages of MS recording is automatic mono compatibility because of the M channel. For tracking with MS recording, I ignore the S channel on the headphone playback since all the required musical information is in the M channel.

However, if you want an S-splitter for MS decoding in real-time (no plug-in latencies), you don't need to worry about inverting and non-inverting opamps if you have transformers. Simply buffer the signal and use two output transformers with their primaries in series or parallel (depending on the turns ratios), and then wire their secondaries in phase and out of phase respectively to your two mix channels.

M/S in pro-tools.

Easy.

VERY easy.

'M' panned center.

'S' panned left. Set up an aux send from the 'S' channel to, let's say... bus 1 for example. Set the send level to 0dB (alt-click on the send fader to set it to zero).

Shift-control-N to Create one mono aux return, pan it right, set the input to be bus 1, set the fader to 0dB (alt-click the fader) Now everything on the 'S' channel will sound like it's mono.

Insert a single band EQ on the aux return, set flat, but invert the polarity. Set the solo safe if you like.

Now the 'S' signal will sound "out-of-phase". Flip the polarity on the inserted single-band EQ to hear the polarity go in and out.

Voila! -Instant M/S decode.

I used it htis very weekend to do a live on-the-fly M/S monitor decode from a Soundfield ST250 recording B-format. 'W' and 'X' up the center (With 'W' polarity-flipped) to make a front-facing cardioid, then Y (which is a side-facing figure-8, just as in M/S) gets this "aux-return-polarity-flipped" trick, to give a very nice M/S decode on the fly.

The 'M' channel (or channels, in the case of the soundfield... the beauty of the soundfield in this application is that by balancing the two W and X channels in Pro-Tools, you can remotely change the center 'M' pattern, cardioid, super-cardioid, hyper-cardioid, sub-cardioid, figure-8, omni or anywhere in between... ) controls the mono signal, the 'S' channel (called 'Y' in Soundfield) controls the width of the image. You can also EQ the 'S' channel to boost or rein-in the width at specific frequency areas.

You leave the return fader at zero, even if you increase or reduce the 'S' fader. It's a post-fade aux return, so the 'S' channel fader does both left and right parts of the 'Side' signal simultaneously.

Easy.

Before anyone screams "latency" at me... there isn't any. none. ZERO.

Proof? just put up the 'S' channel and the aux return at 0dB. Now, pan them both dead center. -Silence. This proves that they are perfectly matched in time, with ZERO latency.

Keith

yeah ... nah

each channel has equal latency

even on non compensated systems
if you use the same plugs on each channel there should be the same plug delay and so no differencial latency issues

Thanks, I'll check that out Kieth. I think I bussed the S to 2 auxes and flipped polarity on the -S using trim pluggin. For some reason I thought that auxes had a small amount of latency, When ADC is disabled.

Michael, thanks for bringing me to my senses. I read some post about a splitter loading mics wrongly and thought of using opamps. Well, if I did have latency on my system I would use a cinemag 600:600:600 and have a polarity switch for the -S.

Now I need one of those Soundfields -time to start saving or selling stuff.
The soundfield has front-8, side-8 and coincident-8's?


BTW I am really loving M-S on acoustic guitar.
Anyway, I still have plans for a DOA based PCB for your typical preamp setup. few, I almost fell into a DIY black hole!

[quote author="guavatone"]The soundfield has front-8, side-8 and coincident-8's?[/quote]
the SoundField has 4 channels: W, X, Y & Z.

W = Omni.
X = Front-back figure-8 (Front is + polarity)
Y= Left-Right figure-8 (left is + polarity)
Z = Up-Down figure-8 (Up is + polarity)

Be aware that you can generate ANY in-between angle of figure-eight with two perpendicularly arranged (i.e. at 90° to each other) figure-of eight patterns. All you do is add them in proportion to the desired angle's sin and cosin.

So for a 45°-left facing figure-8, you just add X (front-back) and Y (Left-right) in equal amounts. Do it at -3dB to retain the same level as the original X and Y components. Now you can 'tilt' this up and down by adding in the Z axis... Again, the sin/cosin law keeps the level constant with the originals.

Then you can generater ANY standard pattern (facing in ANY direction, as demonstrated) from the generated figure-8 and the 'W' (omni).

Remember that an omni and a figure-8 will produce a cardioid if added equally, facing in the direction of the figure-8's positive lobe. Vary the proportion and you have anything from omni through sub-cardioid, cardioid, super-cardioid, hyper-cardioid to figure-8.

You can do this as many times as you want, and generate 2, 4, 5, 8, 16, 32 or however many channels you would like. -Played back over speakers placed at the matching positions, (called "periphonic" playback) the soundfield is unreal. Start adding overhead speakers ("voice of God") and things like helicopter takeoffs recorded with a Soundfield become stunningly real...

All with 4 channels.

It's basically like M/S where M/S has one controllable dimension, whereas B-format has three controllable dimensions.

Keith

Just came across this link and remembered this thread.
So FWIW:
http://emusician.com/tutorials/emusic_front_center/
Also note the related downloads.

[quote author="clintrubber"]Just came across this link and remembered this thread.
So FWIW:
http://emusician.com/tutorials/emusic_front_center/[/quote]

The cool thing about that article is that if you go to the web clips, they have PT sessions to get started.

[quote author="guavatone"][quote author="clintrubber"]Just came across this link and remembered this thread.
So FWIW:
http://emusician.com/tutorials/emusic_front_center/[/quote]

The cool thing about that article is that if you go to the web clips, they have PT sessions to get started.[/quote]
No ! :evil:
Some PT2Cub or PT2CEP-converter out there ? :roll:

But I figure it's just convenience, right ? As in tracks already ready to go, like in other programs only taking a minute from scratch ?

[quote author="clintrubber"]
No ! :evil:
Some PT2Cub or PT2CEP-converter out there ? :roll:

[/quote]

Huh? :?

[quote author="guavatone"][quote author="clintrubber"]
No ! :evil:
Some PT2Cub or PT2CEP-converter out there ? :roll:

[/quote]

Huh? :?[/quote]
As in a conversion-script-something to translate PT-songfiles to stuff readable by other software.
But it'll be fine I assume. Haven't looked at the soundfiles yet, but I assume it's just a matter of putting them on a few tracks in ones fav seq-prog.

Bye,

Peter

Clint, don't bother looking for those converters. Kieth stated everything u need to know really. The Emag versions is actually way over complicated.

I like to send all signals to an aux and compress on mixdown. Looks like I am going to be doing most of my mixing in the box for now. My clients need more bang for their buck. ...i ramlbe.....

I checked 5-6 samles of latency for an insert on an aux channel Kieth. Not the kind of thing people can hear really. I think it's in the nanosecond range.

Truth be told, this delay was heard by my client since I souldn't hear the source, I would have no idea if there was latency. It started with the compressor accross the mix fader, and when that was removed my cleint claimed to hear latency, which was most likely the reverb from altiverb.
-Funny how peoples perceptions aren't very changeable once they hear something they expect to still be hearing it and their mind tricks them to think they are hearing it, especially if they are in a bad mood.

I think half of my job is wranling with issues of perceptions. My favoriete is the bass player who things they need more pressence to come through the mix. (brewer material)