Albini electronics test

[Jayro_Rockola]

Every one knows that Steve is a crotchedy opinionated bastard (in a good way) who many people think makes really cool sounding records. What does he think is necessary knowledge to call yourself an engineer? Evidently a fairly advanced electronics knowledge.

How many do you know?

Knowledge of electronic circuits to an extent that will allow selection of appropriate signal paths. This means more than knowing the difference between a delay line and an equalizer. Which has more headroom, a discrete class A microphone preamp with transformer output or a differential circuit built with monolithics? Where is the best place in an unbalanced line to attenuate the signal? If you short the cold leg of a differential input to ground, what happens to the signal level? Which gain control device has the least distortion, a VCA, a printed plastic pot, a photoresistor or a wire-wound stepped attenuator? Will putting an unbalanced line on a half-normalled jack unbalance the normal signal path? Will a transformer splitter load the input to a device parallel to it? Which will have less RF noise, a shielded unbalanced line or a balanced line with floated shield?

I must say I consider myself a pretty accomplished engineer and am embarrased to say I have complete understanding of none of these. I guess now is the time for edification.

Anyone want to run these down for us?

[joelpatterson]

And here's the bonus question:

Which foot do you put your LEFT SHOE on?

[Jayro_Rockola]

I'm actually interested in the answers to these questions!

[Frost]

I dropped out of EE school and that was 10 years ago, but here is my best answers... DO not take as fact and lemme know where Im wrong, I like learning too:

1.Which has more headroom, a discrete class A microphone preamp with transformer output or a differential circuit built with monolithics?

Generally, the class A, but it depends on which monolithics vs which transformer as well as the available voltage reserves to each. There is no 'all things being equal' as a class A circuit uses a lot more power.

2.Where is the best place in an unbalanced line to attenuate the signal?

at the beginning if you are attenuating to prevent clipping, at the end if you are just making it quieter. A hotter signal is better at longer runs as long as it isnt clipping.

3. If you short the cold leg of a differential input to ground, what happens to the signal level?

You lose 6dB in theory.

4. Which gain control device has the least distortion, a VCA, a printed plastic pot, a photoresistor or a wire-wound stepped attenuator?

Assuming high quality resistors, the stepped attenuator will be the highest quality.

5. Will putting an unbalanced line on a half-normalled jack unbalance the normal signal path?

Yes, I think so, but I might have to think about this some more.

6. Will a transformer splitter load the input to a device parallel to it?

Not if its the correct transformer.

7. Which will have less RF noise, a shielded unbalanced line or a balanced line with floated shield?

Depends on the type of noise present, but generally, I think balancing will have a greater effect than the shield. Floating a shield isnt necessarily a bad thing.



I hope I passed this quiz. If not, Im interested in where I failed.

Thanks to Mr Albini

Frost

[cornboy7x]

Originally Posted by Frost

I dropped out of EE school and that was 10 years ago, but here is my best answers... DO not take as fact and lemme know where Im wrong, I like learning too:

1.Which has more headroom, a discrete class A microphone preamp with transformer output or a differential circuit built with monolithics?

Generally, the class A, but it depends on which monolithics vs which transformer as well as the available voltage reserves to each. There is no 'all things being equal' as a class A circuit uses a lot more power.

2.Where is the best place in an unbalanced line to attenuate the signal?

at the beginning if you are attenuating to prevent clipping, at the end if you are just making it quieter. A hotter signal is better at longer runs as long as it isnt clipping.

3. If you short the cold leg of a differential input to ground, what happens to the signal level?

You lose 6dB in theory.

4. Which gain control device has the least distortion, a VCA, a printed plastic pot, a photoresistor or a wire-wound stepped attenuator?

Assuming high quality resistors, the stepped attenuator will be the highest quality.

5. Will putting an unbalanced line on a half-normalled jack unbalance the normal signal path?

Yes, I think so, but I might have to think about this some more.

6. Will a transformer splitter load the input to a device parallel to it?

Not if its the correct transformer.

7. Which will have less RF noise, a shielded unbalanced line or a balanced line with floated shield?

Depends on the type of noise present, but generally, I think balancing will have a greater effect than the shield. Floating a shield isnt necessarily a bad thing.



I hope I passed this quiz. If not, Im interested in where I failed.

Thanks to Mr Albini

Frost

the answer to number 1 is who gives a shit. transistors and transformers rule.
question 3, i thought you gain 6db? just dealt with this problem...i should know this. but i guess im just being an asshole
5. depends on how its wired. if cold gets tied to ground then yes. if cold is not connected then no.
7. if you have a balanced line with a floated shield, wouldnt that mean no shield? i could understand floating at one end, but this assumes floating both ends. which would most likely cause (not cause but recieve...you get my point) more noise i would think?

i hope this went better than calculus one at my first college...

[hectc7]

you introdunce an unbalanced line to attenuate?? how does that work?

[Martin Kantola]

But please remember that this kind of knowledge should not be what we use to make our musical decisions in the recording/mixing process. How sad if we would chose the right mike pre because of specs.

IMHO it's the kind of stuff you learn properly and then keep in the back of your head. Finding it extremely useful in helping to understand why a device sounds like it does, speeds up the process of selecting the right tool for the music I'm working on. Not to mention how much it helps technical troubleshooting, which should never get in the way of the flow of the creative process.

Martin

[Francis Vaughan]

Neat questions. This is my take on the answers. Do we get marked?

Which has more headroom, a discrete class A microphone preamp with transformer output or a differential circuit built with monolithics?

Depends a bit on how the 0db point is defined. But since the transformer output pre is able to go into some level of soft clipping, all other things being equal, it will have a naturally greater headroom in so far as it can be overdriven without drastic clipping and sound good, unlike the monolithic differential one, which will simply transit to hard clipping directly it exceeds its rail limited range.

Where is the best place in an unbalanced line to attenuate the signal?

At the reciever. If you attenuate at the source you have essentially thrown away noise resistance to stuff picked up in the link equivalent to the attenuation level.

If you short the cold leg of a differential input to ground, what happens to the signal level?

Nothing. A pure differential line doesn't care. It is the difference between the lines that matters.

Which gain control device has the least distortion, a VCA, a printed plastic pot, a photoresistor or a wire-wound stepped attenuator?

Wire wound stepped attenuator.

Will putting an unbalanced line on a half-normalled jack unbalance the normal signal path?

This seems a trick question. Since the line insertion disconnects the normal signal path it can't affect it. But "putting a line on a jack" is funny English, so I'm not sure I'm interpreting the question correctly.

Will a transformer splitter load the input to a device parallel to it?

Yes. The transformer reflects the impedance of all the windings to one another. They all appear in parallel (approriately multiplied by the square of the winding ratios)
with one another. So the impedance of each device connected appear in parallel to one another.

Which will have less RF noise, a shielded unbalanced line or a balanced line with floated shield?

The shielded unbalanced line. RF does not follow the same rules as audio frequencies. The CMRR is meaningless at RF, since even with a transformer the RF just hops across the windings through the inter-winding capacitance. At RF a floating shield on a balanced line does nothing, and the balanced pair just looks like an antenna coupling the RF energy straight into the next stage.

[scottwilson]

Here's some that I have opinions on - and I'm assuming literal questions with all else being equal as opposed to 'Well I know this sounds better'

1.Which has more headroom, a discrete class A microphone preamp with transformer output or a differential circuit built with monolithics?

The IC based differential circuit will have a greater linearity assuming the voltage rails are the same between the two.

4. Which gain control device has the least distortion, a VCA, a printed plastic pot, a photoresistor or a wire-wound stepped attenuator?

The VCA will have a distortion figure. The others are passives, and while each may affect the signal, they won't technically have any distortion.

7. Which will have less RF noise, a shielded unbalanced line or a balanced line with floated shield?

Balancing will not protect against RF noise, only LF induced hum/buzz. The shielded unbalanced will have better RF rejection.


-s

[Eganmedia]

[quote=I must say I consider myself a pretty accomplished engineer and am embarrased to say I have complete understanding of none of these. ?[/QUOTE]


Herein lies the problem. An Electrical Engineer has a piece of paper with the title printed on it, as well as the schooling and knowledge behind it. A "Recording Engineer" is a guy who decides to call himself one. That's not to say that there aren't bucketlaods of gifted, skilled recording engineers with no formal training whatsoever. It would be nice, though, to have some professional spec to distinguish those who are adept at the skill set needed to make a good recording from those who simply have the money to buy the gear.

When I go t the doctor I am somewhat reassured by his diplomas from Yale and Duke. They don't guarantee that he'll be a good doctor, but at least I know he passed some tests. The again, GW Bush went to Yale. Yikes.

[terminal3]

I had a very serious conversation a couple months ago with a very well respected mastering engineer who agreed that being a recording engineer has shifted in the past 10 to 15 years from being an actual electrical engineering sub-pursuit to that of an I.T. - and if anything - computer engineering-related field.

I love that Steve Albini sticks to his guns though, it's like being a holdout of film cameras in a digital world, and he gets away with it because he's so damn good at it.

[Martin Kantola]

Originally Posted by scottwilson

The IC based differential circuit will have a greater linearity assuming the voltage rails are the same between the two.

We also don't know if it's a single ended "class-A" circuit or not. Dynamic range is more important than headroom, because we can always chose our signal level...

A photoresistor does introduce distortion, definitely more than a good VCA.

Martin

[The Reel Thing]

well done boys.
now go start listening.

[scottwilson]

Originally Posted by Martin Kantola

We also don't know if it's a single ended "class-A" circuit or not. Dynamic range is more important than headroom, because we can always chose our signal level...

The question doesn't ask anything about what sounds better or what has more dynamic range. It's only asking which has more headroom. Assuming each circuit has the same reference 0dB and the same absolute maximum output, then they each have the same headroom.

Originally Posted by Martin Kantola

A photoresistor does introduce distortion, definitely more than a good VCA.

Distortion is an introduced non-linearity, and the only non-linear device in the list was a VCA. All of the others were passive devices, even the photoresistor. Maybe you're thinking of an optocoupler?

-s

[malice]

I know perfectly fine AEs (I mean, world class ) that don't know shit about EE.

malice

[Martin Kantola]

Originally Posted by scottwilson

The question doesn't ask anything about what sounds better or what has more dynamic range. It's only asking which has more headroom. Assuming each circuit has the same reference 0dB and the same absolute maximum output, then they each have the same headroom.

My point was that there's a lot of assumptions there (topology, rail voltages etc.) and that the headroom expression on its own doesn't tell us much in this case. You can have lots of headroom if you make sacrifices with the signal-to-noise ratio, in other words shift your "0dB" voltage.

Distortion is an introduced non-linearity, and the only non-linear device in the list was a VCA. All of the others were passive devices, even the photoresistor. Maybe you're thinking of an optocoupler?

No, a photoresistor (LDR) will distort an audio signal, it's not linear. There are many non-linear passive devices (like transformers).

Martin

[Karl J]

Originally Posted by joelpatterson

And here's the bonus question:

THE LEFT ONE!!!!!!!!!!!!

What do I win?

[joelpatterson]

Originally Posted by Karl J

THE LEFT ONE!!!!!!!!!!!!

What do I win?

The envy of your peers-- who haven't dared to hazard a guess!

[matta]

I seriously believe that having a working knowledge of electronics has made me a better engineer

I started years ago as musician, became a studio owner/engineer who in turn became a tech.. cause and effect if you will....

I started delving in electronics just to keep my gear running as I own a fair amount of vintage/old pieces of kit and I reckoned teaching myself this skill would save me a boodle of cash... little did I know it would cost me more in tools, gear and parts, but the knowledge has been invaluable.

Besides being able to repair/rack/replace parts/gear having a working knowledge of the HOW and WHY has made me a better engineer, training myself to HEAR the differences between say an OPTO, VCA or FET based Compressors and their interaction with an audio signal. Having serviced many of my pres it has given me a better feeling for each of their pros and cons and helped me make informed decisions about what to use when and where.

If nothing else learn how to solder your own cables...

Cheers

Matt

[Francis Vaughan]

Assuming each circuit has the same reference 0dB and the same absolute maximum output, then they each have the same headroom.

I would say therein lies the subtlety of the question. That is making two separate assumptions. The second one you can't make since that depends upon the amplifier design, and in effect is nailing the terms of the question. Absolute maximum output isn't going to be the same thing for a transformer stage and a direct output stage. You will typically design an amplifier so that you can tweak the 0dB point to be where the amp is still linear and free of noticeable distortion.

The question is comparing the classic Neve style pre with a modern transformer-less. The trick is in working out how to design the 0dB point.

The transformer is going to be the limiting part for that pre-amp - at noticeable distortion it has begun to run into its non-linear part of the BH curve. For an op-amp driven output distortion begins to occur right at rail limiting. It is already clipping the tops off the waveforms and the harmonics generated sounding bad. It has no headroom.

The subtlety now becomes the question of how the transformer pre designer sizes the output transformer relative to the class-A amplifier rails. He could design things so that the class-A amp is right at the edge of its rail limits too, but he doesn't need to. There is nothing stopping him from providing wider rails and allowing the pre to continue to drive higher and higher levels with increasing levels of distortion. Which of course is exactly what Rupert Neve did. It is worth remembering that the output transformer is typically providing some of the voltage gain itself, so for the same rail voltages the class-A design will have a greater final output swing than the op-amp based one.

[boxymoron]

funny, I'm not an "audio engineer", but these questions would seem very elemental for any electrical engineer.

I think "audio engineering" is somewhat of a misnomer. Engineering is about a practical application of science to solve a problem. Making music/recordings sound good is more of an art, it is a subjective endeavour, and perhaps more respectable imho. You are not solving a problem, unless chasing down some nasty 60Hz hum. Personally, I can't mix for shit, but I can design equipment (engineering) which could help make recordings sound better (subjective). Would you consider a sculptor an engineer? not more than you can call the person who engineered the tools of that trade an "artist". Dan Lavry comes to mind, lol. I dunno, there IS a fine line between art and science sometimes, but to me audio engineers make music/recordings sound good, they don't "engineer" audio. I don't think you can "engineer" something that is based on subjectivity.

[Mike Caffrey]

He's wrong, you need a locomotive, some overalls and a striped hat.



It's easy for him to make an arguement seem strong when it's fundamentally flawed.

What type of engineer? Train, sanitation, audio, electrical?

He's conflating audio and electrical.

You can certainly be a mix engineer and not have a clue what a mic preamp is.

[scottwilson]

Originally Posted by Martin Kantola

My point was that there's a lot of assumptions there (topology, rail voltages etc.) and that the headroom expression on its own doesn't tell us much in this case. You can have lots of headroom if you make sacrifices with the signal-to-noise ratio, in other words shift your "0dB" voltage.

Granted... I still feel that the point of the question "Which has more headroom, a discrete class A microphone preamp with transformer output or a differential circuit built with monolithics?" is to simply test if you're paying attention and understand from an engineering point of view that headroom has absolutely nothing to do with the class of the circuit or whether the components are discrete or integrated.

Originally Posted by Martin Kantola

No, a photoresistor (LDR) will distort an audio signal, it's not linear. There are many non-linear passive devices (like transformers).

Okay, do I get kicked out for backing down off that one? (photoresistors are semiconductors whose conductivity is created by photons knocking electrons free and has a figure called breakthrough distortion) However, if you're lumping in xformers into non-linear devices, then I'll say the actual answer is the printed plastic pot since wire-wound attenuators will exibit some hysteresis just as the xformer does.

-s

[projectMalamute]

Originally Posted by boxymoron

funny, I'm not an "audio engineer", but these questions would seem very elemental for any electrical engineer.

I think "audio engineering" is somewhat of a misnomer. Engineering is about a practical application of science to solve a problem. Making music/recordings sound good is more of an art, it is a subjective endeavour, and perhaps more respectable imho. You are not solving a problem, unless chasing down some nasty 60Hz hum. Personally, I can't mix for shit, but I can design equipment (engineering) which could help make recordings sound better (subjective). Would you consider a sculptor an engineer? not more than you can call the person who engineered the tools of that trade an "artist". Dan Lavry comes to mind, lol. I dunno, there IS a fine line between art and science sometimes, but to me audio engineers make music/recordings sound good, they don't "engineer" audio. I don't think you can "engineer" something that is based on subjectivity.

This is an interesting point, but not one which I am sure I agree with. I think in large part an audio engineer does in fact 'engineer' audio, but we are clearly dealing with a case on the border of art and engineering. While perhaps a sculptor does not need an engineering background I would say that an architect clearly does, but neither of these analogies translate directly to what an audio engineer does.

I also feel that in many cases things are not really that subjective, using the available technology to as accurately as possible convert an acoustic event to a document of it on magnetic tape is an engineering problem. The process by which this happens should not be a mystery to someone who claims to be an audio engineer and charges money for providing that service.

All things being equal I would always choose to record with someone who has a deep grasp of how and why every piece of equipment in a studio operates over someone who is able to 'feel it' and get 'that hit sound'. In fact, I think that is the essential difference between what I would call an audio engineer and what I would call a producer. And I have no use for producers.

[joelpatterson]

Originally Posted by projectMalamute

I also feel that in many cases things are not really that subjective....

Wha? Every single solitary move you make as an audio engineer is completely subjective.

You must decide out of thin air which mics to use, where they go, what the signal chain is going to be, and then you must turn the knobs on all those gears to some indeterminate, subjective setting.

And that's only tracking. When you mix down, you are bird-free to set levels and pans and subgroups and side-chain compressors and ride faders and... everything.

I myself thought boxymoron had an excellent "take," if you will.

[projectMalamute]

Originally Posted by joelpatterson

Wha? Every single solitary move you make as an audio engineer is completely subjective.

You must decide out of thin air which mics to use, where they go, what the signal chain is going to be, and then you must turn the knobs on all those gears to some indeterminate, subjective setting.

And that's only tracking. When you mix down, you are bird-free to set levels and pans and subgroups and side-chain compressors and ride faders and... everything.

I myself thought boxymoron had an excellent "take," if you will.

Listen to the sound in the room. Listen to the recording of it. Have you captured that sound in an accurate manner? Granted this is not the only standard to apply in whether or not something sounds 'good', but being able to do this consistently with whatever equipment is on hand is to my mind more of an engineering problem than an artistic one.