Recording Studio Equipment

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
In another thread, it was mentioned that recording studio quality started a decline about 1974.
http://www.diyaudio.com/forums/showthread.php?s=&postid=260806&highlight=

One of the reasons, apparently, was because of the use of multi-point microphoning (rather than plain stereo microphone recording). I've seen recent articles touting the use of Sennheiser microphones for multipoint recording. For me, I guess that stereo v. multipoint each has advantages and disavanatages.

Being that I would like to build an amplifier, I've wondered if studios use op amps or not. An answer, in one case, was not too difficult to find. I just purchased a new Deutsche Grammophon recording (of Boulez conducting Berlioz "Romeo et Juliette").

From the CD liner:
Recording facilities: Polyhymnia International.
From their website: "Polyhymnia International was formerly the Philips Classics Recording Centre. The Philips Classics Recording Centre was founded in 1950, and moved to Baarn in 1973." Interesting that that is about the year analog_sa and Christer write that things (in recording studio in general) started a decline.

Further, under "New Developments":

"Mixer: The next item to be modified was our existing PolyGram analogue mixers. These mixers, built in the 1970’s especially for PolyGram’s classical labels, had already been modified several times. The most recent modifications included replacing all of the op-amps with low-offset *Burr-Brown* models, and removing all of the capacitors in the signal chain, resulting in a fully DC-coupled mixer. "

Spec-wise the Burr-Brown OPA627/637 is probably the best audio op-amp there is; however, it answers for me (in this particular case) whether they use discrete circuitry or not. On the other hand, I’ve recently noted that Forssell Technologies makes discrete (solid state and tube) mixing equipment. (Forssell's schematics page is interesting.) I mention Forssell to acknowledge that there is discrete equipment used in recording studios. Again, I'm not in a position to say one is better than another.

http://www.polyhymnia.nl/
http://www.forsselltech.com/

Anyways, I will see what other information that I can learn of my other CD liners. The type of circuitry used in the studio seems more practical to me. (Although learning about different recording techniques are also interesting.) Fortunately, my favorite composer/conductor has recordings dating from 1951 to this year. So, I ought to have a good variety of recordings to experience.


JF
 
Ran across the article (mentioned above) that briefly discusses the layout of microphones in a recording session.

http://www.zioshow.com/viewnews.php?id=2553

Recording the Cleveland Orchesta:
- four (MKH 800s) microphones were used across the front,
- one binaural (KU 100) was used over the conductor,
- a double set of MKH 30 and MKH 50 was used in the mid-side and
- a pair of Sennheiser MKH 20 microphones in midway back in the hall.

Sound Engineer Michael Bishop states, "you want the whole surround image to work cohesively, and that's only going to happen if you use a family of microphones. I need a complete sonic signature from front-to-back, and for me that means using nothing but Sennheiser microphones."

The article reveals a bit of the professional effort that goes into audio recording.

Again, my interest is learning what type of electronics is used for recording classical music. (That's why I started the thread in the Solid State section.)


JF
 
Studer A807 professional analogue two-track stereo tape recorders are full of NE5534's or NE5532's in the signal path. An old (late 1960's or early 1970's, I guess) professional Studer mixing console I've repaired a couple of times already contained LM301A's, with an external discrete input stage connected to it for the low-noise sections.
 
MarcelvdG said:
Studer A807 professional analogue two-track stereo tape recorders are full of NE5534's or NE5532's in the signal path. An old (late 1960's or early 1970's, I guess) professional Studer mixing console I've repaired a couple of times already contained LM301A's, with an external discrete input stage connected to it for the low-noise sections.


Interesting, I was reading through the posts of forum member mhennessy (because it seems like he works at a sound studio) and noticed his schematic at his webpage (link below) that he also uses NE5532. I wondered why. I've read about OPA627/637 and AD8610, but the NEs are older parts. I guess they have a good reputation in the audio industry. I've read otherwise good comments from other people, too. Thanks!

http://www.mhennessy.f9.co.uk/

schematic
http://www.mhennessy.f9.co.uk/preamp/Analogue.gif


JF
 
Bearing in mind that I am not a pro but have only played at FOH and PA when I was younger (and mostly as the volunteer gofer for the hired gun), I can distinctly recall an attitude that the equipment must be tried and true.

You all know... consoles durable enough to withstand slamming the faders down when a performer feeds a mic, preamp buffers that don't pop or buzz when the little gofer pulls a plug or patches a DI box, tape decks with transports strong enough to deal with a reel that got all unwound during cue up, line outs with enough power to drive headphones. I kind of miss it, in a masochistic kind of way. Stage mothers, performance cassettes made by facing a boombox mic towards a boombox speaker, "No... not that song, the one that's two before it!" Such warm fuzzy memories.

That said, I'm not surprised to see the NE5532 or a TL07x inside good equipment. These are old standbys that are clean, have good headroom and can be had for cheap. One day, the OPA627 might be commonplace, but if I were a pro audio designer and trying to balance performance with profit, those two old chip families are sure hard to beat, especially compared to discrete. I am kind of surprised that the OPA2134 isn't more common in newer gear as it is almost as cheap and a bit better than the other two. The only place where I can see a tube is in the mic preamps where it can pretty up all the little divas singing their little hearts out. What's that brand now... Neve?

:)ensen.
 

PRR

Member
Joined 2003
Paid Member
> wondered if studios use op amps or not.

Chip opamps, module opamps, discrete opamps, or non-opamp?

The classic transformerless mike amp can't be done right with a single opamp, and until recently not well with any chip opamps. But a lot of designs use a couple fat transistors with 3 or 4 good chip opamps.

The classic transformer mike-amp, tone-amp, and especially the summing amp beg for opamp simplicity. The repair department is much happier when everything uses the same card, strapped in the back-plane. There were scores of audio-optimized opamp cards and modules; there are people at recording.org tech talk restoring or recreating the classics (and the crap).

One stand-out is the Jensen 990. This was a discrete design sold as a potted module. The potting was for thermal tracking: Jensen published the insides, no big secrets (some small secrets). As a DC opamp, it sucks: huge input bias current. As an audio amp it is so good it is still in production today, for both old and new systems, from several makers. But using it to best advantage requires somewhat different design (much lower impedances) than chip-users tend to use.

I first saw chip-opamps in a 1972 RCA broadcast console full of 709 (yes, 709!). It was real gee-wizz at the time. In retrospect, the 709 is hardly the best opamp ever made but, if properly compensated (a pain) perhaps better than many of the things that came after it.

The Gately was full of 301, which is a 741 except you can de-compensate it for better gain and slew. Compared to most low-price discrete designs of the day, it is far cleaner and more predictable.

The better discrete amps of that era were better than 709 or 301, but a LOT more expensive. I dunno about the cost-no-object crowd, but in affordable gear chips often had less-crappy pots and switches than legacy discrete designs, because they could afford to.

And then came the TL072. A very good chip, within its limits, but so darn cheap that designers went crazy and used them by the bucketload. Here is when I noticed that a 2-transistor preamp sounded better than a 40-chip mixer. There is nothing much wrong with TL072, except it is too easy to use too many of them.

The 5532 is about as good as any discrete design, but is a chip. For 150Ω sources we would like the input transistors even fatter, and for very-high Pro power levels we would like more voltage and dissipation. But within the limits of a mass-market chip, it is still hard to beat, with chips or discrete. It does have high bias currents and has been around long enough to look old-fashioned. And it is possible to make a bad mixer out of good chips. But mostly the 5532 is an excellent choice for many audio tasks. Once source says that Rupert Neve had some casual input into the 5532. So many other "audio opamp chips" just-miss excellence; maybe Rupert did tell the chip-heads where they were just-missing excellence.

"Opamp" typically implies differential inputs. That is not always necessary in audio. A summing amp can be built from a single-input inverter, and a 1-transistor input -may- give less noise than a diff-pair. However, while we can block DC with caps, power rail and output swing and general stability tend to favor a 2-input opamp over a simple inverter. And using one amplifier (card, module, or chip) for most or all of a board is a real advantage in the hustle and bustle of a commercial studio, which can't go "down" for long without losing BIG money.

Neve did a lot of good designs in single-ended discrete, some of them without complementary types (all NPN). These have mad fans. I'm not sure the actual amp is special (except from having only small sweet flaws): his magic seems to be in transformer selection and a lucky pick of capacitors. And fans debate the "color" of these amps, which tells me they are imperfect but in musical ways.

If you go way back: the classic Langevin and RCA tube consoles used plug-in amplifiers. They were not called opamps and you almost never ran external feedback around them (it was fixed-gain), but the audible difference is moot. And a large mix board would have just two types of modules: a 35dB gain +18dBm preamp and a 40dB gain +30dBm line driver. And there were two slots next to the power supply where you could keep a spare of each type hot and ready for instant swapping.
 
PRR said:
Chip opamps, module opamps, discrete opamps, or non-opamp?


Guess, it's ingrained in me that opamp = chip opamp. (Even after reading about Forssell's discrete opamps.)

http://www.forsselltech.com/

And I forgot how close the subject of audio recording it seems that you are PRR. (Though, in fact, I was hoping for your slant on the use of discretes and chips.) So, again, thanks another long reply.

In general, from what I gather, one advantage of discrete circuitry is that the passive components can be better. Apparently, quality resistors and capacitors don't lend themselves to small integration. Quality being things like high Q, linear, and with low parasitic components. Seeing distortion graphs of capacitors leads me to believe that it is good to minimize use of low quality capacitors. Of course, discretes provide flexiblility for true class A topology.


JF
 
I just looked up the A807 manual. Besides NE5532's, there are also quite some RC4559's in the signal path, as well as AD7528's (apparently multiplying DAC's) used for automatically setting all kinds of levels and equalisation time constants.

Regarding passive components on chips, except in some exotic IC processes, it is true that integrated passive components are usually not as good as the better discrete ones when it comes to accuracy, quality factors and/or linearity. Besides, high value capacitors and inductors above 20nH or so are simply not integratable on an acceptable chip area.

Regarding capacitor and resistor linearity, modern processes with polysilicon resistors and poly to poly, poly to metal or metal to metal capacitors are MUCH better than old-fashioned processes having only diffusion resistors and junction capacitors. The absolute accuracy is still bad, though, 15% to 30% tolerance being typical.

However, I don't think that the quality of integrated passive components affects circuits with op-amps much, because in these circuits the really critical passive components, such as feedback network components, are external anyway. I can only think of one exception: the Miller compensation capacitor may affect the linearity at high frequencies to some extent if it is very bad.

I think one inherent disadvantage of op-amp circuits is the fact that you cannot optimise them for a certain application to the same extent as you could with discrete circuits or with application specific integrated circuits.
 

PRR

Member
Joined 2003
Paid Member
> I hoped to hear *your* slant on discrete v. chips

As for the active components: they are about the same. They all come off the same process.

Dollar per square inch, it is hard for a chip designer to justify very large transistors. The 5532's transistors are about as big as the cost department can justify, and silicon has not improved much since then, which is one reason the 5532 still stands as a good audio amp.

Until recently, NPN and PNP on the same die was always a compromise. At one end, you can do an all-NPN design like many early classic discrete boards or the LM387 (couple PNP for bias, all NPN in the signal path; and BTW not a high-end audio chip and now discontinuted). Going discrete allows random polarity selection without worrying how you can grow it on one lump of silicon. (Newer processes have greatly reduced the cost of equal-quality NPN and PNP in one chip.)

Resistors: well, in the end of the discrete era, everybody was getting resistors out of the signal path to raise gain. You butt two collectors and a base together. Low-value silicon resistors are not SO bad (high-value ones are pretty strange creations).

I'm sure, especially on modern processes, you could build a chip as "good" as any collection of stray parts. But there are already dozens of very good and wide-market chips. To get better audio performance, you probably have to sacrifice other parameters, which reduces your market, and the top-end audio market is already too small to justify a chip-design effort for another audio-only chip.

There are several things chips are not real good at. Very large area low-Z lo-noise input devices and large-area high-current output devices eat die-space by the yard. That's one place you want to go buy bulk-silicon outside the chip process (ie, discrete).

Capacitors and resistors are very limited on chips. Of course the trend is to eliminate ALL signal-path caps. Mostly the signal caps are to block DC. The excellent matching of devices on one chip allows DC cancellation without caps. Probably a good thing, though some people select caps to tailor sound.

The intimate thermal coupling on a chip is good for some things but does pose one problem. If you have high output power, that heat gets back into the input stage. Heat is slow, but some chips show a distinct rise in THD below 100Hz, which can't be explained electrically but makes perfect sense when you study the heat. The same problem plagues DC amplifiers, like load-cells and thermocouples, so solutions are well known. The most direct is to move the power stage off-chip, to a buffer (chip, discrete, anything) or move the input stage (LM194 pair feeding a LM318 or similar). But careful attention to input device layout can give a major reduction in thermal distortion.

In discrete assemblies, the current and signal has to flow out of one part, over to another, and into it, with several changes of conductor, joints, and lots more stray capacitance and inductance than an integrated circuit.

Everything is a compromise. The big nut in the discrete versus chip issue is: you can knock-up a new discrete design with a $5 soldering iron, a new chip design costs $5,000-$500,000 for the first sample.

> I hoped to hear *your* slant on discrete v. chips

It is a non-issue. The wise designer will keep an open mind and do the best job possible with any parts available.
 
Do you really believe that a 5532 is a good part by today standards

PRR,

I read a number of things you wrote, so that is why I am having a difficult time believing you mena what you wrote. I stopped using 5532 and 5534 back in the late 80's. There are just so many better parts for all types of applications. Now I understand these parts on cheap, and some time we have to build to a price, but stating these are some of the better parts just does not fly.

Have you used the new Burr Brown and Analog Device stuff?
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.