When did recording studios switch to digital recording from tape?

Denon’s earliest digital recording is from 1972. Released on LP that very year.

http://www.aes.org/aeshc/pdf/fine_dawn-of-digital.pdf

 

I think the most vinyls pressed before 1980 are mastered from analog master tapes. Obviously, the exceptions strengthen the rule.

 

Well, sort of. By the late '70s digital delay lines were popping up in vinyl mastering - it was first introduced in 1973. It's possible the majority of all vinyl releases were going thru a DDL by the early-to-mid '80s.

 

I've never read anyone speak negatively about the sound quality of Dire Straits - Brothers in Arms. That was a digital recording and only recorded at 16/44.1 if I'm not mistaken.

 

Remember, Digital recording was 2 channel only until 3M released their digital recording system, then Sony and Mitsubishi later on.

 

I have spoken about it negatively. I think people love the production, but I think the sound is thin. Great album, by the way. It just sounds “80s digital “.

 

Yes, exactly. It was 2-channel-only for a long time. That's why it was first used mostly for classical music, and occasionally jazz. It's safe to assume that many of the recording engineers who began to dabble into digital at the dawn of the 1980s had started their career in the '60s, and were used to mixing stereo on-the-fly, so that was no big problem when it came to classical. In pop, however, the need to do multitrack recordings and the need to be able to edit them led most studios to still use analogue tape well into the 1980s.

 

I think the thin sound has to do more with pop production tastes of that era than with something inherent in early digital recordings. Although I'm sure the clean sound of digital did its part in shaping those tastes.

There were lush and warm sounding digital recordings earlier than that, but most came from the classical and jazz realm. For instance, UK Decca's first digital recording, the 1979 Willi Boskovsky New Year's Eve live recital of Viennese Waltzes sounds amazing—you'd never guess it was recorded digitally.

 

I think you’re correct about the coinciding of 80s production and digital sound. Still, analog would have fattened up some of those desperately thin pop and rock albums, if only by the nature of the bass bump of 15 ips tape.

The improvements in digital sound through the 90s was probably better converters and people getting used to how to work with digital’s strengths and weaknesses.

 

Fully agree. Very good points.

 

I'm not sure that's true. I think 3M's 32-track system blew away anything else on the market in '79, but if memory serves Soundstream had at least a 4-channel recorder out in '77, and they offered an 8-channel unit at one point that could work with 3M's system, so you could use Soundstream's advanced editing capabilities. It was the first digital audio workstation, where you could enter commands and see waveforms on a computer screen - the precursor to ProTools and all that later stuff.

 

Folks, just read the paper I linked to at the top of page 2. It goes through the pioneering work and first commercial systems (both 2-ch and multi-track).

 

A lot of the 'thin' sound heard today on pre 1990 recordings is often the use of DolbyA, but not properly decoding it. I have been suspecting (and detecting) undecoded material for the last few years, and just got some confirmation of that fact on the 'gearslutz' forum when discussing some matters about the DolbyA issue. (the confirmation came from someone who actually was doing digital releases in the 90's using the earlier material.) Undecoded DolbyA -- even after some EQ, tends to sound thin, tends to flatten the stereo space (stereo is still there, but much of the depth is obliterated), and the material will have that '80s' harsh digital sound. After decoding, there are a LOT of old 'diamonds in the rough', and many of my attempts at decoding feral DolbyA material provide results very similar to direct vinyl rips made from vinyl of the era.

John

 

I've always wondered about that. I also wonder if sometimes different "masters" of a given album are or aren't Dolby A encoded, and things get mixed up. Like the original studio 2-track mixdown master is encoded, but the copy made while mastering the original LP isn't. So the LP master doesn't need to be decoded, but the mixdown made in the studio does, and that information gets garbled somehow.

Or the labels were just lazy effers and didn't bother with decoding - just played the tape back and digitized it.

Does anyone know how well the presence of Dolby is indicated on the various "master" tapes floating around out there?

 

The situation you describe is *exactly* the kind of thing that the mastering engineer told me. Some record labels aren't all that careful about the documentation on their tapes, and so the encoding isn't always known by the person doing the mastering. They try to do the best job that they can do, but with the metadata messed up, and the limited resources allowed, the correspondent told me that the consumer release 'feral' DolbyA happened, quoting him: 'a lot'.

From what I have seen -- many recent sales of older material have DolbyA encoding (or have been processed further without decoding.) One specific example 'The Complete Studio Recordings' of ABBA sound non-decoded (I am pretty good at detecting that now), but also very heavily compressed. The compression was so severe that there was no real benefit of decoding the DolbyA anyway (DolbyA provides compression between about -20 and -40dB in 4 bands -- the 4 bands are combined in an eccentric way.) So, you take the DolbyA encoded material, do a -3dB@3k/Q=0.707 or -6dB@3k/Q=0.707 or maybe add in a -3dB@9k/Q=0.707, and that kind of situation is what I am finding. Most often, it seems like just the -3dB at 3kHz to do the decode on the cheap (or maybe the mastering engineer was just trying to make it sound correct, not sure if it was DolbyA encoded.)

Decoding DolbyA imparts some excess distortion coming from the fact that the 'inverse encoding process' (which is what DolbyA does to decode) is actually not 100% accurate. The DolbyA encoded version of material has most of the original -- before encoding -- quality, but the normally available decoding methods are not as accurate what one really wants.

I can actually understand some reasoning to avoid decoding DolbyA material, doing some EQ instead, then use a good quality current technology NR mechanism. I wrote a very clean DolbyA decoder because of the not-quite-correct decoding from an actual DolbyA HW unit. Our project's DolbyA decoder is definitely very good (it is called DHNRDS, being done in conjunction with a fairly well known restoration/recording engineer), but that isn't really the subject of the discussion here. (Developing the software on this project is why I know a little bit about DolbyA decoding issues.)

John

 

One more follow-up note (just my opinion) -- if you have the choice between a unmodified, 'Feral' DolbyA copy of an older recording, versus a molested, hyper compressed remastered version -- you are better off with the 'Feral' DolbyA version. First -- the material (sound) is all there in the recording, but might need a little EQ, and if REALLY REALLY motivated, you can get an old DolbyA unit or get some kind of DolbyA decoder software (I mean, a good decoder -- not just something that 'just works'.) The hyper-compressed remastered stuff is often damaged beyond recognition. In the very best case, try to get a properly remastered release -- one that has been treated with great care and properly decoded, and NO excess compression beyond the master.

John

 

This was the first digitally recorded album ever made, recorded by Denon in 13-bit PCM digital in January 1971. Direct to 2-track, no editing (because they couldn't edit digital yet). Released on vinyl in 1972; re-released on CD in 2013.

 

Sounds great. Interesting hearing something from that era without all the hiss...

 

I think Bowie's "Lets Dance" was the first totally digital album.

 

It is wonderful to hear an early (the early) recording, and it shows that 13bits isn't all that evil (not as good as what have gotten used to in good recordings today though.) However, I do have a few potentially eye-opening examples about hiss from that timeframe. My hearing is recked by tinnitus, however I did check the hiss levels on these examples a spectogram, and these recordings (fairly common at the time) seemed not to be all that hissy. Tell me what you think... Remember my rants and raves about not properly handling some of the old material -- I did NOT master these recordings, but I did do the proper first steps for mastering. These just might be less hissy than other normally available copies, but absolutely ZERO special single ended NR was added, just doing standard things needed for NR/EQ/etc. Added no 'love', just bog-standard processing.

Again, caveat: I truly can not accurately judge though... These snippets will be up for about 1-2wks -- probably long enough for this discussion. This is not using the standard Youtube thing, so I must only supply snippets. Again -- tell me if these are hissy :-(.
Also -- sorry about the mp3, but they are good enough to judge hiss -- if you really need flac, I can supply them.

Dionne Warwick -- You'll Never Get to Heaven -- 1964 (I think, at least the 1960s)
Dropbox - 05 Dionne Warwick - Walk On By - You'll Never Get To Heaven (If You Break My Heart)-V0.9.4T2-xxp-13.90-45secs.mp3 - Simplify your life

Carpenters (early) -- Reason to Believe -- 1970
Dropbox - 04 - Reason to Believe-V0.9.4T2-55secs.mp3 - Simplify your life

John

 

With regards to Dolby etc., in my experience anything properly recorded in a decent commercial studio with a good engineer started out with all the information on the tape sheets or boxes, after that things often went downhill with sheets lost, boxes swapped and information lost, great fun for anyone who specialises in transfers and that's not to mention baking, then there are the fishing expeditions where someone hands you a stack of reels all missing sheets and tells you the track you want is somewhere on one of then, probably.

 

After reading your comment and re-reading mine earlier. I must apologize about suggesting that the label/studio or whatever didn't care about the metadata. More (like you wrote) it sometimes gets lost, misplaced or just the situation gets confused. Reading your comment -- I respect and accept your information along with my other correspondents who have been kind enough to tell me what is going on. Thank you.

John

 

I did some research on this previously and the best I could find is that The Beach Boys “Summer In Paradise” (1992) is the first true all-digital recording. As in, it was not only recorded on a digital multitrack and mixed to a digital format, but also mixed using a digital mixer. Previous “full digital” recordings would have undergone several A/D/A conversions as they were mixed through analog boards. It was recorded with an early version of pro-tools.


I would say that digital didn’t become the dominant professional recording format for popular music genres until the late 90’s or very early 2000’s. So many big records were still done on 2” up to that point. Doubt any major label did an analog classical session after 1983 though.

 

Actually, that recording from the early '70s was "all digital", as it was recorded direct to two-track digital with no mixing. Denon did that with many classical recordings prior to 1980.

The first digital consoles appeared in 1986 according to the AES webpage, so it's highly unlikely that Beach Boys album was the first all-digital recording, even if you take into account multitrack mixdowns.