Denon PCM Encoding in 1970s. Is it different than Sony CD PCM?

Yeah, that's what I was thinking - must have been a prototype.

I wonder when 3M embarked on building that first all-digital multitrack recorder? Must have been around 1975.

 

My guess would be 1976 since Soundstream was first started in 1975. I assume Stockham's entry into digital audio got the attention of US companies more than Denon did from Japan.

 

Ever listen to the Telarc SACD reissues made from the 16/50 Soundstream masters? They sound really good, and can give modern digital recordings a run for the money.

Heck, The Nightfly SACD (available on a Japanese import), derived from the 16/50 3M original master, could very well be the best sounding album in my collection.

 

I'm just curious if any of these 16/50 recordings (or other early oddity digital formats) will ever be made available in their native form? I'd love, for instance, a digital clone of the original Nightfly PCM master. Or many of those Telarc recordings. Maybe these can be sold with the caveat that many hardware/software combos may not play nicely. Bundle in a high quality upsampling to a standard rate like 24/88.2 so everyone can have the next best thing if the real files give them trouble.


But I suppose that would put an end to a lucrative remaster business if they really gave us *the* master.

 

I have just been trying to keep an apples- to apples comparison as much as possible since these pre Redbook PCM converters were all output to LPs only. So my frame of reference here is comparing those to Redbook sourced digital LPs.

However speaking of SACD and the DSD process it used, there was a proto delta sigma recorder in the mid 80s, the DBX 700. It had a sampling rate of 644 kHz (bit less than 1/4 of the eventual Sony DSD) and output in video data format. I have only found a few LP recordings made through it exclusively:


A Solid Brass Christmas - Musical Heritage Society (1986) MHS 7418W.

Beethoven The Violin Sonatas Vol. 1 Robert Mann/Stephen Hough (ASV AMM 151R 1986_

 

I'm not an expert on digital conversions but my understanding is that in the last few years current sample rate converters are basically transparent unlike the non-transparent rate converters of the past. Of course one work-around in the past was to use exact multiples eg 88.2 kHz or 176.4 kHz so that samples could just be discarded without sample rate conversion.

 

Re: The Nightfly, I think that 16/50 transcoded to DSD is the best we will ever get. I am not sure if there is any consumer hardware made that will support playback of 16/50 material, so I doubt the raw PCM file will ever be released in that format. The DVD-A was 24/48 which probably entailed resampling from 50 to 48 kHz and then padding the 24 bit words with zeroes for the 8 least significant bits. Those 24/48 files are available on HD Tracks and are probably the closest we are going to get in the PCM world. Anyway, the 24/48 files lose some (inaudible) data vis a vis the 16/50 original master, so I prefer the DSD which although transcoded has all of the original recorded frequencies intact.

 

I would guess you are right. My own preference would be DSD x 2 or 5.64MHz. My subjective impression of SACD that it does not sound like PCM, but still has some oddities of presentation probably due to the partial loss of high frequency information from noise shaping with that 2.82MHz sampling rate.

 

Is that the multichannel version? I hear it's great - I should grab it sometime.

No, I read an article by Roger Nichols regarding The Nightfly and getting it off of the multitracks for the multichannel remaster. I'm assuming the two-channel high-res master was produced via the same process. Anyhow, as I recall there's no way to get multitrack digital data off of that thing and onto anything modern and even if you could, you wouldn't want to. That's because those early A/D converters had linearity issues, and so the A/D converters for each track were paired with matching D/A converters that corrected for their linearity issues! So when Roger did the multichannel masters, he converted each track to analog using carefully adjusted D/A converters on an original 3M deck, then redigitized every track using state-of-the-art converters.

I found the article here: Roger Nichols: Digital-To-Digital Transfers | Sound On Sound »

It's possible the two-channel masters were directly transferred, as I believe the best A/D converters were reserved for the two-channel recorder, and so wouldn't have those linearity issues. I don't recall. But I thought his account was a fascinating peek into early digital recording.

 

If that's the case, I could see the 2 track mixdowns being fine… but I also could see those potential linearity issues being fixed in the digital domain pretty easily nowadays. I would hope that they could find a way to archive and preserve the original pcm data. That's audio history right there.

I'll have to check out the hdtracks, PCM->DSD seems a little more convoluted to me partially due to all the filtering necessary for DSD. Seems like that 48khz version may be as close as it gets though, if it really is a straight resample of the original master. I have the old WG Target cd and even that of course always sounded pretty good to me, so I can only imagine what getting even closer to the master would be like.

 

The 3M converter seems to have had quite a few practical and design issues and problems. The WB Director of Engineering Lee Herschberg recounted the process of recording Ry Cooder's Bop til You Drop with the 3M and it seems very frustrating and dodgy. I have not heard similar complaints about the Denon or Soundstream converters.

 

The Soundstream system was only 4 track though (really 8 I think, but each track had a duplicate for redundancy). 3M's system was full 32-track digital. That introduced quite a bit of complexity.

Realistically, the thing was probably a couple of years ahead of its time. If 3M had just waited until 16-bit A/D converters were a bit more standardized and available, the system probably would have been a lot less twitchy and a lot more successful. I haven't heard of similar problems with the early Sonys, but they apparently didn't sound quite as good, either.

 

Good point. Denon was a 4 track. Of course the sheer number of tracks, while a big reliability and usability issue, is a sonic issue only if you are doing a lot of processing and bouncing. I believe digital mixing was primitive in the early 80s so they had to constantly flip from digital to analog to digital. I am having trouble finding LPs mastered through the 3M system except for the 4 that are widely known. Maybe people just found it too difficult.

 

I suspect that's exactly what happened. Also, the consoles were all analog at that time IIRC. So the multitracks from the 3M were digital, and the final 2-track was digital, but the mixing would have been done I presume in the analog domain.

Of course, mixers were waaaaaay cleaner than any analog tape. I don't even know if you could tell the difference between something mixed on a good analog console and a good digital console - I'd imagine both would be 99.99% transparent.

 

Those early telarcs are some of the only early digital examples I know of where the recordings are "pure" digital as they were basically mixed on the fly or live to 2 track, etc. multitrack recordings were still going through multiple conversions in and out of analog boards until the 90's.

 

I wonder if the Denon's were fairly pure as well, particularly their classical titles. I am in the process of acquiring more pre Redbook Denon PCM LPs and the ones I have gotten recently sound very clean and unfutzed with.

 

Multichannel and stereo.

 

Although I believe the Denon masters may be the other rare "pure" exceptions compared to other digital recordings of the time, I don't believe their initial CD releases are "pure" - Denon mastered a lot of their discs with pre-emphasis and unless I'm mistaken, both the emphasis and the pre-emphasis take place in the analog domain - that's always been my understanding anyhow but I've never seen any documentation on how it was performed at the mastering stage, perhaps someone can clarify. I always thought it involved playing the digital master back as if it were a tape master and inevitably recapturing it.


If any of those recordings were reissued later on but without the pre-emphasis, it's possible that they haven't gone through any analog conversion prior to the final product. Of course for an LP release, an analog conversion stage is inevitable, but they should still have gone through a few less conversion stages than was typical of other recordings.



I don't believe those early Denon Jazz discs were mastered with pre-emph - those may be pure digital, which is cool being that they date back to the 70's

 

The 2 channel Redbook of The Nightfly, at the very least, was sample rate converted from the original 3M digital master. How do I know this? Because mistakenly, the very first German Target CD mastering (the -02 mastering) was taken from the analog LP cutting master and was quickly withdrawn and replaced with a mastering derived directly from the 16/50 digital master after Nichols heard it and was not happy. Rumor has it that Stevie Wonder was one of the first people who complained about the sound quality and brought it to Nichols' attention. A second German Target (the -03 mastering) was issued, as well as a Japanese Target, with the correct mastering.

All that being said, I have a copy of that original -02 mastering and, although it doesn't sound AS good as the -03, it's still pretty dang good!

 

See here:

Keith Hirsch's CD Resource » Blog Archive » The rare Donald Fagen The Nightfly “02” Target »

 

I agree which is why I am only citing digital source LPs that could not have been processed through Redbook converters at some point. According to Billy Budapest the pre Redbook Sony PCM 1 is their best sounding converter which makes one wonder if Redbook and its progeny are the real problem with digital audio.

 

This is what Roger Nichols said about The Nightfly 02 debacle on his now-defunct forum:

 

I'm not sure I understand. Are you saying that the CD -2 source was an RIAA analog master without correcting the RIAA EQ and then sent through a Redbook converter?
But then the -3 50kHz digital file had to go through a Redbook converter too didn't it?

 

Hold on a minute--I'm not sure I've heard anything made from PCM-1 or PCM-10 tapes. It took me a while to pin down what I've heard matched to the correct ADC. The good Sony's that I've heard are the PCM1600 and PCM-F1. The later PCM-1630 and DASH machines are the ones people usually find objectionable.

An example of a good sounding Sony PCM-F1 master is Madonna's first album. It has a fat, almost analog sound that is totally missing from her later albums mastered from PCM1630 tapes.

 

No, RIAA EQ isn't on tape, it's done internally via the cutting head amplifier. Rather, it was a 30ips tape copy used as an LP cutting master. It had other EQ adjustments such as shaving off some of the bass and perhaps even summing the bass to mono, as was common in LP cutting.