CD-4 half-speed transfers on an Akai 360D from 1978

[ndiamone]

When I was in middle school, a neighbor man of ours back in Michigan was working as an engineer at the Air Force base alongside everybody's father. He was a little younger than our fathers and hence unmarried and therefore considerably more well off since all he had to spend his engineers' salary on was himself.

He got an Akai 360D 4-speed reel to reel (plays 1-7/8 all the way to 15 IPS) for a song from one of the guys on base when he came back from Thule, AFB Greenland.

By then, CD-4 was starting to fall off in popularity. As a result, earlier that summer in May, the Base Exchange was blowing out all their old stock, including dozens and dozens of quadraphonic titles, a number of them Japanese CD-4 imports guys had ordered and never picked up, all for a dollar or two apiece and Mel bought the whole lot. I remember him bellyaching later on that summer that it was a shame you couldn't tape the CD-4 onto a reel to reel and be able to demodulate off of that and save the LP as you could for stereo and SQ.

Enter endless cases of still-sealed BASF LP 35 CR tape, the CRO2 reel to reel designed to give 7-1/2 IPS fidelity at 3-3/4 IPS. The local ice-skating rink had been using it in that fashion to get 6 hours of music on one tape, but had recently changed from taping their own music off records to subscribing to a music service which supplied their music digitally in 6-hour blocks using PCM F-1 format on VHS tapes. Therefore, once the rink dumped all their analog gear and tape, Mel got all this expensive tape donated to him along with the seriously worn out leftover music reels.

He had the bright idea that since the CD-4 LP's were recorded at 16 RPM, he could play them back at 16 RPM on his Thorens TD-124 4-speed turntable, record them at 7-1/2 on the Akai with the LPR 35 CR tape, play them back at 15 IPS and demodulate. So he invited me over just about everyday that summer, and we spent the days taping the likes of Emerson, Lake and Palmer, Carole King, the Jackson Five, Joni Mitchell, Aretha Franklin and all kinds of other great records.

But he never got to find out if you could demodulate off that or not, because over Labor Day weekend he was killed in a car wreck. His mom kept everything of his pretty much just as it was, and when she passed away earlier this year, she had her grandson (Mel's nephew) call my Dad back in Michigan who gave him my number out here.

He called, and told me Mel's mom who we all called Nana Reenie (for Irene) had said before she passed that if I wanted all Mel's gear, records and tapes I could have it. So I went home over 4th of July weekend to bring it all back, and it turns out you CAN demodulate CD-4 from an LP transferred at 16, recorded on LP 35 CR tape at 7-1/2 and played back at 15.

So, if all the transfers are as clean as this one, I may use them for discrete 4.0 channel BR-D conversions vs. going back and transferring from scratch off the 5 cases of LP's I got in the deal.

 

[zabble]

Interesting story. For one thing, I'd sure like to see that copy of Emerson, Lake, and Palmer in CD-4.

 

[JonUrban]

Are you sure you don't mean 1978?

 

[ndiamone]

Dunno, I thought it was a late-70's deck myself, but all the Google results I get tell me it was released in `68.
And as far as the LP's I haven't taken them all out of their crates yet, so dunno what there is.
I just know the first tape I tried was marked Judy Collins: Colors of the Day and if you play it
at 3-3/4 you can plainly hear the carrier wave.

After some ditzing, my old JVC demodulator decided it could see the carrier wave too, the RADAR
lights up and you hear material from the rear.Not having handy access to the stereo LP I couldn't
compare it side-by-side, but I know it's one of those that decodes real good in QS, so who knows.

I just remember the Lucky Man was the first record we played on it while taping and at real-time speed
(for taking a level, and then taking the level down for 16-RPM) the synth at the end swirls around in circles.

 

[Doug G.]

So that means the playback head has response out to...say...40-45kHz?

Doug

 

[BananaSlug]

You could always transfer them to digital at half speed first, that would preserve the carrier wave.

 

[ndiamone]

So that means the playback head has response out to...say...40-45kHz?

Spec-sheet says 28K at 15 IPS but that's with the normal tape that was available at the time.
Like I said, this is LPR 35 CR chromium tape, so when you record with it on normal bias, the
resulting recording is excessively bright-sounding, like recording with Dolby NR and playing back
without. So it would not surprise me if the playback frequency response went out to 35K, with
chrome tape i.e. enough to pick up the carrier wave and get SOME if not all modulation out of it.

I'm guessing that since CD-4 cuts off the base-band with a low-pass filter at 20K or lower
as stereo FM does so as to have program material not interfere with the carrier wave, that
in actual practice you'd only need 7-10 KHz above and 7-10 KHz below the 30 KHz carrier
wave, so maybe all you'd need is a top end of 35K or 37K to be able to get carrier playback.

John French Magnetics (he used to be a head designer for Ampex here in California)
has all the tech specs and all kinds of never-used or partially-used head designs in
his lab in suburban New York City (North Jersey I think), so I'm going to write or call
him and see what he says.

You could always transfer them to digital at half speed first, that would preserve the carrier wave.

You wish. Even recording half-speed at 32/192 your then-15 KHz carrier wave would only
give you 12 samples per cycle of the carrier wave, certainly not enough to demodulate it.

You'd have to do something akin to what Jamie Howarth is doing at Plangent Process trying
to retrieve the bias tone from recorded tape and use that as a sync tone to re-lock a tape
either back up to itself to eliminate the endemic analog wow and flutter on the master, or lock
it back up to other production elements. With time-coded tape or sprocketed film that is possible
vs. trying to sync up ``wild'' elements by hand over days and days in a music-editing program
as the boys over on the Stereo Chat Board are fond of doing.

Or, record in analog onto the video track of a videotape, akin to Beta Hi-Fi, but then you'd have
to make sure that the audio-frequency modulation of whatever kind of Beta Hi-Fi you were using
didn't interfere with the 30KHz (or in the case of half-speed transfer 15 KHz) carrier wave.

 

[BananaSlug]

Even recording half-speed at 32/192 your then-15 KHz carrier wave would only
give you 12 samples per cycle of the carrier wave, certainly not enough to demodulate it.

You're the engineer, but really?

 

[Doug G.]

You could run a sweep from a signal generator into that deck, record it, and play it back to see what comes out.

The low band in CD-4 is cut off at 15kHz and then the modulating frequencies start at 20kHz.

Also, initially, JVC cut masters at 1/3 speed and later went with 1/2 speed. They also had no doubt that full speed mastering would eventually be possible with improved cutters. Of course, we never got the chance to see if that would have come about.

Anyway, it's an interesting way to get quad out of a stereo tape deck.

Doug

 

[ndiamone]

You could run a sweep from a signal generator into that deck, record it, and play it back to see what comes out.

Yeah, recording at 7-1/2 , wiping out one of the Reader's Digest box set tapes recorded at 1-7/8,
when I speed up to 15 IPS, I get a more or less fairly consistent readout at about 36KHz especially
if I record with the Dolby on and playback with it off.

The low band in CD-4 is cut off at 15kHz and then the modulating frequencies start at 20kHz.

So, it's the same as FM Stereo pretty much, which also cuts off at 15KHz to make room for the 19 KHz pilot tone.

Also, initially, JVC cut masters at 1/3 speed and later went with 1/2 speed. They also had no doubt that full speed mastering would eventually be possible with improved cutters. Of course, we never got the chance to see if that would have come about.

Which is one reason I'm trying to get a grant and do my masters in engineering on
What Would Have Happened to CD-4 if DMM Had Been Invented During Its' Tenure?
People get grants to study the mating habits of exotic animals all the time, so... .

Supposedly there's some Ortofon DMM prototype cutterheads around that were
optimized for half-speed as well as DMM at the same time.

Since former RCA CD-4 engineer Greg Bogantz has the last surviving CD-4 Modulator,
if and when we get our grant, if and when we can get access to the last half-speed
DMM lathe operating in the U.S. (at the Scientology plant in L.A.) and if and when
we can get one of these Ortofon prototype DMM/half-speed heads, maybe we end
up with something historically interesting. I have another thread here I discuss in
more detail about half speed DMM CD-4 if you're curious.

Anyway, it's an interesting way to get quad out of a stereo tape deck.

Yeah, that's what we thought. A HUUUUUUGE waste of expensive tape to be sure but interesting nonetheless.
I mean what did EE tape cost when it came out, 9.99 a reel when regular tape was 2.49?

 

[ndiamone]

Does your reel-to-reel tape recorder have 3 heads (play, record, erase), or does it have a combined play/record head? Play heads might not play back the 30kHz signal too well.

Four-head deck, ERP and an extra R/P for Cue and Reverse-Track monitoring.
Comes in handy for cutting apart 14-inch reels into two 10-inch for easier handling

Also, the electronics usually are designed to do equalization on what's recorded and played back.
One way to get the CD-4 signals is to make your own playback preamps and try using the record head if the play head doesn't work. Record heads are designed to handle the ultrasonic bias signal (typically 75-150 kHz) so maybe they can play back the 30kHz CD-4 signal.

I wish. That only works for constant i.e. unmodulated tone, not the frequency-modulated tone from CD-4.
The gap in a recording head is simply too wide to be able to recover bias tone.
However, the erase head that applies its' own bias to the tape prior to recording doesn't do too bad either.

Another idea is to try to connect your CD-4 demodulator to the tape head via a cable without any tape preamp.

Yes, this one does in fact have a HEAD output jack, so the MAG IN level to the demodulator would be roughly the same.

When recorded, tape equalization (NAB?)...

...Or CCIR...

was used so the playback equalization wouldn't be exactly right (not the same as RIAA).
Adjusting the carrier level and separation on the demodulator and adjusting the equaliztion should help.
Also, apparently, when the records were played while being recorded, RIAA equalization was used when
playing the records (reversed compared to when the record was cut).

Could be, dunno. Mel was kind of a purist that way. Knowing him, he'd've figured out a way to tape the LP
raw onto the tape, no RIAA decoding before the NAB/CCIR encoding for tape recording. Like I said, these
recordings are EXCEPTIONALLY bright almost to the point of being brittle-sounding, like what used to
happen in the early days of CD when they'd just use the eq'ed-for-LP master and just cut a CD from that.

I just thought it was from being Dolby-processed and recorded at 50 uS pre-emphasis instead of EE's 35
(i.e same as recording a chrome cassette at 120 instead of 70) and being recorded at half-speed besides.

I can think of more problems.
When the record was played back at half speed the equalization probably wasn't right.

Which is another reason I think you may be right about being recorded flat originally i.e. no RIAA de-emphasis.

So if you want all of the equalization to be right, get out those engineering
calculators/computers, the RIAA, etc. NAB or other tape EQ charts and figure out how to equalize
it properly, unless you want to do it by ear and maybe through the use of two stereo graphic equalizers.

Yeah. Adobe Audition can do all that. I just wish there was a such thing as a sound card with a MAG IN.
Then I could transfer EVERYTHING flat and add in and/or tweak the RIAA, NAB, CCIR, FFRR, etc etc etc curves myself.

Another problem I thought of: if the audio wasn't equalized properly on tape
which could probably be fixed with a genius IQ created electronic circuit designed based on all of
the calculations I explained already,

Like the ones in Adobe Audition or Diamond Cut Pro...

then mixing the demodulated audio from the carriers with
the F+B audio won't match exactly right to get complete separation.

Probably not, but it's still interesting to hear ANY kind of sounds from the rear speakers from
CD-4 demodulated off tape no matter HOW poor the demodulation and/or resulting separation.

Reminds me of when LaserDiscs came out right around the same time, and one of the demos
showed how far videodiscs had come in 50 years. The British had a 30-line black-and-white
videodisc system which was completely contained within the normal audio spectrum as recordable
by conventional audio mastering equipment in the 30's. These are shown being played off what
appears to be a normal 10-inch 78-RPM disc, next to an early CED Selecta Vision test from the
early 70's next to a brand new 1978 Pioneer-Philips crystal-clear AND stereophonic sound LaserDisc.

To get an idea, remember, color TV in the U.S. was only 12 years old at the time, a ``big screen TV''
was 21 inches measured diagonally, stereo sound for broadcasting would not hit for another six years
and FM-Stereo simulcasting for concerts and TV shows such as Bert Sugarman's Midnight Special was
a brand-new thing still being tested out.

Another problem: when the tape was being recorded at one speed and played back
at twice the speed it was recorded at, the equalization probably won't be the right, either.

To sum up all the EQ's:

1) playing the record at half speed EQ isn't RIAA for half speed records
2) recorded at one tape speed -> EQ used here
3) play back at twice the tape speed it was recorded at -> EQ most likely different compared to recording
4) if the audio could be made flat using a specially made EQ system then it would have to go through reverse RIAA EQ
before it goes into the CD-4 demodulator (audio EQ only)
5) the carrier level most likely wouldn't be the same relative to the audio compared to the record, but could be
adjusted using the carrier level adj on the demodulator.

Which, again, makes me think that Mel taped these straight to the HEAD IN from the CARTRIDGE OUT on the turntable
because as I've said, the recordings are EXTRAORDINARILY bright to the point of being brittle-sounding.

Another possible problem: intermodulation distortion between carriers and audio,
and maybe even the bias signal, because of non-linearities in the tape and electronics.

Another vote for the direct cartridge-to-tape-head transfer supposition.

Notwithstanding of it's technical reproductive accuracy, the whole idea that someone tried it in whatever capacity they did,
coupled with the fact that anybody 40 years later can retrieve it also in whatever capacity is interesting in its' own right.

 

[MidiMagic]

If it can record CD-4, it must have a really high bias frequency. Most recorders use 30 to 40 KHz, which would wipe out CD-4 with aliasing. It would have to be at least 45 KHz to prevent aliasing with a half-speed CD-4 recording. And the recorder must not record its own bias frequency, or that would also cause trouble.

Actually, this is not too much of a stretch when recording. The CD-4 frequencies went up top 45 KHz, and at half speed, that would be 22.5 KHz - well within the range of a good recorder. Playing it back is another issue. The playback head would have to go out to 45 KHz.

CD-4 didn't exist in 1968. It wasn't fully available until 1974.

Or is 1968 the date of the recorder manufacture.

 

[ndiamone]

If it can record CD-4, it must have a really high bias frequency.

It could record CD-4 at 16-RPM the same way it was cut on disc in the first place.
We haven't figured out yet if it was a non-pre-amped cartridge-out to direct-head-in arrangement
to avoid all the various pre-empases noted above or if it was simply a pre-amped-but-flat
(non-RIAA decoded) signal fed to the tape deck at 7-1/2 IPS for playback at 15, but it sounds
like both SciFi and Doug G. have that impression.

Most recorders use 30 to 40 KHz, which would wipe out CD-4 with aliasing.
It would have to be at least 45 KHz to prevent aliasing with a half-speed CD-4 recording.
And the recorder must not record its own bias frequency, or that would also cause trouble.

Like I said above, all I know is I hear rear-channel information out of the demodulator when demodulating
off a 15-IPS chrome reel recorded at half-speed from a linear tracking turntable. The radar lights up the same
as if it were an LP and the somewhat-fuzzy rear channels activate.

Actually, this is not too much of a stretch when recording.
The CD-4 frequencies went up to 45 KHz, and at half speed, that would be 22.5 KHz - well within the range
of a good recorder. Playing it back is another issue. The playback head would have to go out to 45 KHz.

Well it says here on the AES whitepaper that although the carrier wave modulation does in fact go out to 45K or so,
since it has 10K to work with from the start of modulation at 20K to the carrier wave center of 30K, technically you'd
only need another 6 or 7K above that to get a decent signal, and like I said above, we can get 35K or 36K fairly reliably
out of it by recording a tone generator at half speed and checking it on a waveform monitor.

Of course it falls off fairly rapidly as the heads clog with this old tape, so you might only get 4 or 5 minutes at a stretch
for transferring or playback, but as I said, it's still interesting in its own right that it works at all.

CD-4 didn't exist in 1968, it wasn't fully available until 1974. or is 1968 the date of the recorder manufacture.

The date of the recorder according to Web sources.