1/2 speed remastering CD-4 records

[Groovy Daniel]

Hello all. I'd like to play my CD-4 vinyl at 1/2 speed, capture the playback in my DAW through the mic input on my interface, and then double-speed and apply RIAA equalization digitally. The end game of my scheme is to capture each channel digitally, remove noise and disc anomalies within the digital realm, and burn the recording to either a multichannel DVD-A or DTS-encoded CD. My thought is, at 1/2 speed playback, the phono cartridge will provide excellent traceability of the sum signal and especially the carrier signal. The obvious problem is I have no means of demodulating the 1/2 speed carrier at 15kHz. I can think of three ways to accomplish the demodulation, but I lack the technical tools to do so. (1) I could perform demodulation via software. But to the best of my knowledge no software exists. (2) I could tweak an outboard CD-4 demodulator so all the circuitry is tuned an octave lower, and perform the demodulation while recording to my DAW. But I don't have the technical skills to perform such a modification to a demodulator. (3) I could double-speed the sum and modulation signal (i.e., the signal as originally cut to the CD-4 LP) in the digital realm after recording, then burn the signal to some medium capable of capturing and playing (in the analog realm) a 30kHz signal, and then play the analog recording through an outboard CD-4 demodulator. But I know of no medium capable of reproducing a 30k signal except the CD-4 disc. Thoughts?


Sent from my iPhone using Tapatalk

 

[ArmyOfQuad]

Stereo dvd-a could do 192kHz or 96kHz, which is capable of holding a 30k signal. However, you'll need to be careful not to send too high a level into the demodulator.

 

[Groovy Daniel]

Hello ArmyOfQuad. DVD-A can sample at a high frequency as you suggest, but it can't output in analog anything above 20k or so. So I don't believe your answer is a solution.


Sent from my iPhone using Tapatalk

 

[ArmyOfQuad]

I know of no such restriction, if dvd-a couldn't output over 20k, that would make the high sampling rates pointless.

Alternately, you could just playback your files off a computer, if you have a high enough quality sound card for computer noise to not be an issue, since you'd need to playback at a low level.

To be honest, I'm a bit doubtful this process will gain any quality in cd4 performance...but I certainly am curious what results you get if you find a way to do this.

 

[Groovy Daniel]

The purpose of the high sampling rate is to better capture the audible frequency range, not to output supersonic frequencies. Such frequencies would wreak havoc on downstream audio equipment, not to mention loudspeakers, and so are filtered from the output.


Sent from my iPhone using Tapatalk

 

[proufo]

If you record at half speed and 192k and convert to 96k I believe you'll have all the frequencies contained in the CD-4 record. But you can't feed that to a demod, as they take phono signals.

You could modify the demod so you feed the DVD-A/AVCHD player's line-level output signals at a point where it would be equivalent to what the earlier stages of the demod would deliver, probably with some reduction in level.

You could continue the demodulating process in the digital domain (in the PC) but then you'll probably get stuck at ANRS decoding.

You could beg the Involve Audio people to include a digital input in their demodulator so you can feed your digital recordings directly to their demod. Their demod could have an analog box and a digital box, the latter with such digital input.

 

[proufo]

DVD-A can sample at a high frequency as you suggest, but it can't output in analog anything above 20k or so. So I don't believe your answer is a solution.

First time I hear this.

Such frequencies would wreak havoc on downstream audio equipment, not to mention loudspeakers, and so are filtered from the output.

First time I hear this.

Could you please elaborate?

 

[proufo]

To be honest, I'm a bit doubtful this process will gain any quality in cd4 performance...but I certainly am curious what results you get if you find a way to do this.

Hello AoQ.

This scheme is interesting as it avoids all the hassle regarding special cartridges and a second turntable in some/most cases. Of course getting a regular turntable to play at half-speed is a mystery, although I'm sure there are people that can do it easily. In my case it would involve either a second motor (seems the easiest way) or adding a couple of pulleys and doing a lot of things to get the feedback circuitry of the motor controller to believe the platter is turning at 33 rpm.

A damper for the tonearm may be needed, as low frequencies may fall in the resonant band of the tonearm-cartridge assembly.

It might be more elegant to record with no RIAA at the phono preamp. But I guess it is certainly acceptable to undo and redo RIAA digitally.

Afterwards I believe the best way to continue is to produce two AM channels and two FM channels. But other options could be the same or better.

At one point someone will find a way to do ANRS in a PC. And then we'll be happy. Wouldn't it be nice.

 

[Groovy Daniel]

Actually, the speed was easy but inelegant. I have a journeyman Technics belt drive table. I simply sanded down the drive shaft like a lathe until I hit 16 2/3 on a strobe disk.


Sent from my iPhone using Tapatalk

 

[winopener]

IIRC someone else in the past already did the option n.3, record a CD4 lp at 96k stereo realtime, placing the file on a DVDa then play it back on a modified demod.
Going the half-speed way is not difficult for the speed itself - you can find DJ turntables that have a +- 50% slider - but for the eq that change. Probably a riaa-less recording is easier to work in digital.

 

[proufo]

Actually, the speed was easy but inelegant. I have a journeyman Technics belt drive table. I simply sanded down the drive shaft like a lathe until I hit 16 2/3 on a strobe disk.

How far have you gotten re the whole procedure>?

 

[Groovy Daniel]

I'm ready to roll. The consensus is that I should be able to play back 30k on DVD-A. So I'll try it.


Sent from my iPhone using Tapatalk

 

[Jim the Oldbie]

The purpose of the high sampling rate is to better capture the audible frequency range, not to output supersonic frequencies. Such frequencies would wreak havoc on downstream audio equipment, not to mention loudspeakers, and so are filtered from the output.

Sorry, but this is incorrect. Any properly-designed 96K soundcard or other digital audio device will record and play back frequencies right up to near 48KHz with little or no filtering or rolloff. (My little Focusrite Scarlett 2i2 can do this just fine.) It is possible for ultrasonic output to cause issues downstream as you mention, but that's a problem for the downstream gear, not the sound card.

Your original proposal of half-speed CD-4 playback to improve hi-freq. tracking sounds interesting. Perhaps you could do it in this order:

Record (digitize) half-speed vinyl playback at 48K sample rate, direct to your DAW, no RIAA EQ, as you've described

Double the sample rate to 96K in software (Audacity could do this if your DAW doesn't have this capability)

Apply RIAA EQ (Audacity again?)

Demodulate CD-4 info

That last bit I know not much about, having never owned any CD-4 gear. There may be software solutions for this? If not, the 2-channel 96K track (now at proper speed and freq. range) could be played back at this point into an analog CD-4 demodulator, assuming RIAA EQ could be bypassed since it's already done? Or if none of the hardware demods has a way to bypass RIAA EQ, perhaps you could just skip that part in software and play the un-EQ'd file into the analog demodulator. You'd have to provide some sort of attenuation to phono cartridge levels so as not to overload the input, but that wouldn't be too difficult. But then you'd have to do a multitrack recording of the demod ouput.

This is a lot of steps, not to mention a possible extra A to D conversion. But most of it could be done with very low "loss." I would imagine the weakest link here is the ultrasonic CD-4 "track" itself, so if your half-speed idea yields any solid improvement in accurately pulling that off the vinyl, it could well offset the other stuff and leave a net improvement in sound quality.

Let us know how you get on if you have time.

-- Jim

[EDIT] Pablo makes an important point about tonearm resonance coming into play at these reduced frequencies. That will be something to watch for.

[EDIT 2] When I mention doubling the sample rate to 96K above, you'd want to make sure that whatever you're using will simply change the sample rate with no resampling, of course.

 

[Groovy Daniel]

So here's where I'm disconnecting with the advice. No prob getting the sound digitized into Adobe Audition and doubled up to real time with the 30k carrier intact. Essentially what I'd have at this point is a digital copy of the contents of the LP (no RIAA eq, just like the LP), only hopefully cleaner because I should have a better copy of the sub carrier and I'll also have done digital noise reduction and tick/pop removal. Now here's the sticky wicket. I have to get that signal out of my DAW through my m audio 610 interface into the analog realm so I can run it through an analog CD-4 demodulator. The only analog output in the m audio 610 is headphones. Are you suggesting that the headphone output will pass analog frequencies up to half the sampling rate, which at 96k would be 48k? My contention is that the digital to analog conversion would not allow content at that frequency to pass. Do you believe otherwise? I'm happy to be wrong in this instance.


Sent from my iPhone using Tapatalk

 

[Jim the Oldbie]

According to the specs on the M-Audio site, the headphone outputs are only rated to 22KHz. But this unit also shows eight TRS-balanced line outputs on the rear panel, which are rated up to 80Khz at 192K sample rate.

 

[proufo]

Record (digitize) half-speed vinyl playback at 48K sample rate, direct to your DAW, no RIAA EQ, as you've described

The range of the FM signals in CD-4 records is 30kHz -10kHz +15 kHz. A sampling rate of 48k will not do. 96k could be too close. I believe 192k and time-compressing to 96k may be the best approach.

 

[proufo]

... I should have a better copy of the sub carrier and I'll also have done digital noise reduction and tick/pop removal.

Maybe it is possible to preprocess the FM signals and make them better. For example, maybe there is a software library somewhere that would allow to homogenize the carrier exactly to 30kHz and its level.

 

[proufo]

The purpose of the high sampling rate is to better capture the audible frequency range, not to output supersonic frequencies. Such frequencies would wreak havoc on downstream audio equipment, not to mention loudspeakers, and so are filtered from the output.

Thought about this and I wonder if some power amplifier designs that use negative feedback would be affected by signals beyond a certain frequency.

 

[Groovy Daniel]

Hot diggity dog! Jim, you're right! So now I need to figure out how I can D>A and A>D simultaneously, or do I need a second interface? (I want to suck the post demodulated signal back into Audition to convert to DTS.)


Sent from my iPhone using Tapatalk

 

[ArmyOfQuad]

The range of the FM signals in CD-4 records is 30kHz -10kHz +15 kHz. A sampling rate of 48k will not do. 96k could be too close. I believe 192k and time-compressing to 96k may be the best approach.

30kHz + 15kHz puts the top range at 45kHz. But, playing it back at half speed halves it, 22.5kHz. Nyqist states we need twice the sampling rate, which brings is back to 45 for the required sampling rate. Which 48 is above.

In theory a 48kHz sampling rate is enough to capture enough detail in the carrier signal at half speed, or 96kHz at normal speed. But in practice? I don't have any audio evidence, but I do have this picture of a waveform I recorded in of a cd4 record played in realtime with a 96kHz sampling rate, which does show the carrier.