Working CD-4 (software) Demodulator!

[Jerfo]

Jerfo:

Please post the exact software name and version number of the Mac emulator/virtual machine
and also please the version number of Windows 10 you're using.


(this looks like my most promising option for archiving my CD-4 discs using my Windows 10 PC)


Although extended (phono cartridge) frequency response (to 45kHz) is important for CD-4 decoding,
channel separation in the 20kHz to 45kHz range needs to be at least 12dB and preferably 20dB.

Kirk Bayne

The software that I used was Oracle's VirtualBox, which is free to download and use. You can grab it from here: https://www.virtualbox.org/wiki/Downloads. This is also do-able using the VMware Player or Workstation, if you prefer VMware.

I'm running the April 2018 build of Windows 10 (10.0.17134.345), although VirtualBox should run similarly on anything dating back to Windows 7.

I used the process detailed here to build the macOS virtual machine: https://techsviewer.com/install-macos-high-sierra-virtualbox-windows/ . Please note that this enters into a legal gray area, as Apple really doesn't want their OS to run on anything other than Apple-brand hardware.

 

[Jerfo]

Quick update: I just picked up some more CD-4 LPs (two Tomita LPs and Frank Zappa's Over-Nite Sensation) and figured that I'd give this software another shot. I tried the FZ tune "50/50," which seems likely to have some interesting movement happening in the mix and should be a good test candidate.

Once again, the software complained that no RADAR was detected. There's an option to decode only the CD-4 subcarrier to the rears for troubleshooting purposes, so I tried that. Sure enough, the resulting file had no sound in the rear speakers. Yet, when I run the same rip through SDR#, the demodulated subcarrier is completely audible. I'll keep goofing around with it, but right now I'm at a loss.

I can post more sound clips and pictures if anybody is curious. I'm also re-doubling my efforts to get a hardware CD-4 demodulator. Fingers crossed that I'll find one soon...

 

[fredblue]

Quick update: I just picked up some more CD-4 LPs (two Tomita LPs and Frank Zappa's Over-Nite Sensation) and figured that I'd give this software another shot. I tried the FZ tune "50/50," which seems likely to have some interesting movement happening in the mix and should be a good test candidate.

Once again, the software complained that no RADAR was detected. There's an option to decode only the CD-4 subcarrier to the rears for troubleshooting purposes, so I tried that. Sure enough, the resulting file had no sound in the rear speakers. Yet, when I run the same rip through SDR#, the demodulated subcarrier is completely audible. I'll keep goofing around with it, but right now I'm at a loss.

I can post more sound clips and pictures if anybody is curious. I'm also re-doubling my efforts to get a hardware CD-4 demodulator. Fingers crossed that I'll find one soon...

fwiw (and only imho) the Tomita's should be ok for CD-4 testing purposes, with decent enough pressings and active mixes so you should be able to figure out if its demodulating alright.. but the Zappa Quadradiscs aren't the best pressings, nor are they the best Quad mixes by any stretch (I'm being kind, OverNite Sensation I found to be rather a poor mix tbh) anyhoo, just my twopennuth

 

[Brice]

Quick update: I just picked up some more CD-4 LPs (two Tomita LPs and Frank Zappa's Over-Nite Sensation) and figured that I'd give this software another shot. I tried the FZ tune "50/50," which seems likely to have some interesting movement happening in the mix and should be a good test candidate.

Once again, the software complained that no RADAR was detected. There's an option to decode only the CD-4 subcarrier to the rears for troubleshooting purposes, so I tried that. Sure enough, the resulting file had no sound in the rear speakers. Yet, when I run the same rip through SDR#, the demodulated subcarrier is completely audible. I'll keep goofing around with it, but right now I'm at a loss.

I can post more sound clips and pictures if anybody is curious. I'm also re-doubling my efforts to get a hardware CD-4 demodulator. Fingers crossed that I'll find one soon...

Hi Jerfo,

The Stereo Lab software will say "No RADAR" when the level of the subcarrier is below -30dBFS. That's a very low level and suggests something is wrong. But we can lower it (I just chose -30dBFS as a level at which we might get problems.) Are you setting the overall level of the rips a bit low? If so, that might be the cause. You are not using a Shibata stylus, so the carrier will be below the usual value.

Even with no RADAR, the software will try to make something of what it receives. What are you hearing as a decode?

Finally, the option to output the subcarriers to rear does just that, the L/R signal is just the 30kHz signals, so you certainly won't hear anything. It's for diagnosis. If you could cut a short section of the subcarrier to rear as a .wav, I can take a look and see what's up. Please send via Psptial Audio tech support: [email protected]

Best wishes,

Richard

 

[Jerfo]

Hi Jerfo,

The Stereo Lab software will say "No RADAR" when the level of the subcarrier is below -30dBFS. That's a very low level and suggests something is wrong. But we can lower it (I just chose -30dBFS as a level at which we might get problems.) Are you setting the overall level of the rips a bit low? If so, that might be the cause. You are not using a Shibata stylus, so the carrier will be below the usual value.

Even with no RADAR, the software will try to make something of what it receives. What are you hearing as a decode?

Finally, the option to output the subcarriers to rear does just that, the L/R signal is just the 30kHz signals, so you certainly won't hear anything. It's for diagnosis. If you could cut a short section of the subcarrier to rear as a .wav, I can take a look and see what's up. Please send via Psptial Audio tech support: [email protected]

Best wishes,

Richard

Ha! I didn't realize that the output to subcarrier option didn't also demodulate the subcarrier. My hearing is OK, but not superhuman.

I normalized the rip that I made of the FZ tune and then ran it through Reaper's built-in spectrograph plugin. The subcarrier is clearly there, but it looks like it's -40dB at best. Look at the bottom right corner of the screenshot.

Here's a link to the raw unconverted rip of the first several minutes of the song: https://drive.google.com/open?id=1feqR-yJE9Keja3dNJhZ-asGO5qBgmIR5 , in case you want to look at it yourself.

Here's a link to the subcarrier-only decode: https://drive.google.com/open?id=1xplPqfBfPMgiqhcmmY59rKYIfkpr5lkj . The rear, center, and sub channels are all silent. Looking the the spectrogram of the front channels, there's a very faint (-100dB) signal remaining at 30kHz.

I'm happy to admit defeat and accept the fact that I'll need to purchase a shibata stylus in order to make this work. But, if there's anything that you can do to make this work, it would be greatly appreciated!

I'll give this another try with one of the Tomita records, since it sounds like I picked a dud with the FZ record. I'll let you know if I have any better luck.

 

[Brice]

Ha! I didn't realize that the output to subcarrier option didn't also demodulate the subcarrier. My hearing is OK, but not superhuman.

I normalized the rip that I made of the FZ tune and then ran it through Reaper's built-in spectrograph plugin. The subcarrier is clearly there, but it looks like it's -40dB at best. Look at the bottom right corner of the screenshot.

View attachment 36120

Here's a link to the raw unconverted rip of the first several minutes of the song: https://drive.google.com/open?id=1feqR-yJE9Keja3dNJhZ-asGO5qBgmIR5 , in case you want to look at it yourself.

Here's a link to the subcarrier-only decode: https://drive.google.com/open?id=1xplPqfBfPMgiqhcmmY59rKYIfkpr5lkj . The rear, center, and sub channels are all silent. Looking the the spectrogram of the front channels, there's a very faint (-100dB) signal remaining at 30kHz.

I'm happy to admit defeat and accept the fact that I'll need to purchase a shibata stylus in order to make this work. But, if there's anything that you can do to make this work, it would be greatly appreciated!

I'll give this another try with one of the Tomita records, since it sounds like I picked a dud with the FZ record. I'll let you know if I have any better luck.

Hi Jerfo,

I looked at the original track and bypassed a load of processing so I could look at just the subcarrier part of the signal. This is - in effect - the output of the subcarrier filter in Stereo Lab. I have taken a screen-grab of the resultng "audio" in Adobe Audition (it's at the foot of this post). As you can see, the subcarrier is there, and just below my -30dB threshold. So, I can see why this is getting flagged as "No RADAR".

To give you an idea the subcarrier level is 6 times lower (-15dB) below the recorded level. This does suggest a real problem. I suggest this is a cartridge and (possibly over-capacitive) lead/pre-amp problem.

However, if I boost the subcarrier level by 20dB and run an RCA Square test on it (second screen grab). It does show that it's actually in pretty good shape. Just very low.
.

So, before you rush out and buy a Shibata stylus, the thing you could try is to run a high-shelf boost from about 15kHz upwards on the original needle-drop. It appears to need about 20dB shelving boost. So as not to have this clip, the whole file needs to be reduced by about 6dB prior to the filtering. You can even use the FFT filters in Audacity to do this.

I've tried it here and this gets the needle-drop over the threshold. There's certainly plenty of quad effect with this done and all the extra boost doesn't seem to have upset the quality too much.

Longer term, I can look at lowering the threshold in the software. That just needs a proper release and that takes time.

Best wishes,

Richard

 

[proufo]

I have always believed that if your stylus is a line-contact type you should be able to adjust the stylus rake angle (both for stereo and CD-4) to perfection by measuring the carrier level as you adjust the SRA. Highest level (or cleanest tracing perhaps) should correspond to optimal SRA.

 

[Brice]

I have always believed that if your stylus is a line-contact type you should be able to adjust the stylus rake angle (both for stereo and CD-4) to perfection by measuring the carrier level as you adjust the SRA. Highest level (or cleanest tracing perhaps) should correspond to optimal SRA.

That's absolutely true. Moreover, I believe there are good reasons NOT to play CD-4 records with anything but a Shibata stylus (at least, not too many times) for the reasons given here: http://pspatialaudio.com/CD-4 wear.htm

Nevertheless, I wanted to let Jerfo get something out of his experiements. I'm not suggesting that Shibata and an appropriate cartridge and pickup aren't desirable.

Jerfo, Here is a screen grab of the subcarrier uplift filter I tried on your track.



Any DAW will offer this type of function. Be sure to use a non-causal FFT filter. A normal HF lift EQ will "mess" with the phase of the lower sidebands.

ATB,

Richard

 

[Jerfo]

Regarding a Shibata stylus: I'm well aware of the potential dangers of using anything other than a Shibata / microline stylus on a CD-4 record. I do have a spare p-mount turntable, and I'm hoping to find a reasonable Shibata stylus / cart combo to use on it. If anybody has any recommendations on that, fire away. I was just hoping to be able to temporarily get away with what is otherwise a nice stylus (Ortofon 2M Blue) with low tracking force; I run mine at 1.5g, which is at the low end of what's recommended on it.

The good news is that subcarrier boost filter worked! The software was able to detect the subcarrier after that and render what I assume is a much more accurate result. That said, there's still an "AM radio" quality to everything that I assume has to do with my recording setup and not the software. Before I pass final judgement, I'll see if I can modify a spare phono preamp to not do RIAA EQ and use that instead of my audio interface to get everything to line level.

I'll post some samples later so that you guys can hear the results.

 

[Jerfo]

Quick update: I picked up a CD-4 demodulator (a JVC 4DD-5) off of eBay and it works perfectly with my Ortofon 2M Blue stylus. I've had nothing but a good solid RADAR light so far, and the sound quality (at least on the Tomita "Snowflakes" LP) is outstanding.

At this point, I think that my difficulties with getting the software decode to work were due to impedance issues going from my turntable directly to a sound interface. I'll re-test the software once I've had a chance to do a proper rip to my PC using a flat phono preamp.

 

[kfbkfb]

Jerfo:

Have you had a chance to compare the output of the 4DD-5 and the Stereo Lab software decoder
when playing the same CD-4 disc (presuming a suitable non-EQ needle drop is provided to the
Stereo Lab decoder)?

Kirk Bayne

 

[Jerfo]

Jerfo:

Have you had a chance to compare the output of the 4DD-5 and the Stereo Lab software decoder
when playing the same CD-4 disc (presuming a suitable non-EQ needle drop is provided to the
Stereo Lab decoder)?

Kirk Bayne

I have. There's no real comparison using my current setup - the hardware demodulator sounds significantly better. I don't blame the software, though. My RIAA-less rips are clearly not good enough; even a stereo rip sounds muddy through any software RIAA EQ, not just theirs. Between that and the fact that I had to boost up the subcarrier level significantly, the overall quality is suffering. The decode does sound accurate, so I think that people who have a setup that allows for a good quality flat rip will have good results. Maybe one of these days I'll mod an old phono preamp to bypass the RIAA stage (which should lead to better flat rips), but I don't think I'll have the time to do it in the near future.

 

[Mendemender]

As the Pspatial Audio CD-4 decoder is being discussed here, I thought it was time to join the forum because I am responsible for the decoding algorithm development. Let me know if there are any questions I can (try to) answer.

I was really pleased to see that there might finally be a software decoder for CD4. I wish you every success in getting it working well. There is one thing I don’t understand however. If I read your description correctly you are implementing an FM demodulator. However according to the technical spec. of the CD4 system that I have, the modulation is a mixture of FM and phase modulation. On page 3 of the manual I have (which I assume is the same one you have) it states that “Regarding the difference signal, the ranges lower then (sic) 80 Hz and higher than 6kHz are frequency modulated (FM), and a range between 800Hz and 6kHz is phase modulated (PM)” (I assume that either 80Hz or 800Hz is a misprint). How is the range 80Hz to 800Hz coded? How does your demodulator cope with the different types of modulation?

 

[Mendemender]

Spurred on by the prospect of getting a working decoder, I finally unpacked my AT-OC9ML/II and installed it. Unfortunately I'm still working on building a decent MC preamp but I can get some sort of needle drop using a flat MM preamp and turning up the volume on the A/D converter. I expect noise and hum will be high. Also my collection of records is of unknown provenance and they probably need a good clean. However if I can help with any unequalized 96kHz (or any other frequency) needle drops please let me know. I can offer various things such as Red Octopus, Spitfire, Dragon Fly, Tubular Bells, 461 Ocean Boulevard, Apostrophe, Overnite Sensation, Machine Head, On The Border, One of These Nights and a few others, but not all at once

 

[proufo]

What I have (All You've Ever Wanted to Know about the CD-4 Disc System, probably a later edition of what you have) says:

"Modulation, although usually referred to as FM, is actually a combination of FM and PM, taking advantage of the merits of both methods. Below 800Hz, frequency modulation is used, phase modulation takes over from 800Hz to 6,000Hz, to be replaced again by FM above 6kHz. This combination of modulations in fact offers the advantage of achieving even wider dynamic range."

A can of worms.

Thanks Mendemender.

 

[Mendemender]

Thanks for that proufo. That helps to explain it.

For anyone who wants to play, here's part of Red Octopus.
Red Octopus, side A
It's quite a large file, over 300MB. Unfortunately no MC preamp at present so it will probably be a bit noisy but there's plenty of sub-carrier there.

 

[Brice]

I was really pleased to see that there might finally be a software decoder for CD4. I wish you every success in getting it working well. There is one thing I don’t understand however. If I read your description correctly you are implementing an FM demodulator. However according to the technical spec. of the CD4 system that I have, the modulation is a mixture of FM and phase modulation. On page 3 of the manual I have (which I assume is the same one you have) it states that “Regarding the difference signal, the ranges lower then (sic) 80 Hz and higher than 6kHz are frequency modulated (FM), and a range between 800Hz and 6kHz is phase modulated (PM)” (I assume that either 80Hz or 800Hz is a misprint). How is the range 80Hz to 800Hz coded? How does your demodulator cope with the different types of modulation?

Hi,

The frequency/ phase modulation copy in the CD-4 brochures is just marketing hype really.

Phase modulation and frequency modulation are close cousins. Frequency modulation is the rate of phase modulation. In other words, the amount of frequency modulation is the time derivative of the phase modulation (f = dPHI/dt). This means that FM and PM are really the same thing except for a frequency response issue. All the "phase modulation in the range between 800Hz and 6kHz" means is that is the region of audio which is on the rising part of the pre-emphasis curve (and thus has a frequency response which rises at 6dB/octave). This is all "un-done" by the de-emphasis filter in the decoder.

You could equally say that FM radio broadcasts are phase modulated in the audio region above 3kHz because that is the region in which the audio is boosted 6dB/octave prior to modulating the carrier.

Best,

Richard

 

[Mendemender]

Hi Richard,

many thanks for the explanation. It’s a while since I learned about A cos(ωt + φ) and all that but now you mention it I remember the relation between FM and PM although I can't remember the exact details.

You could equally say that FM radio broadcasts are phase modulated in the audio region above 3kHz because that is the region in which the audio is boosted 6dB/octave prior to modulating the carrier.

I need to go away and think about that First I need to get MacOS running in Virtualbox so that I can try out the software.

Thanks
Stuart

 

[kfbkfb]

So...

Anyone else tried the Pspatial Audio CD-4 Decoder?

What Phono Cartridge/Stylus Shape are you using?

I'm upgrading my Record Playback Equipment, I've used my Shure V-15 Type III since 1976-08 (it does a fair job
of CD-4 decoding with my SH-400), I think it's time for a new cartridge and I'm considering the AT VM540ML.

https://www.audio-technica.com/cms/...8b64/index.html#prPaContainere7a3d4bc8b248b64
^^^
Audio Technica thinks the VM540ML will work for CD-4 (Q&A section).

Kirk Bayne

 

[winopener]

Trying to set up a mac VM... AT 540ML is a small upgrade of the 440ML that works perfectly (here... and everywhere on the QQ members' houses) for CD4 decoding.