1 (edited by dir 2020-03-26 08:25:46)

Topic: Digital recording from Mini Disc deck

I am engaged in transferring from MD archive to the hard drive. Sony's MD-player is connected to RME via optic cable. The ADI-2's interface describe the incoming SPDIF signal as 24/44. I made a USB recording to my laptop in the same dimension 24/44 using Wavelab and ASIO. ADI's clock was set from the deck (SP opt). The 24/44 files obtained in this way, during playback from laptop, shows 18 bits only (via DIGIChek).
When I do the same except clock (switch to Internal), the recording files shows the full 24-bit scale on DIGIChek already. Of course, the pairs of rips do not match.

I know about different generations of MD recorders used 20- and 24-bit algorithm for recording both, but I never received enough information, what happens on the digital output from that decks. And how to make the correct bit-perfect copy from the mini-disc today, when MD data encoded between 16 and 24 bits? As I understand it, the first way to transfer is "correct." But where do the extra bits come from in the second method?

2

Re: Digital recording from Mini Disc deck

First of all: did you disable SRC for the SPDIF input?

Regards
Matthias Carstens
RME

3 (edited by KaiS 2020-03-26 07:32:15)

Re: Digital recording from Mini Disc deck

Method one, SPDIF as master clock is correct, with:
Setup/SPDIF.../SRC set to "OFF", as Matthias mentioned.
SRC outputs 24bit, no matter what the input is, and the files will be different every run.


Test:
Do the same rip twice.
Then bit-compare the two files (phase reverse file B, sample-precise align both file's, then bounce/mix the tracks.
If the result is null, your ripping is Bit-perfect.

Every kind of error will show up this way.
For transfer from DAT I do this on a regular base.

Re: Digital recording from Mini Disc deck

Yes, i never touch SRC.

5

Re: Digital recording from Mini Disc deck

If nothing changes but the clock, and this clock is incorrect (MD has to be master, right?), then the 'extra bits' are just distortion/artefacts.

Regards
Matthias Carstens
RME

Re: Digital recording from Mini Disc deck

So what happens technically when I set master clock to Int ? Why ADI's display doesnt't shows true 18 bit?
I use Sony MDS-JE640 deck as transport. Anybody knows about other models - any chance to get a more than 18 bit from MD source?
How accurately such a type of stream with 6 empty bits will be decoded by conventional DAC's?

7 (edited by vinark 2020-03-26 12:26:07)

Re: Digital recording from Mini Disc deck

I would suspect that minidisc uses 24 bit conversion at recording to create 18 bit files at playback. What is recorded is a question (and data, not audio) since MD is a lossy format. The bitrate of ATRAC  compression is 292 kbps AFAIK.

How accurately such a type of stream with 6 empty bits will be decoded by conventional DAC's?
No problem. And when you think of it these last 6 bits only influence bit 7 which is already at -108db (18X6db).
There is no way to know what Sony does but no DAC that plays 24 bit (or even only 16 bit, it will just truncate at 16bit) will have any issues playing them.

Vincent, Amsterdam
https://soundcloud.com/thesecretworld
Babyface pro fs, HDSP9652+ADI-8AE, HDSP9632

8 (edited by dir 2020-03-26 14:10:38)

Re: Digital recording from Mini Disc deck

This is not my music archive, so I don’t know for sure the list of equipment that recorded and dubbed these mini-discs. Presumably everything was done on portable recorders. I began to put on DIGIChek the rest of the records again. Some tracks are 18-bit, and some are 24-bit. The strangest thing is that it happens even on the same track. I switched the digicheck to the input of the spdif from the playback MD. Most often there are 18 bits, but sometimes a full scale of 24 bits is turned on.
I understand that a mini-disc is a low bitrate lossy format, but I would not want to aggravate problems with a truncate. From a mathematical point of view, truncate has a tiny level of errors, but nevertheless it negatively affects to sound of the master record and we would like to avoid it.

Re: Digital recording from Mini Disc deck

Maybe the bit depth is dynamic. Did you ever get different results for different location of the same song *like the fade out.
Still if you hear nothing strange on the fade out of a song, here errors would be exposed, I am sure you will not loose anything.
And of course if some are 18 bit and some 24 bit repeatable, then that is what it is. No worries then.

Vincent, Amsterdam
https://soundcloud.com/thesecretworld
Babyface pro fs, HDSP9652+ADI-8AE, HDSP9632

10

Re: Digital recording from Mini Disc deck

As I understand it, the recording bitrate does not depend on the settings of the clock. The situation described in the first post was a pure coincidence. I can get (with different probability) an 18-bit , and a 24-bit sample from same track & settings. Which one record should I choose for archiving? Do I need to treat a dither for such audio, when I go to downsample for 16 bit releases? I would not want to do this by ear.

11 (edited by KaiS 2020-03-26 22:39:04)

Re: Digital recording from Mini Disc deck

Don't let yourself confuse, it's simple.
If you get 18bit in a bit-exact transfer (test see above) that's what's stored in the file - basta.

Archive it like it is.

For the 16bit release version use dither to make the 2 extra bit available, if you truncate them they're gone.

I prefer PowR2 as dither scheme, but even simple triangular dither does the job very well.

Just keep in mind:
perform intersteps like mastering and level changes at highest available bitdepth (not "bitrate" that's something different), 16bit dither only in the final output step.

I guess you already know that.


REMARK:
Every slightest change in your original 18bit file will activate the lower 6bits, simply as result of a calculation.
This does not mean your audio content has gained more information, it's just the extended precision of divisions and multiplications that fills the bits.

12

Re: Digital recording from Mini Disc deck

Thank you for you attention to my issues. I was put 24bit and 18bit rips into iZotope. As I said, for these test I didn't change any settings on RME. Its just making like Sony decide.
18bit rip has a maximum peaks on 0dB, despite 24bit rip has a maximum at slightly lower -0,04dB and minimum slightly higher. So it seems like 18bit was 'more autentic' stream.

13 (edited by KaiS 2020-03-27 00:20:34)

Re: Digital recording from Mini Disc deck

There is no more or less authentic.
It's either authentic or not.

The 24bit is simply the result of SRC processing or Clock errors.
18bit is the correct one.

Try the bit compare test - the 24bit will be different in every run.
The 18bit should be always the same.

Looking at the levels isn't an evidence for either version.

14

Re: Digital recording from Mini Disc deck

I guess you're right. But I never knows about 18bit audio on MD digital out. No one was told about it.

15 (edited by KaiS 2020-03-27 13:09:15)

Re: Digital recording from Mini Disc deck

Why not - over the times we had all types of AD-converters.
18bit was the 1st step up from the standard 16bit of that time.

I guess Sony used it as a marketing argument.
"Better then CD" was quite common for anything that had the slightest notch up in the numbers anywhere.

In fact I didn't like the practical audio quality of the ATRAC data reduction system too much.
But the biggest culprit was it's low reliability, not better then DAT.

16 (edited by KaiS 2020-03-28 20:47:30)

Re: Digital recording from Mini Disc deck

When using a hex editor - did you remember to skip the file header?

For me it's easier to use 2 tracks of a DAW and phase reverse one file for substraction.
The result gives a hint what's different, as you can listen to it.

Maybe it's just some very low level information, like dither that's different and could be ignored.

Clock skips are obvious this way too.


One more idea:
Maybe the 18bits output of the MD player are fake, and the 2 lower bits just filled with noise.
What if you record or truncate to 16bit?

17 (edited by dir 2020-03-28 21:07:03)

Re: Digital recording from Mini Disc deck

I was delete previous message, because I was made more rips. Plus i include to experiment yet one PCM-recorder for Sony deck. Finally I got bit-identical 5-sec samples on both.