MiniDisc media differences and trivia

[R.Daneel]

Hello girls and boys! Well, mostly boys anyway...

As promised, here's the scan of an MD shootout from 1999.

http://i53.tinypic.com/ofoggk.jpg
http://i52.tinypic.com/1r3q52.jpg
http://i52.tinypic.com/14cdzx5.jpg
http://i56.tinypic.com/2hocmj8.jpg

[Eminence1963]

Hello girls and boys! Well, mostly boys anyway...

As promised, here's the scan of an MD shootout from 1999.

nice read*bigthumbup*

[Glantoir]

Thanks for posting this.Antun

[Elite-ist]

Thanks for the scanned article, Antun. This is the first comparison of MiniDisc recordings I've read. At first, I thought there wouldn't be much variation in the quality of the recordings, but it's apparent by the results there are differences in the sound quality of different brands.

Nando.

[R.Daneel]

Well, I don't think there are any significant differences between the discs and they are certainly not as great as with tapes. In the early days, Sony issued engineering norms for MD production as well as mastering procedures. Naturally, Sony DADC Austria was the first to embrace it and issued official instruction book for MD mastering. Also, the error correction system used in the MD format is a lot more advanced than in the CD format so the discs themselves are extremely reliable and cross compatibility with different brands of machines and discs as good as it gets. In the early days of CD burning on PCs you were always faced with a possibility that some discs might not work properly for who knows what reason. Minidisc never had such problems even though sound quality wise, the earliest machines to hit the market were less than spectacular and sounded very compressed and harsh.

However, just to be sure, I will test the discs I have. I doubt there'll be much difference between them but there were times I noticed differences between Sony Premium And TDK RXG discs. I don't specifically recall what the figgeences were so I will have to verify that. I won't promise anything though, it will most likly be a wild geese chase... sort of like looking for hairs on an egg.

[Scorpion8]

Hairy eggs? Not a good mental image. But thanks for the scans, I never knew BASF made MDs. Thanks!

[Wilhelm]

BASF did not make MiniDiscs. Only a very few companies made them. Although the reviewers made strong claims about sound differences, my guess is that all the versions in the test were from the same three manufacturers--Sony, Maxell, and TDK.

BASF manufactured ALL the DCC cassette tape and supplied every DCC brand with that tape. I suspect that there is some review somewhere that rated DCC cassettes "by ear" and found large differences in sound quality, too.

We bought all of our optical discs from other suppliers and even some specialty tape products such as DAT, 8mm, and Hi-8. In turn, we supplied everyone with DCC tape and almost every video tape manufacturer with chrome pigment. TDK Luxembourg was also a ferric oxide customer of ours. It was simply too expensive and too risky for each tape manufacturer to attempt to make every existing and every potential storage format.

[R.Daneel]

Hi Wilhelm and thank you for another of your informative posts!

So now we know that BASF didn't manufacture the discs themselves. It's not surprising since BASF had only one model called MD Maxima.

If there are differences in sound quality on different discs, then it would be interesting to know what produces those differences. Is it the way the magnetic particles are applied to the disc surface or something else? Does the actual disc mechanism have any influence? So much questions and no answers. I guess the format reached it's sunrise so it makes the whole thing irrelevant.

[Wilhelm]

MiniDiscs are fascinating media because they seem to be a hybrid of multiple formats. The recording layer is sputtered, similar to that of rewritable CD and DVD discs; but instead of a semi-metal alloy, the coating is terbium-ferrite-cobalt. It is magnetic, but the magnetism is not read with a magnetic head. The MiniDisc uses a process known as Magnetic Field Modulation in which a laser uses high power (6.8 milliwatts) to heat the terbium-ferrite-cobalt layer to its Curie point (185 C./365 F.) from the bottom up while a magnetic head does the recording at the surface of the disc. This process simulates that of TMD (Thermo Magnetic Duplication) used to record VHS chrome pancakes at high speed. The magnetic recording records a digital data pattern in the magnetic layer, and that pattern remains as the surface immediately cools beyond the laser.

For playback, the process uses the laser at much lower power to read the magnetic patterns according to the Kerr effect, which causes a polarization of laser light when played across different magnetic orientations. This is the process used in magneto-optical discs, but the MiniDisc was the only consumer version of this method of recording/playback.

As to why the discs sounded different, there a number of possibilities or combination of them:
1) error rates - error correction differences. ATRAC encoding and compression at 292 kilobits/second would the be same for all the discs, but if there were any scratches or other physical anomolies that caused error correction to react differently, that might be audible. Compression schemes for digital audio were very new when MiniDisc and DCC were introduced, and although both formats were failures in the consumer market, their legacy is MP3 and other codec schemes.

2) psycho-acoustics - the ability of people to hear differences when they expect to hear differences, even when there are no differences or when the source is exactly the same source. Blind testing does not reveal this phenomenon; double-blind testing does. The ear is no more reliable than the eye when gathering information, particularly with regard to localization. It is the brain that does the processing, and it is easy to fool the brain. Few people take advantage of this human weakness better than high-end stereo and cable sales people.

[R.Daneel]

I couldn't have said it better myself!

Error correction algorithm might be the answer here. Similar to bit reconstruction in jitter elimination system aka Digital lens. Pioneer DVD player use this technology incorporated into the main DSP while on the other hand, LaserDisc may achieve a virtually identical trace on the scope without that tehnology. Still, there are some audible differences so the only thing that make those differences is the actual interpolation algorithm. I observed this on several occasions but always wondered how that might be if instrument showed virtually the same trace.

As for blind tests, I completely agree. The differences in this respect might be the same as the differences between different cables. One time I was at this high-end audio show. One particular room was full of audio journalists from different magazines and they were all writing stuff down how something sounds. But one closer look revealed that the left tweeter wasn't even hooked up to the amplifier! I always remember that and it's one of the reasons I stopped buying magazines. I never do A-B tests but rather take the time to get a good grasp and feeling of how something sounds. Hearing is a delicate instrument that adapts very quickly and small differences can be annihilated by our own internal equalization system. It's a bit of an off-topic here I know.

[PhilippeC]

https://picasaweb.google.com/minidisc.channel is the best MD blank picture database, but I guess we are a lot you know this precious link.

PhilippeC (also member of Sony Insider Minidisc)

Please sign our petition :
http://forums.sonyinsider.com/topic/27022-petition-to-keep-mds-here/

[Nick Sunn]

re: 8-17-2011 post by R. Daneel....."On The Various Brands of MD's..."
.........U.K. magazine 'HI-FI CHOICE' oct 1999 article....

HERE IS ONE MEDIA VENDOR THAT
SELLS ONE OF THOSE "seldom seen" MD's (hhb)
http://www.tapes.com/hhb-md80-80minute-minidisc-p-7665.html


There is another seller that has some
TDK Music Jack 74 3pk @ 7.99 BIN free ship #110744319264
SONY NEIGE 74 10pk @ 19.99 BIN free ship #120778124255 4avail hist36sold
also has 5pk of NEIGE 74 @ 12.99 BIN free ship #120778124255
(Ebay store: newdealaday ) seller:newdealady1 (1217+) 99.6
They all appear to be Japanese market MD's with Japanese text on wrappers

-------
I'm just pointing these out, should someone like to see different ones still in the packages, or perhaps someone is seeking an alternate source for new MD's or just curious. I do not know these sellers.

CRUTCHFIELDdotcom is a source for SONY blank MD's.

[Pentium100]

MiniDiscs (well, at least on my MZ-RH1) are buffered, so jitter of the reading is completely eliminated, the disc does not even spin some of the time. Error rates remain, so unless the data is damaged beyond the ability of error correction and re-reads to get it before the buffer runs out, it will play fine. Otherwise the player will skip.

[R.Daneel]

MiniDiscs (well, at least on my MZ-RH1) are buffered, so jitter of the reading is completely eliminated, the disc does not even spin some of the time. Error rates remain, so unless the data is damaged beyond the ability of error correction and re-reads to get it before the buffer runs out, it will play fine. Otherwise the player will skip.

Buffering of MD data into D-RAM is a standard MD feature and is actually an inherent specification of the MD system. In newer generation portable units firmware is designed to provide maximum battery life and one way to achieve this is to reduce the number of revolutions once the buffer is full or to stop the disc completely. This also allows for immediate playback even after the device has been turned off because the data remains stored in the buffer. It is unclear what is the actual size of the D-RAM for the newer generation units marketed as "G-protection" but it is a 16bit system for the older "10 second" units. In any case, one has to try very hard to make an MD player skip.

Error correction system of the MD system is a highly complex mathematical operation including FFT (fast Fourier transformation) which is essentially a system of multi-variable differential equations arranged in a matrix. I don't want to bore you to death but this is what makes MD a very reliable format with consistent results. Remember that the format was released in late 1991 and back then data corruption on storage media was a common problem. What's more, the hardware in an average MD machine is a lot more powerful than the one in a CD player which is what makes real time error analysis and correction possible.

It actually reminds me of something called "Digital lens" introduced in the mid 1990s by a company called Genesis. It's a digital processor intended as a time-axis corrector in a digital chain such as the one consisting of a CD transport and a D/A converter. It was essentially a buffer with a highly refined jitter elimination technology. The "Lens" was considered a strictly esoteric piece of kit and I guess many didn't even know what it did. It is generally unknown that some major companies like Pioneer use the very same technology but for a completely different purpose and it is to enable the DVD player to read even the most damaged discs. We did some measurement and scope tracking a few years ago and I have to tell you that a cheap DVD player like the DV-340 has an almost perfect square trace... Something even a Sony with the otherwise superb BU-1 transport cannot match. And mind you, this transport was used in some Accuphase CD transports. I was shocked to see some others have discovered this too:

http://lampizator.eu/LAMPIZATOR/REFERENCES/Goldmund/goldmundizator.html

Oops I got carried away!! It happens when I start talking about MD!

[ke4mcl]

interesting read. i cant help but wonder though how can a media "sound" different when all its doing is ones and zeroes? bits are on or off.

[R.Daneel]

interesting read. i cant help but wonder though how can a media "sound" different when all its doing is ones and zeroes? bits are on or off.

The article suggests there are measurable differences between the media. I doubt the reasons lay in the media itself but rather how the error correction system handles the media imperfections. It could be that the writing head is not always successful in changing the polarity on the medium. Error correction system would immediately fill in the missing bit according to it's interpolation algorithm which is adaptable. So there is no way to guarantee it would always work in the same way. This would explain the difference in sound the reviewers observed. I am not so sure the difference is noticeable.

Interpolation is a tricky business. Two CD transports can have a virtually identical scope trace and only a careful visual inspection would confirm that in fact we're talking about two different machines. However, they might sound different from one another. It is impossible to tell but interpolation might be responsible for this.

[Pentium100]

Interpolation (essentially "guessing" the missing data) is used as a last resort on CD players, only after the normal error correction fails. The regular error correction can reconstruct the missing data perfectly.

As MD uses compressed data, interpolation probably won't work as well as on an uncompressed data, for example CD.

Interpolation is not used at all in data media (hard drives etc). If error correction fails, the drive reports a read error and the software reading the drive can do what it wants - report the error to the user, interpolate, crash etc.

[close652]

During the past months I collected some blank MDs. It all started with a need of a dozen to record... Most of them are from ebay and a local auction site.
It is interesting to see the different brands and packaging from the early 60min pieces to the very latest ones still available in some (web)shops.

I am also interested who really produced MDs and which brand gave only its name (like BASF). There were four plants in Japan and two in Europe (and I do not know anything about USA). Sometimes it is obvious (Maxell) sometimes we only have hints.

https://lh4.googleusercontent.com/-ZCM4yVP6hSs/T10kpm-4JLI/AAAAAAAAAU8/KFeE3oexfXM/s640/blanks-earlymarch2012.jpg

https://lh3.googleusercontent.com/-MnLLm0cDxyE/T2HA4CtNXTI/AAAAAAAAAZc/yILhdu9jHks/s640/blanks2.jpg

[close652]

another source of blank media (UK) (http://www.stanleysonline.co.uk/search.php?frm_id=searchform1&search=minidisc&submit_searchform1=Search+Products)

[Scorpion8]

It could be that the writing head is not always successful in changing the polarity on the medium. Error correction system would immediately fill in the missing bit according to it's interpolation algorithm which is adaptable. So there is no way to guarantee it would always work in the same way. ....

Antun -- Doesn't this also suggest that the unit might sound "different" on two different plays of the same track? If the interpolation algorithm adapts, then it might "guess" different on two separate passes. I realize we may be talking single bits here in a digital "word" used to describe a portion of the music, but a collection of them being different on two different passes of the same track would essentially result in a different digital waveform being generated, and thus a different sound. Thoughts?

[Scorpion8]

During the past months I collected some blank MDs. It all started with a need of a dozen to record... Most of them are from ebay and a local auction site.

Oddly I find myself in the same boat. I have enough MDs now to last a long, long time.....*eyepop*

[R.Daneel]

Antun -- Doesn't this also suggest that the unit might sound "different" on two different plays of the same track? If the interpolation algorithm adapts, then it might "guess" different on two separate passes. I realize we may be talking single bits here in a digital "word" used to describe a portion of the music, but a collection of them being different on two different passes of the same track would essentially result in a different digital waveform being generated, and thus a different sound. Thoughts?

Hi Jim!

It is possible. However, I was referring to the recording process, not playback. During the recording, the machine has to cope with different situations resulted by imperfections in the media (inherent flaws or caused by damaged surface of the media), integrity of the incoming signal (high levels of jitter) and the shortcomings of the machine itself. An error correction system would apply the proper code given according to it's evaluation of the incoming digital stream. Once the recording is complete, the blocks of digital data on the disc are already formed and displaced according to the error correction system of that recorder. So, playback would not not be affected in any way.

This is in theory of course. In reality. situation is slightly diffeent and same rules that apply for CD players apply for MD players as well. This means that just like with CD players used as digital transports, there are differences between MD decks as well. However, with CD players, mechanical performance of the transport is a paramount since the signal is read and reproduced in real time. This is not a rule though. MD decks usually have a few Megabytes of RAM used as a buffer in which the data is stored before being handled by the encoder and then read from that same memory. While this minimizes the impact of transport on the performance, it still involves interpolation so differences are possible.

Given the hardware power of an MD deck, I'd say the differences are much smaller than between typical CD players.

[Pentium100]

However, with CD players, mechanical performance of the transport is a paramount since the signal is read and reproduced in real time.

This is only true for old and/or "high quality" CD players. Portable and car CD players buffer the playback - anti-skip protection and (for portable players) not having to spin the disc all the time saves energy and extends battery life.

[R.Daneel]

This is only true for old and/or "high quality" CD players. Portable and car CD players buffer the playback - anti-skip protection and (for portable players) not having to spin the disc all the time saves energy and extends battery life.

That's quite correct. Like I said, it is not a rule.

[Scorpion8]

It is possible. However, I was referring to the recording process, not playback. During the recording, the machine has to cope with different situations resulted by imperfections in the media (inherent flaws or caused by damaged surface of the media), integrity of the incoming signal (high levels of jitter) and the shortcomings of the machine itself. An error correction system would apply the proper code given according to it's evaluation of the incoming digital stream. Once the recording is complete, the blocks of digital data on the disc are already formed and displaced according to the error correction system of that recorder. So, playback would not not be affected in any way....

It seems that you just made my point, although reversed. If the interpolation is occurring in the recording stream, then it's entirely possible that, due to media differences, dust content of the ambient air that day for submicron particulates that may affect the laser, and subtle vibrations of the surrounding environment that induce jitter and errors, that each recording might be a unique event and that quite possibly no two recordings of the same digital data stream may be exactly the same.

So is it possible that digital is not, and can never be, an exact copy?

[Ghitulescu]

So is it possible that digital is not, and can never be, an exact copy?

No, the engineers provided every digital stream with CRC and other error identify- and/or correction- algorithms.

What happens, for instance, with the CDR is that there are several EFM codes for the same value - opinions are that some recorders (in particular Yamaha) that use a certain set sound better than the others. Maybe all this can be traced down to a better noise immunity, which is the mother of all digital signals.

[trott3r]

Thanks for the MD blanks pictures.
I never knew there were so many brands.

I tended to stick with sony and some tdk although JVC were my first purchase as they were cheapest by far at the local supermarket in the UK.

[Pentium100]

So is it possible that digital is not, and can never be, an exact copy?

No, digital is an exact copy. My MD recorder allows me to copy files to a Hi-MD formatted disc - so far none of these files have been corrupted. It is a bit different with CD-Rs, but if the disc was properly recorded then the files are OK (some time later they might become corrupted or unreadable though).

However, Audio CDs do not use a lot of error correction, so a few errors can lead into the interpolation. This was probable chosen to make more space available for the music. Data CDs use a third layer of error correction, that is why you can fit less data (700MB) than music (80min ~845MB) on a CD.

So, if your player can play, wav files from a CD, recording the music as wav files will result in better error correction and less space.

[paperclip]

I've been doing some tests of my own to check my CD ripping processes. I was surprised by the results. I'm an all-digital guy, I love MiniDiscs & CDs. I am in the process of ripping the new CDs that I let stack up and thought I'd check my processes.
Mostly I rip with a Plextor Premium drive, the thing is just a truly great drive - Fast and accurate. But sometimes I get lazy and use the LG Multi instead of firing up that machine.
I'm probably one of the few people who actually use Sony's Sonic Stage. On a fast machine it's ok and it has features I really like. Normally I use EAC for ripping but I thought I'd skip a step and use Sonic Stage. Sonic Stage doesn't have much for error correction, it's operates much like EAC in burst mode, so I assumed it did not do as good a job. I ran some comparisons from:
LG Multi using Sonic Stage & EAC
Plextor using Sonic Stage & EAC
The bottom line was all files checked exactly the same using EAC's file comparison. I was impressed as the CD itself was scratched and by no means pristine.
Then, not happy without some kind of error I burned the wavs to a CD-RW in CDDA mode. The disc itself was VERY used and I had no great hopes. Upon ripping those wavs, they also checked out identical.