Realtime vs NetMD Recording

[markwalsham]

OK - so my journey back to MD land happened a few weeks back, when I managed to obtain a lovely Sony MZ-R35 and Sony MDS-JE530 deck.  I have another thread on the various experiences I have found with the sound quality of digital to digital copies and analogue.

Anyway - I recently acquired a Sharp IM-DR420H recorder, which happens to be a NetMD unit and using the excellent guidance on here, I have managed to get Sonic Stage 4 (Ultimate) running perfectly well on Windows 10. After transferring some music to the Sharp via SS, I realised that the restrictive nature of Sonic Stage is probably not worth the effort.

Why?

First I thought, ah, I can make use of the Atrac Advanced Lossless codec, which I can of course, but this didn't yield any noticeable differences between that and an optically captured recording done in real time.  Secondly, the whole (and I am not opening the debate) of check in and check out is a major pain.  You cannot just wipe a MD with tracks from SS on it, as it says protected, so you have to check them back in.  This is a pain and probably not worth the effort.

I was also not impressed by the transfer speed of the SS to MD process.  Especially on the check back in from MD to SS, which was considerable!

Finally, it means I have to have to fiddle with music file formats, as I have a variety in my library, but the dominant one is AIFF, which SS does not understand!  Yes I know I can transcode and covert to WAV or FLAC, but that is another step in the process that isn't worth it.

So - I have come to the conclusion that using a CD copy of the music and copying this optically to the Sharp MD unit is just as easy, with the only downside being track naming and time.

I guess I am stating the obvious - but it's why SS came to have a bad reputation!

Is there any other 'Benefits' of using SS to improve sound quality that I am not aware of?

Cheers, Mark. 

[sfbp]

Just a few corrections, Mark:

1. With the 4.x Sonic Stage, no one cares about checkout count
2. Checking in is the same as deleting off the media (from within SS). Nothing is copied
3. You can use your trusty pre-NetMD devices (R35 and JE530, both) to delete files or format disk.
4. AAL (Atrac Advanced Lossless) is not transferred to the device (there are some portable flash devices which appear to play back the full file, but I seriously question whether a human, no matter how good their ears, could tell the difference).
5. Where AAL shines is as a storage format
    a. it is not encrypted at the end of ripping (so no need to do the SS File Conversion Tool thingy - I am sure you read all about that and how it allows you to transfer files out of your PC and play them in another, absent the encryption which is placed during rip)
    b. if you are not sure what the eventual destination will be, you can postpone the evil day and store in a format (I recommend 256kbps ATRAC3+ or if you can be absolutely certain you won't play on HiMD device, then 132kbps MDLP Atrac 3) which can be varied when transferring to the USB device. SS will make the correct decision and fix up the output for you based on the destination you choose. Of course the transfer of the native format without the lossless extra will be faster than an actual conversion, but with today's PCs there's little downside anyway.

BTW you can still title the disk, sitting in your portable unit, from SS, regardless of how the music got there.

Oh, one more thing. ACTUAL SP format does not happen with NetMD. Many people therefore swear by optical recording of SP. I am not one of them, as I know from (other folks' scientific) tests that MDLP is the better codec, and what we regard as "warmth" is actually a bit of a fake.

Another hint.... do NOT make the mistake of ripping WAV with SonicStage and hope to transfer it to the portable. It doesn't work worth a darn. At the root of this is that 16-bit rips are not really a true reflection of the contents of a CD, whereas an ATRAC3 (or AAL!) 24-bit rip will be awfully close. The software will take your 16-bit rip and convert it to MDLP., which then gets sent to the unit (when they did NetMD, USB wasn't fast enough otherwise). So now you've had 3 conversions, 1 to rip, another to transfer, and a third to pad back zeroes to make MDLP back into SP (we refer to this as "Fake SP"). The result is terrible and almost certainly accounts for why MP3 beat ATRAC in some lab where supposedly fair tests were done.

There's probably more, but that's a first pass.

Welcome to my world.......

Finally, ALWAYS have your Sharp powered by the mains, not just the battery, when transferring. NetMD units in general (this does not apply to HiMD units running in NetMD transfer mode, I think) do not get power from the USB connector, putting you at the mercy of a battery which might be only NiMH at the lower end of its operating range (1.15 volts instead of a nice 1.6 from fresh Alkali).

Ya pays yer money and ya takes yer choice..... (many folks buy a HiMD unit and transfer MDLP to it!!!)

Stephen

[PhilippeC]

8 hours ago, sfbp said:

many folks buy a HiMD unit and transfer MDLP to it!!!

As for me with my Net-MD work team MDS-JB980, MZ-NH600 & MZ-N710 used for my LP2 recordings.

[markwalsham]

Stephen,

Thank you for your corrections and updates, it has certainly been a learning experience over the last few weeks.  Funny how this was all oblivious to me back in the day when I would just squirt a CD over to my Sony MZ-R909 and off I would go (after frantically titling all the tracks)!

I have today, done some additional testing - and I think I am starting to come to some conclusion on what I think works, and for me, what is not quite as good!

• So the first test I did today, was to import my reference CD (Queen, The Game, Deluxe Remaster 2011) into Sonic Stage using ATRAC Advanced Lossless Codec, and I then transferred this to a blank MD.  I then took the CD and ran a CD to MD copy through my JE530 (via Optical connections).  When I transferred through SS, I used the standard (Default) transfer option, which is to copy tracks over in LP2 mode - I made no changes to the default SS settings other then where the music files where to be stored.
  
  I then did some close listening between the two recordings and what I found was that:
  
  The SS transfers were almost as good as the original source material, and most people would be hard pushed to hear a difference, but I can hear that there is a little bit of squishing at the top end of the audio, and a slight loss of focus.
  
  The direct digital copy was near identical to the source material, with sharper treble response and a more focused soundstage.
  
  The differences are marginal to be honest - and for everyday, sitting around a pool listening to music, the SS transfers in LP2 would be fine.  I wonder if over time I would tire of the sound though, since the digital copy seemed more neutral to my ears.
   
• That got me to thinking, so I fired up Exact Audio Copy and I ripped the source material again, this time directly into WAV.  I have used EAC before, and it does a good job of ripping to WAV, and lets you have a lot of control over the settings.  So I took these files, imported them into SS and transferred them to another MD, but this time I did a few things.
  
  First I transferred with the default settings > LP2 track.
  Second I changed the settings to use SP only > SP track.
  Third I changed the settings to use LP4 > LP4 track.
  
  I then compared by listening back (through a Sharp IE-DR420H) with the same settings as prior to each in turn, and here is what I found:
  
  LP2 Track - I could not discern any real difference between this track (from a well encoded WAV) to the SS AALC track.  They sounded very near to the source material, but with the same squish in the treble and slight loss of focus.
  
  SP Track - I didn't like this at all - it sounded a bit flat and the treble was a bit splashy too as if an over harsh EQ had been applied.  Not worth listening to.
  
  LP4 Track - run for the hills.
   

Conclusions

So now I have tried a number of ways of getting audio data into SS and out of SS to MD, using different input file types (yes have also tried MP3 and M4A files too and they yield poorer results than above - which isn't surprising) and have come to the conclusion that for "My Ears Only" that I prefer the sound of the reference material recorded directly to MD from the CD master using my home MD deck in SP mode.  This offers the most neutral, focused and sharp sounding recordings that I have found so far.

Yes it means only one album per MD, yes it means realtime dubbing, yes it means carrying around more discs, but for me, if sound quality is key, then that is a trade off I am willing to take. 

This isn't a put down of MDLP and it's associated codec's, as LP2 using AALC is very very close 2nd in these tests, and as I said, it's likely that 90% of people who listened to that would not notice any difference at all.  I am very familiar with the music I am using to test the sounds with, so I can spot those nuances.

So for me, important albums will always be in SP and those which I consider to be 'casual listening' - I shall go the LP2/AALC route.

Good tip on using SS to title tracks - hand not appreciated that, and now I can use it for that purpose, and yes, I did make the schoolboy error of assuming the NedMD device would draw power via the USB port, and did not have my AC source connected and it lasted a minute before giving up the ghost on a standard AA battery!

Thanks for your help everyone, and I hope others might find reading these tests useful.

As usual, your mileage may vary, because you have different ears to me!

Cheers, Mark.

[sfbp]

What happens if you try transferring by USB to lower data rate is this: the sound spectrum gets "mastered wrong". The proper frequency for cutoff goes down as you use lower data rate. If you don't use the right cutoff, too much data gets spent trying to process higher frequencies instead of being used on the main body of the sound. Result- distortion/squishing, just as you describe.

The surprising thing is that if you record direct to optical at LP4, USING A DECK, (at least in my experience) it works great! I think doing that with a portable is not so cool and guaranteed, which is probably one of the things a deck does with all that "extra" circuitry. I have thousands of hours of classical recorded at LP4. Admittedly it's for use in the car, but I am pretty sensitive to any sort of distortion, and this gets it right, IMHO.

The other interesting factoid is this: when you record in MDLP (either LP2 or LP4) and transfer to the PC (using USB and the MZ-RH1) or from the PC, the data is completely untouched. As it is when you move Hi-MD music files around. If you can't do that, then the ability of Sound Forge to process all forms of ATRAC except AAL, is quite useful. SF seems to get the mastering right.

[markwalsham]

Stephen,

Thanks again, and alas my home deck is only a Type-R deck, and does not support MDLP, but I am searching for a reasonably priced MDS-JE780 or if I find the funds a MDS-JB980, and then I can test LP2 and LP4 as well.  I understand that ATRAC Type-S does nothing for SP mode playback, but improves LP2 and LP4 endcoding and decoding! So it would be one to try and see if the later devices handle MDLP better than the earlier versions.

Your last comment, does this mean that any MDLP recording when transferred in any direction, there is no decoding or re-encoding?

I ask as I have seen that SS always seems to go into “Converting” when sending data to a NetMD  when using MDLP - this implies there is some processing going on - or am I missing the point?

Cheers, Mark.

[sfbp]

The JE640 is my reference deck (when I started making "good" LP4s). Both the 980 and 780 come with the complication of TrProtect which you will have read about. 770 is probably the most up to date deck without Type-S and NetMD.

Now - you are right about improvements, Type-S "INCLUDES" Type-R, so the Type-R recording improvements are there. But Type-S is a playback technology. Just in case you don't want to wear out decks, you could use one for mainly recording (a Type-R) and one for mainly playback (a Type-S)

The MDLP data DOES get there untarnished. It's just that they like to encrypt and decrypt the data in order to prevent someone sticking a USB-hoover on the end of the wire!! So that is why it says it's converting. Even the RH1 cannot upload MDLP NetMD tracks, because they put something we cannot get at into the TOC on the minidisc itself. The actual (unencrypted) data plays ok even on a non-MD deck, though. This was one of the stumbling blocks when trying to figure all this stuff out. Someone actually convinced the RH1 to upload SP but that never really saw the light of day; the code is in the linux-minidisc# thing, but there are too many other things wrong with the way it works for practical usage, at least when I tried it out. I don't think the other direction is ever going to work, though, it would require a firmware change on the NetMD target. As I mentioned, NetMD was designed back in the days of 1.1 USB, which simply is not fast enough to cope with 1411kbps.

 

[markwalsham]

Me again!!

All my listening tests have been conducted at home, in the comfort of my usual environment, but this evening I had to go out to play in a local band, so I had a 45min drive to get there!

So, I took the MD player, connnected to the car through the auxiliary input, and took two MD’s.

One is an LP2 recording made through SS from an CD encoded into AALC, the second is a SP disc burned to MD from CD via optical.

Both these discs, when listened through the supplied headphones, sound pretty identical to me, except the slight squashed high end, but it’s barely noticeable!

So the results from the car are very interesting - playing the MDLP recording at admittedly quite a high volume, it was clear that the squashed high end was even more present - especially in cymbals which sounded harsh and sibilant. I only lasted for 15 mins before I had to turn it off.

So I switched to the SP recording - and again, it sounded like the sound stage opened up, the treble had the crispness and focus was much better.

I have learned then two things - MDLP through headphones is very convincing, and secondly MDLP when played at higher volume levels can expose the weaknesses of the codec more easily.

I also did some comparison listening between the MD and the lossless files on my iPhone using AirPods, and I came to the conclusion that the latter won, but then it isn’t a fair comparison is it?

i am still enjoying going down this MD rabbit hole though!!!

More listening to be had, and more observations will be made too!

Cheers, Mark.

[sfbp]

Sony did some clever stuff in their car units to do with voltages and noise suppression. I'd wager that most of it sounds MUCH better if you actually had a (Sony-made) MD head unit, The 3,5 mm jack is pretty unreliable, and would be my best guess as the source of bad things in the car.

[markwalsham]

Stephen,

Yes, ordinarily I would agree, except that other 3.5mm output devices when used sound pretty good, such as an old iPod, and even a Bluetooth device to allow music streaming from my phone to the car.

I am now toying with a Sony NW-A45 to see if I can put DSD/Lossless music files on that and try some listening tests. I had grown tired of the smartphone output, hence the diving back into MD, but it’s not quite hitting the mark for me (sadly), so I wonder if that might be a way to go?

As usual, time and money are the limiting factors!

Cheers, Mark.

[sfbp]

Depends on your car. A head unit from around 10 years ago is quite affordable. I can even sell you one for not a lot if you're willing to pick it up from Surrey the week after Easter, as it's time to start to rid myself of "spares".

Modern cars have Bluetooth that is way capable (early BT was not, quite) of reproducing. My favourite BT device is the **Japanese** version of NW-F887 (and /6 and /5). But not the NWZ-F887. Why? Because the domestic version is the only one that still supports ATRAC. These little beasts sit in your glove compartment quite happily, given that is one place you can find USB power these days. My entire ATRAC library (well, not all at once) can be copied there. It even speaks to Sonic Stage as an ATRAC Audio Device.

Oh yes, did you activate the "line out" option on the R35 (I haven't checked but most portables have one)? Experiment has shown that if line out is not active, turning the volume on output up all the way actually increases beyond the "line out" point, which is, IIRC, about 85% of max. So there's another possible source of distortion.