bug80

Nice article, but it doesn't really answer the question asked in the introduction. In fact, it doesn't answer any questions at all.

First of all, I recommend that you convert the files from MP3 to WAV to ATRAC instead of MP3 to MP3 to ATRAC. Every "lossy to lossy" conversion (called transcoding) will give you quality loss. Second, I recommend foobar 2000 for all your conversions. It needs some getting used to, but it's a great audio player too.

My European MZ-N510 is only 2.5 mW and I can set it as loud as I want for listening with earbuds. I have to set it at a volume of 30/30 to get a to a decent level to hook it up to my hi-fi stereo, unfortunately.

You could try downloading this zip file from the guys of MDcenter.nl. It contains a registry file that modifies your registry, such that SS will point to Connect EU. I guess from there SonicStage will find out where you're living.

But compression algorithms choose the most optimal bit-depth for every block of data that passes. Or, in the case of CBR, it allocates the most bits to the blocks that needs to the most and fewer bits to less "important" blocks in a perceptual sense (both in the time and in the frequency domain). That's what I meant with bit-depth.

XP's default firewall can be a problem. It was a problem on my PC, at least. I had to turn it off to get past the Registration stage. After that, I could turn it back on and everything worked fine.

Yes that's what I mean. Gapless MP3's don't exist, but gapless MP3 playback does. CDex doesn't include information that is necessary for this to happen. I wonder if SonicStage will read it anyway, but that's another story.

Oh and yes, fake sp@132 = lp2@132, recorded as SP. So it gives you no improvements on quality.

Haha, well my "new" earbuds are definitely not A class (Sennheiser MX500, what are they, around $20 or so). Right now I am not thinking of going back to the Sony ones. SP (true SP, that is) sounds great through the Sennheisers. I will not give up that sound, just to mask the artifacts of LP2 This all doesn't mean LP2 can't be good enough for other people, mind you. It is just that I am very picky on sound quality, and a sound that I consider to be "bad" will ruin my listening pleasure, as is the case for LP2. Most other people don't find sound quality that important, or they just don't hear the artifacts. The popularity of MP3@128 kbs is a prove of this statement. It's just that LP2 isn't good enough for me (anymore).

Well, I don't agree I always thought that LP2 was comparable with MP3 >= 160 kbs too, untill I swapped the stock Sony earbuds for some decent Sennheiser ones. Now I believe LP2 is just horrible on most of the music I listen to. It lacks clarity, sharp attacks are smeared out way too much and the vocals sound as if they were recorded under water. Furthermore, an extended listening test at hydrogenaudio.org pointed out, that LP2 is worse than MP3 @ 128 kbs.

If I understand it correctly, you still need a special program (BeatJam) to transfer the files, if you want to be able to play them back. I think this is a special move from Sony. Because Mac users will be able to use Hi-MD with this unit, the sales probably will go up en Hi-MD is more likely to become a "standard". Furthermore, it was time for some cheap/low budget units to enter the market, so that people who just want a player to listen to their MP3's can jump onto the Hi-MD train. However, Sony has a high reputation (hardware wise! I repeat, hardware wise!) and probably choose to outsource the production of such low budget units for that reason.

But the final result can be optimal or sub-optimal, compared to what would be possible when memory, speed and power consumption weren't an issue.

Ok, that explains something, because CDeX doesn't encode gapless in my experience. In other words, the MP3's you transfered to your unit weren't gapless, that explains the short silence.

I did a quick listening test in the past to find out if this is true and the result was: I heard no difference between LP2 encoded on my MD unit and encoded in SonicStage 2.3. Furthermore, LP4 encoded in SonicStage sounded better than encoded with my unit. The thing is, the encoder in your unit can't evolve, because it is fixed in a chip (plus, Sony and firmware upgrades don't match). SonicStage's encoder improves over time (see the same topic mentioned above). You mention SP versus Hi-SP and SP being better, but that is not a fair comparison because we are talking about two different codecs here, whereas you claim hardware encoded audio sounds better than its software equivalent. Therefore, you should compare hardware encoded Hi-SP versus software encoded Hi-SP (and do a blind ABX test by the way, because the placebo effect is with us all ). Another thing I'd like to add is: a hardware programmer always has to live with a lot restrictions in terms of chip memory, speed of the chips and the like. A software programmer doesn't have these restrictions, or at least not that much. Therefore, it is more likely that the hardware encoder is sub-optimal, not the software one.

I don't see why ATRAC at a certain bitrate will sound better on a Hi-MD unit than its MP3 equivalent, when MP3 outperformes ATRAC when listening on a PC at the same bitrate. The latter is a dangerous statement, I know, but it is my experience (done a lot of testing).

Huh, shouldn't Hi-LP be totally gapless? Or at most something around a couple of milliseconds? kungfool, two questions: 1) what was your original signal, an audio CD or two wave files? 2) did you encode the files in one run, or seperate?

Properly encoded MP3's (CBR or VBR) should not give any problems when importing in SonicStage. The EAC/LAME combination should be fine (I mean, if THAT doesn't work, what else?). Some crappy encoders don't write the headers (metadata) like they should, especially with VBR files and I think SonicStage chokes on them for that reason.

It would be very interesting to see if something similar can be done with Hi-MD. It would not only mean we can bypass SonicStage when transferring MP3's, but it would also mean Mac and Linux support is possible. Am I right?

zulu2002, there are some reasons why people choose PCM (or lossless) for their music: * In the case that you get better equipment (or ears!) that reveal artefacts that you've never heard before, it's good to know that you have the original, full quality recording. You'll be able to re-encode it to higher quality bitrates, if necessary. * Another important point is, that it is always a bad idea to use a lossy encoded file for further processing. For instance, if you've recorded a concert of your band to minidisc and you want to apply some EQ, compression or maybe reverb before burning it on CD, lossy is not safe. The original encoding of the audio was such, that the music sounded as good as possible as it was. If you do processing afterwards, nasty things may happen to the sound.

I did a quick test in the past. There were definitely some minor improvements, but in the end it still sounds horrible indeed.

I just did a quick test, transfering the famous "castanets" sample in SP mode. I can assure you: it was definitely NOT true SP. My ears still hurt.

I used to encode my CD's to LP2 using SonicStage, for listening while traveling, but the artifacts became simply too anoying for me. So, I switched over to real-time recording in good old SP. SP is transparent to me. Real-time recording takes a lot of time, unfortunately. If I'll ever buy a Hi-MD unit, I think I'll switch to Hi-SP or even MP3's encoded with LAME --alt-preset standard. Because I always keep the original CD quality loss is no big deal to me, as long as the result is transparent to my ears.

And the fact that in most cases MP3 will give you the same quality as ATRAC at lower bitrates.

OK. That will be 212 minutes (approx 3,5 hours), then.

Well, let's see. A 192 kbs MP3 will give you 60*192000/(8*1024*1024) = 1.373 MB per minute A 80 min MD disc reformatted as Hi-MD is 300 MB So, with 192 kbs MP3's, 218 minutes of music will fit on the disc. However, the real value will be lower because some extra information besides the audio data will be stored on the disc. A 1 GB disc will give you 1024 MB / 1.373 MB/min = 746 minutes