The Ghost In The Machine...

[dispher]

hello everyone...

i need your help in something really annoying in *ALL* versions of SonicStage that i used starting from the version bundled with the MZ-NH1 ... (i think it's 2.1) up until 4.0. it is the way the encoder in SonicStage AND SimpleBurner uses the ATRAC3plus CODEC ... Not ATRAC3 or WMA or MP3...just the

ATRAC3plus CODEC.

When you rip a CD or convert files or rip using SimpleBurner it seems like the 2 programs have a way of achieving "gapless" rips or converts by streaming the audio signal as one long stream & cutting it where one track ends & the other begins (simulated track marks) but due to an error in the way the 2 programs

use the ATRAC3plus CODEC it doesn't split the tracks in the original point in the stream! it splits them before the end of the first track by about 300 to 400 milliseconds (less than half a sec) & the remaining part of the first track goes to the second track. i checked this using my ears & sound forge to examin the output.

what dose this mean??? you will always have to hear the last milliseconds of the track that was playing before the track you will listen to next ... on some

CDs this will have no effect but on gapless CDs it will be EXTREAMLY ANOYING

there are no solutions for this that i know of...except for

1. ripping the CD

2. combine all the tracks using the combine function

3. split them using the divide function

4. re-name & re-tag everything (everything was tagged & named perfectly using gracenote in the first place)

of course ... when you record using the MD unit (NH1 in my case) you don't have this flaw.

i have tried deferent PCs & deferent windows versions (freshly installed) heck, i even tried virtual PCs & i still get the same result!

i have made 3 WAV test tones for you to checkout...they are:

#1 (300Hz) // #2 (500Hz) // #3 (700Hz)

please pop them in SonicStage and:

1. select them all (all 3 signals) <=== IMPORTANT

2. convert Them to ATRAC3plus (any bitrate)

3. delete the WAV files.

4. listen to them...or better yet, open them in sound forge or (after converting to WAV) in any other WAV editor of choice.

please let me know if you get part of signal #1 in #2 and part of #2 in #3.

i really need to know if this is a widespread issue or something caused by some kind of flaw in windows (or SonicStage)

peace

ps: please rename the extension of the attached file to RAR

sample_tones.txt

Edited May 27, 2006 by dispher

[Avrin]

This used to be a very well-known bug in SonicStage. But today I encountered even more weird stuff in SS4.0.

I downloaded your files and converted them to ATRAC3plus@192. Deleted the original WAVs, then played the converted files in SS. No parts got to other files, but the files had GAPS between them. Opening them in Sound Forge 8.0d clearly shows about a 78 ms gap at the end of each file.

Then I converted the files back to WAV in SS. This added a 54 ms gap to the beginning of the 2nd and 3rd files.

Tried ATRAC3@132 (which used to be faultless). Same outcome but with different gap lengths (37 ms at the end and 26 ms at the beginning).

What's even more weird is that SS4 refuses to play the WAVs as they are. When I try to play one of the WAVs SS just freezes, and even Task Manager is unable to close it (I have to restart my computer).

Now for some more weirdness. I tried ripping a gapless CD in SS4 (the first King Crimson album, which is a horror for most encoders because of its phase shift problems). No gaps but track marks moved as usual.

Took some live MP3s (The Bays gig at Shambala), removed original gaps in Adobe Audition (in the usual way), saved them to WAVs, imported into SS4 and converted to ATRAC3plus@192, removing originals after conversion. Again - no gaps, but moved track marks!

The only idea I have about gaps appearing between your files is that they are too short for ATRAC3[plus].

EDIT: Just tried longer sinewave files. Generated three files with the same frequencies as above, but 20 second long each. Imported into SS4, converted to ATRAC3plus@192. GAPS are still there!

It seems that SS4 does not like pure sinewave forms.

Edited May 27, 2006 by Avrin

[dispher]

This used to be a very well-known bug in SonicStage. But today I encountered even more weird stuff in SS4.0.

I downloaded your files and converted them to ATRAC3plus@192. Deleted the original WAVs, then played the converted files in SS. No parts got to other files, but the files had GAPS between them. Opening them in Sound Forge 8.0d clearly shows about a 78 ms gap at the end of each file.

are you sure that u selected ALL the files when you encoded them to a3+ ? because in my case the gaps will appear when you convert each file individually. the ripping problem (track marks moved) is exactly happening to me in the same way...

the strange thing is....i suspected that my PC or windows config was wrong or broken in some kinda way....but i have tried it on multiple PCs with multiple windows installs

thanks for trying

[Avrin]

Yes, I selected all the files during import and during encoding. The way I always do when encoding live albums gapless.

EDIT: I have now updated the audio driver. Now SS4 does play WAVs normally, and all inter-track gaps are gone. But the files now DO get into each other when encoded in ATRAC3plus@192. About 385 ms of #1 is at the beginning of #2, about 392 ms of #2 is at the beginning of #3, and #3 has a nice 66 ms gap at the end (after converting it back into a WAV file). Also the waveforms are somewhat distorted at the places where files are joined (as seen in Adobe Audition). These distortions (amplitude changes) are about 50 ms long.

EDIT #2. Just tried ATRAC3@132. The result is somewhat better, though still far from ideal. About 55 ms of file #1 gets into file #2, 61 ms of file #2 gets into file #3, and there is a 24 ms gap at the end of file #3. The distortions are also present.

Edited May 28, 2006 by Avrin

[dispher]

you know...

this problem is sooooo frustrating because of this failure, SonicStage is 99% useless to me now

all my audio (that matters) is on original CDs so now i'm pretty much stuck! (ripping them in another program & importing will have the same effect as you know)

in a desperate effort to find alternative ways, i have tried to use the scripts in Sound Forge to extract & encode (in ATRAC3plus) a CD to import later in to SonicStage & transfer (as is) to my MZ-NH1 but it fell apart because of 2 things:

1. Sound Forge will rip & encode & INTRODUCE GAPS in encoding

2. SonicStage will import ok but will not let me transfer the files anywhere! (prohibited) &$!^@#&^$# <====

i am aware that the files produced by Sound Forge are AA3 ant not wrapped in an OMA container and it won’t help if i apply DRM to them & make them OMA

the thing that surprises me the most is that almost no one is complaining about this ???? Isn’t it a widespread thing?

thanks for trying Avrin

edit: i have just e-mailed Sony customer support (www.connect.com) about this unforgivable mistake! ;P

let's hope they can reply...

Edited May 29, 2006 by dispher

[Avrin]

Seems that the problem is even more serious than we thought. And it has to do with the format itself.

I burned your files to an Audio CD using Nero. Ripping them back with Adobe Audition 1.0 showed that they were burned perfectly (exact 4.000 sec lengths, exact waveforms).

Then I recorded them from my Panasonic DVD-S75 to my MZ-RH10 via optical in. The mode was Hi-SP. After tranferring the files to the PC and converting them to WAVs (SoundForge refused to open the recorded OMAs transferred to the PC) I got the following really horrible results:

1. #1 has 125 ms of silence at the beginning. The track length is 3.985 sec.

2. #2 has 139 ms of #1 at the beginning, and about 115 ms of waveform distortions around the joint. The track length is 3.993 sec.

3. #3 has 146 ms of #2 at the beginning, about 123 ms of distortions, and 2.949 sec of silence at the end. Also, the waveform does not drop to zero, but instead fades out for 28 ms. The track length is 6.780 sec.

The results of the same process but using the Hi-LP mode:

1. #1 has 129 ms of silence at the beginning, then fades in (though not from the zero level) for about 44 ms. The track length is again 3.985 sec.

2. #2 has 142 ms of #1 at the beginning, and about 20 ms of very slight distortions (much less than in Hi-SP!). The track length is also 3.993 sec.

3. #3 has 149 ms of #2 at the beginning, and about 14 ms of almost unnoticeable distortions. It fades out to zero for 28 ms, then has 2.943 sec of silence, and its length is again 6.780 sec.

EDIT: Some forum members may object that 4 sec files are too short for ATRAC3, and do not contain a whole number of frames (75 ms). Tests performed using 12 second tones yielded similar horrible results. Using both hardware and software encoding. Though ATRAC3@132 clearly shows better results in all cases.

EDIT#2: Just tried other available formats:

ATRAC Advanced Lossless - similar results.

WAV - perfect.

MP3@192 - small gaps all over the world with track marks roughly in the middle of those (between files).

WMA@192 - perfect (!).

AAC@192 - files get into each other AND have gaps.

Edited May 29, 2006 by Avrin

[chris_caudy]

Just to confirm the above findings, I also am having problems with encoding introducing gaps and beginning and end of tracks corruption. I agree that this causes great problems for me as I like to record radio drama and use CD track marks for easy navigation, then import to SonicStage. Presumably this also affects the ATRAC HD and memory stick based devices. If so this will affect products that are early in their life cycle. This doesn't bode well for SonicStage when it's up against I-T**es. Will also email my concern to Sony. Grrrr!

[dispher]

what is up MD peoples ?

the e-mail i have sent to sony connect (since no other division wants to help or at least provide an "e-mailabe" address) was the first post in this thread...& after more than a week (9 days) here is what they had to say:

================================================

Dear Customer,

Thank you for your patience.

Sony CONNECT store and Sony Aura devices support gapless recording and

playback. In your case the key point is that you have to use a gapless cd

in

order for SonicStage to rip it as gapless.

We hope this information is of assistance to you.

Kind regards,

The CONNEC Team!

================================================

don't you just hate it when people seem so confident when actually they don't have a clue?

[Avrin]

There really exists such a thing as a "gapless CD". Simply burning gapless WAV files with whole frame lengths is not enough to make a CD "gapless" according to the standard. Tracks need to be linked together (not necessarily all of them, only those, which are meant to be gapless). Feurio CD Manager does this nicely. Nero does not even know about this. Also, Feurio may be used to rip such CDs into MP3. It detects gapless tracks, and rips them as a single file (this may be switched off).

What is really funny is that SS does not differentiate between "gapless" and Nero-burned CDs. It always shifts track marks. And I wonder whether the CONNECT store really supports gapless...

Edited June 7, 2006 by Avrin

[Richard]

I have been following this thread for a while and am a little confused. I have literally hundreds of cd's of music that is gapless. For example one long piece of music lasting 60 minutes but with track marks on the cd at 5 minute intervals.

I have ripped most of them using SS or Simple Burner and transferred to Hi-MD. I have had not issues with any of them no longer being gapless or having track marks in the wrong places? Am i missing something completely obvious or have I not understood this thread properly?

[dispher]

the problem is in the encoder...not the CD ripping component

richyhu: just download the files & pop them in SonicStage & encode them in the same steps that i mentioned in the first post & check the outcoming file ether directly using Sound Forge or by converting it to WAV then opening it in any WAV editor. most probably you will get ~380ms of the last part of file #1 in the beginning of File #2 & ~380ms of the last part of file #2 in the beginning of file #3.

i kindly ask you to double check if the tracks were encoded 100% perfectly (the track marks exactly correspond to there original location in the source material)

Avrin: it's the way the encoder works...it's not about ripping gapless CDs (that's what i mentioned to the e-mail to Sony but they keep insisting to treat it as a CD ripping problem witch will put more variables to this problem!!!)

i'm starting to guess that this is somehow an OpenMG Secure Module error more because:

1. SonicStage is just a GUI for OpenMG Secure Module (witch explains why this is happening also to Simple Burner).

2. it doesn’t affect any other file encoding (only OMG files).

the error in the ATRAC3 codec is very undetectable since it is ~50ms that gets mixed up with the next file & is hard to notice in terms of audibility.

my only fear is that this problems becomes one of "those" problems that never get detected by quality control (if there is such a thing for this software) & never get fixed

Edited June 7, 2006 by dispher

[Avrin]

This is an ATRAC3[plus] format problem. Including hardware.

Gaplessness is not a problem. Trackmark misplacedness is. Very simple tests show that trackmark misplacement is up to 70 ms in ATRAC3, and up to 400 ms in ATRAC3plus.

And I should probably add that audio CDs burned by SS do not comply with any standard at all.

Edited June 7, 2006 by Avrin

[chris_caudy]

I've received this reply from Sony Connect this week concerning this problem

Thank you for your patience.

Much to our regret we must inform you that there is no solution available yet.

But your issue has already been forwarded to the responsible department for

furhter investigation as other users have the same problems.

We will inform you as soon as possible.

We are sorry for any disappointment this may cause.

Kind regards,

The CONNECT team

I pointed out that if this problem exists with the ATRAC3plus encoder in Sonic Stage, then it would affect products very early in their life cycle such as the hard disc devices as well as the more mature products.

I live in hope ...

[Avrin]

Now this is something!

[Avrin]

This is not an experiment, just sad experience.

I had to record a non-stop CD to my RH10 in PCM via optical in. The CD itself is a bit defective (the last few tracks) and does not not rip completely on my computer, but plays fine on my DVD player. So I spent 68 minutes recording it, then another 25 minutes transferring it to the PC, then a few more minutes converting the files to WAVs, and guess what I got when I opened these in Adobe Audition? Misplaced trackmarks! In PCM! Between all files!

Now I wonder where these came from. Possible sources:

1. The unit does not put trackmarks correctly when recording.

2. Trackmarks are misplaced during transfer.

3. Trackmarks are misplaced during the OMA to WAV conversion.

Don't want to know the exact cause. I'm tired of all this...

[dispher]

indeed...

i just realized that YES this flaw is also in the hardware! but not severe as the software encoder (at least in the NH1)...

i just made a test with a 100% complaint audio CD i just compiled & it gave me random errors (in the trackmark placement) that varied from 18ms to 35ms using ATRAC3plus 256kbps (Hi-SP) but still...it's nowhere near the ~380ms you will get when encoding using the software encoder. i guess that explains why i didn't notice it before & why i noticed it right away when using the built in encoder in SonicStage

sad news for us....

i really hope that the SonicStage people will realize this in there next update.

i just can't help to be completely angry when i put this great technology side by side with this stupid crappy fault!

time will tell

[Karmageddon]

So i guess there still is no solution to this problem?

I too have noticed this, even with old old verisons of SS (2.0)

I really would appreciate Sony looking at this issue as it is misleading when they say the software and hardware are "gapless"

the gapless feature is useless if the track marks are out of alignment.

After having converted 1/3rd of my CD collection in SS to ATRAC3+ @ 192 kbps i have to combine the offending trax and them split them again. kinda annoying.

Sony, Please read this. Please know I am a buyer/owner of the amzing RH1 and fully support MD and have since its inception, despite all the DRM and BS that that brought with it.

You guys (SONY) are soooo close. Don't make us wait 6 months or ayear for a simple software fix.

This really needs to be solved as this affects EVERYONE who uses SS...

Is anyone up for a petition?

Edited July 7, 2006 by karmageddon

[Terence Kelly]

I'm no technical expert, but I seem to recall that the original minidisc format was only accurate to about 1/12 of a second in terms of placing a track. Many of the errors listed here fall within that leeway. Are the errors cumulative? I.e. does the misplacement of the first track affect the misplacement of the next? 30 milliseconds off seems like it could be expected, but 400 ms shouldn't be happening if the software is working properly. On gapless opera cd's I haven't noticed this error, but it would probably have to approach .5 seconds before it would be audible. I'll be curious to hear if Sony can fix it.

TK

[dispher]

This really needs to be solved as this affects EVERYONE who uses SS...

Is anyone up for a petition?

as sad as it seems...

i expected a lot of people to realize this problem after testing it but i'm surprised that there are only a few members reporting this (about 4 that i know of) so it's kinda frustrating...my guess is that it's not audible to them since the error itself is devilish (the audio stream will be 100% gapless with no pops or clicks or anything audible except for the track marks...they will be misplaced) so when you listen to an album continuously, you won't be able to detect it).

Are the errors cumulative?

nope...because it would mean that the final milliseconds at the end of the full stream would be chopped off but instead, you get a random chunk of silence...

ps: it's good to see this thread come to life again...thanks for checking it out!

[Sparda]

I really can't fathom what the big deal here is. The only problem is that the track marks are misplaced by what, lol a few milliseconds? Even a whole second wouldn't be a big problem. I have noticed what you mean listening to one of my albums and can tell you that it doesn't bother me much if at all. As long as it doesn't cut out anything (which if it does then it's a big problem), then I really don't see whats the problem. I personally only experience or notice this track mark misplacement in a few of my albums on my MDs. I suggest maybe to just listen to the tracks in their intended order so you don't experience this minor error? Btw, this thread is being dug up because a recent thread links to here.

[Avrin]

1. If you just play complete albums from start to finish (as I mostly do), there is no problem at all.

2. If you switch tracks during playback or use the Program mode, things start getting annoying, especially if there are loud tracks before low-volume ones.

3. If you want to to re-sort tracks, you are in deep trouble.

Edited December 26, 2006 by Avrin