Jump to content

NEW: SonicStage 3.3 Now Available!


Richard

Recommended Posts

  • Replies 155
  • Created
  • Last Reply

Top Posters In This Topic

Atrac Advanced Lossless:-

http://www.ietf.org/internet-drafts/draft-...c-family-04.txt contains...

"Early versions of the ATRAC codec handled only two channels of audio

at 44.1kHz sampling frequency, with typical bit-rates between 66kbps

and 132kbps. The latest version allows for a maximum 8 channels of

audio, up to 96kHz in sampling frequency, and a lossless encoding

option which can be transmitted in either a scalable (also known as

High-Speed Transfer mode) or standard (aka Standard mode) format.

The feasible bit-rate range has also expanded, allowing from a low of

8kbps up to 1400kbps in lossy encoding modes."

"As ATRAC supports a variation on scalable encoding, this payload

format provides a mechanism for transmitting essential data (also

referred to as the base layer) with its enhancement data in two ways

-- multiplexed through one session or separated over two sessions.

In either method, only the base layer is essential in producing audio

data. The enhancement layer carries the remaining audio data needed

to decode lossless audio data. So in situations of limited

bandwidth, the sender may choose not to transmit enhancement data yet

still provide a client with enough data to generate lossily-encoded

audio through the base layer."

Copyright © The Internet Society (2005).

Here's my amateur explanation, which may be wide of the mark - what this does is to rip a lossy compressed version from your CD, plus all the audio info missing from the lossy version. Add these two together and you get the original. That's what you get when you play back the lossless version on your PC. What goes to your MD (or whatever) is the lossy version without the second part. When you rip the CD, you choose a bitrate for the lossy part. Choose a low bitrate and you will have a small file for transfer to your MD and a large file containing the remainder. Choose a high bitrate and you will have a larger file to transfer to your MD and a less large file containing the remainder. (These may not actually be separate files on your PC but the whole lot bundled into one like having several files in one zip file).

Edited by ozpeter
Link to comment
Share on other sites

What I really want to know is that if I rip all my CD's in AAL 64K, can I use this to convert to Hi-SP later without re-ripping or loosing quality, i.e. does it use the AAL version to convert and giving Hi-SP quality.

I say this because there are times when I like listening to Hi-SP and times when Hi-LP does the job. Its mainly Hi-LP I listen to at present as the background noise makes it pointless for Hi-SP.

Link to comment
Share on other sites

Here's my amateur explanation....

Now that makes sense :lol: !

Thanks ozpeter!!!

That is actually a pretty good idea!!!

My question would be similar to Qwakrz:

What happened when you export toward another bitrate?

Example:

Is Lossless(256) -> 352 the same as PCM -> 352 ?

if so, we can rip lossless, convert to 352 and then transfer to Hi-MD.

Cheers

Edited by LupinIV
Link to comment
Share on other sites

In theory such transformation should be possible but personally I don't know what the software provides. I imagine that the way it's been set up - asking for a bitrate at the outset - is to provide rapid transfer without recalculation. Changing the bitrate would require the whole thing to be redone, perhaps to no advantage compared with ripping the CD again.

Link to comment
Share on other sites

Then shouldn't an ALL file ripped at 64kbps be larger than one ripped at 256kbps because more information is needed to correct the lossy portion?

I could be wrong but back when I was testing it out I noticed that the 64kbps file was still smaller.

Yes if you rip at AAL64 you will have a SMALL lossy file and a BIG lossless file,

however

If you rip at AAL256 you will have a BIG lossy file and a SMALL lossless file. They kind of cancel out.

It could be that AAL64 is getting the encoding closer to the original and needs less correctable bits than the AAL256 when they are finally combined, or AAL256 has a small continuous error that has to be corrected for at all times by the lossless add-on.

Link to comment
Share on other sites

What I really want to know is that if I rip all my CD's in AAL 64K, can I use this to convert to Hi-SP later without re-ripping or loosing quality, i.e. does it use the AAL version to convert and giving Hi-SP quality.

Actually, no.

I've tried converting a 64k-tagged AAL into 352k-Atrac file, then deleted the AAL-file.

Result: It sounded like Hi-LP@64k.

When converting or transferring onto HiMD, only the lossy part is used.

What happened when you export toward another bitrate?

Is Lossless(256) -> 352 the same as PCM -> 352 ?

No, as only the lossy part is used.

Go via PCM, then convert, then delete the PCM-File in the track-properties.

Edited by jadeclaw
Link to comment
Share on other sites

Actually, no.

I've tried converting a 64k-tagged AAL into 352k-Atrac file, then deleted the AAL-file.

Result: It sounded like Hi-LP@64k.

When converting or transferring onto HiMD, only the lossy part is used.

If this is so then what is the point of lossless in SS?

It means that recording in AAL256 and converting to 352K will mean you have a file that is only as good as 256K would be normally. This is pointless then (mind you it is Sony).

One thing I did notice is that if you encode in AAL64 and then choose to force a transfer to Hi-MD at 64K (same applies if you rip in AAL256 and transfer in 256) it goes through a re-encode again of the track. If you just say transfer as is then it will use the lossy part of AAL for transfer.

Link to comment
Share on other sites

Seems Sony has kinda listened to me ;)

That Atrac Lossless had to be converted for 1st and 2nd gen units is in itself understandable. I think it could mean some future Sony device will play back Atrac Lossless. If that's an HiMD playa, it's all mine :)

What about a recorder that can encode to AAL on the fly, and then upload? Come on, Sony. We know you're listening. ;)

Link to comment
Share on other sites

You try the other codecs? I always enjoyed your thoughts on codec quality from a new version in comparison to the last.

Well, there is not much to report, MP3 hasn't improved at all, if you need MP3, stick with Lame,

Generally, the improvements are not dramatic, I don't think, that these small improvements warrant reripping a library at the same datarate.

And Atrac3Plus@128k still doesn't reach LP2@132k

However, if the MD is used mostly at home over a good system,

352k Atrac3Plus is the way to go.

It is the best compromise between playing time/space requirement and sound quality.

That option should have been included in SS2.0.

Ok, what is still needed?

A) 352k on CD-import, B ) Decent home decks with Digital In AND Out. And 352k-standalone recording.

C) Decent Car-units.

Btw, Atrac3@105k & 66k are back via fileconvert.

Edited by jadeclaw
Link to comment
Share on other sites

Well, there is not much to report, MP3 hasn't improved at all, if you need MP3, stick with Lame,

Generally, the improvements are not dramatic, I don't think, that these small improvements warrant reripping a library at the same datarate.

And Atrac3Plus@128k still doesn't reach LP2@132k

However, if the MD is used mostly at home over a good system,

352k Atrac3Plus is the way to go.

It is the best compromise between playing time/space requirement and sound quality.

That option should have been included in SS2.0.

Ok, what is still needed?

A) 352k on CD-import, B ) Decent home decks with Digital In AND Out. And 352k-standalone recording.

C) Decent Car-units.

Btw, Atrac3@105k & 66k are back via fileconvert.

D) losless WMA (or AAL) direct playback on the units. Or do you think 352k reaches pcm quality with all types of music ?

Edited by garcou
Link to comment
Share on other sites

Or do you think 352k reaches pcm quality with all types of music ?

Not necessarily.

But the difference is marginal, I guess, that 99% of all listeners won't hear a difference.

That's why I said, best compromise, as it still provides a decent playing time on standard MDs.

At 352k, I don't hear a difference, at 256k, I still have the feeling, that something is missing, so it is 352k for me.

Link to comment
Share on other sites

I've always wondered about that. How did you test this?

By comparing with the PCM-file of the same song.

SonicStage plays 128k directly.

LP2 is less artefacting than A3P@128k.

And the high range sounds a bit crispier as well.

Or is this based on your personal opinion?

No. And yes as well, as I don't like the kind of glossing over, that A3P@128k produces.

Does the new SonicStage support Atrac3Plus 128k on a minidisc?

No. it has to be converted. Convert to 352k, that should be high enough to preserve the artefacts.

Link to comment
Share on other sites

Intriged and confused by it to behonest. A=Lossless and Bx=Lossy where x=bitrate(64/132/etc)

A and Bx stay on the HD. Only Bx is transferred to the portable device. A+B(64) then B(64) transferred to portable. Transfer is quicker because B(64) isn't actually re-ripped when transferring. However if you pick a different bitrate say B(132) then

A+B(64) user picks 132kps for transfer SS transcodes B(64) to B(132) which destroys the sound quality.

To be honest it sounds like a bug that got through testing. Obviously it should actually encode A to B(132).

Still its not a total loss. We now have a higher bitrate for playback 352kbps, which is great if you use HiMD as a player. Personally I don't, but I will be interested in seeing how this 352kbps compares with other MP3 players I have. We already have lossless with PCM. So all that this update does is give us a new bitrate and speed up the transfer speed. Is speed of transfer really an issue for people?

well, not necessarily true blue... I rip my cd's with EAC onto a 250gb external HD, so it doesn't really matter to me to rip 'em to wav... with 3.3 I can transfer these ripped CD's as 352kbps (which is evn closer to CD than 256 obviously and practically indistinguishable from the source)

so there is a bonus even for creating MD's... a gained bitrate, now if only we could directly record into 352 from analogue in :lol:

I pretty much do the same as 'The Low Volta'. Big external disk, and I don't use SS except to transfer to and from the HIMD. I also would like to record analog in 352kbps. Still its nice to see any development of the technology. Theres still potential in Atrac if they can give us ATRAC, Lossless and a wider range of bitrates and portables that have gapless playback.

Link to comment
Share on other sites

"Early versions of the ATRAC codec handled only two channels of audio

at 44.1kHz sampling frequency, with typical bit-rates between 66kbps

and 132kbps. The latest version allows for a maximum 8 channels of

audio, up to 96kHz in sampling frequency, and a lossless encoding

option which can be transmitted in either a scalable (also known as

High-Speed Transfer mode) or standard (aka Standard mode) format.

The feasible bit-rate range has also expanded, allowing from a low of

8kbps up to 1400kbps in lossy encoding modes."

Now THIS is intriguing. 8 channels of audio, up to 96kHz in sampling frequency... We're talking high resolution surround sound 5.1 (or even 7.1) channels here.

Sony definitely has plans for future professional audio storage in mind, with ATRAC. Combine this with those old forgotten rumors of 2GB or even 5GB Hi-MDs, and things get interesting.

I doubt our cute little consumer MD models will have much to do with this, but ATRAC is certainly evolving.

Link to comment
Share on other sites

How about a completely different theory on AAL. What seems to be the general consensus is that ATRAC Advanced Lossess (AAL) is just a wav pack add-on for the user-defined lossy bitrate file.

I have a few problems with this theory first off.

1) Any lossy file throws out audio based on what is needed and not needed. Therefore a smaller lossy file would need a larger wav-esque file to fill the void.

-FACT: AAL at 64kbps creates a smaller file than AAL 256kbps (contradicting the first statement)

2) Any lossy file introduces artifacts based on the encoder. This would be very hard to [predict] and recreate the perfect lossless audio especially at the speed SS currently rips the song at.

My Theory

AAL is not a wav-pack. In fact there is nothing special about the way AAL encodes the lossless audio (beyond the general universal compression). The only difference between AAL and Windows Media Audio Lossless (WMAL) is that AAL was designed to meet a different need.

A few posts back it was mentioned that ATRAC3+ was designed with many data layers in mind. Only a few of these are used for the lossy data. This was so that if the bandwidth of transmission wasn't great enough to transmit all of the full lossless data it would default on the smaller lossy data instead. This is evident when transferring audio to the minidisc player.

My theory is that AAL consists of two parts:

1) Lossy ATRAC3+ that we are familiar with. This could exist in some of ATRAC3+ data layers.

2) Lossless independent part. This exists in some of the other layers.

Proof:

I have ripped 2 CD tracks from Linkin Park's Hybrid Theory album (don't hate me for my tastes in music, it was the first CD on the desk).

Test Tracks:

Crawling 3:28

In The End 3:36

I then ripped these tracks in WMAL, AAL64kbps, WMA64kbs and Atrac 3+ 64kbps.

File Sizes:

Windows Media Audio:

Crawling WMAL: 24.3MB

In The End WMAL: 24.6MB

Crawling WMA64: 1.62MB

In The End WMA64: 1.68MB

Atrac3+ Audio:

Crawling AAL: 25.94MB

In The End AAL: 26.29MB

Crawling 64: 1.62MB

In The End 64: 1.67MB

Right off the bat it is obvious that AAL is larger than WMAL but this will be explained in the parts to come.

Subtracting the size of Atract 3+ 64kbps from the AAL64kbps (effectively removing the 64kbps lossy file) results in this:

Crawling AAL - 64: 25.94MB-1.62MB = 24.32MB

In The End AAL - 64: 26.29MB-1.67MB = 24.62MB

Look how close these file sizes are to WMAL file sizes:

Crawling

WMAL: 24.3MB

AAL: 24.32MB

In The End

WMAL: 24.6MB

AAL: 24.62MB

Very close although there is a slight discrepancy. My first thought was that the meta data portion of the audio file in Atrac was larger than that of WMAs. That's why I compared the 64kbps versions.

Crawling:

WMA64: 1.62MB

A3+64: 1.62MB

In The End:

WMA64: 1.68MB

A3+64: 1.67MB

This showed that the results were almost identical with a larger file size favoring the WMA64 of In The End.

My conclusion is that either the MB display of the file properties is rounded off too much or that (more likely) the difference in file size is a direct result of the differences in Lossless compression techniques used.

Therefore my theory is that: AAL = Separate(Normal Lossless Packing + Lossy File)

Link to comment
Share on other sites

> Enhancments include:

> As mentioned before, ATRAC Advanced Lossless has been introduced.

> There are four bitrates: 256, 132, 128 and 64kbps at normal and high quality settings.

> When converting from CD to AAL, you can choose from one of those four bitrates.

> AAL cannot be transferred over to an exising ATRAC device; it will be converted to

> ATRAC3plus 256kbps.

Looks like there's a few debates on as to what this new Lossless does...

My knowledge is that the lossless file created includes the lossy file, and when you try to transfer this lossless file to device the SonicStage extracts this lossy file out of the lossless file to save space on the device. That is why you need to choose which baseline bitrate you want to use for PD transfer, i.e. 64, 128, 132, 256kbps.

So, the above description is incorrect. It does not automatically coverts into 256kbps. Again, it extracts the lossy file based on the bitrate selection you've made.

Hope this helps.

Cheers,

Link to comment
Share on other sites

My conclusion is that either the MB display of the file properties is rounded off too much or that (more likely) the difference in file size is a direct result of the differences in Lossless compression techniques used.

Therefore my theory is that: AAL = Separate(Normal Lossless Packing + Lossy File)

I have to say this makes even more sense.

There a similar technology used for digital imaging and often defined as Pyramid

Basically multiple layer at different resolution of the same image are available in a dataset. The final filesize is bigger but the performace is higher since a device can render the image based on its capability without too much resample and processing time.

What is not clear is if converting from AAL(XXbps) to another YYbps, SS3.3 uses the lossles layer or the lossy one.

Cheers

Edited by LupinIV
Link to comment
Share on other sites

One of you technical geniuses please figure out how to get the Hi-MD units to record 352 on the opitcal input!! PLease!!

Still this is all good news. If Sony is able to provide thid and a valid lossless codec on portable devices for recording they will have something valuable on thier hands.

It would be like a recorder encoding .flac directly.

Edited by Jammin72
Link to comment
Share on other sites

One of you technical geniuses please figure out how to get the Hi-MD units to record 352 on the opitcal input!! PLease!!

Forget it.

The internal Software must be changed. And I don't think, that Sony will do something about that.

Still this is all good news. If Sony is able to provide thid and a valid lossless codec on portable devices for recording they will have something valuable on thier hands.

It would be like a recorder encoding .flac directly.

Hmm, I'm wondering, what will be presented at the CES.

why do you need optical input if you can use a PC to convert the audio digitally

For Satellite Radio.

The german public broadcasters have joined forces and have rented a transponder on the Astra-Satellite,

using it to distribute almost all public radio stations available throughout germany, about 80 radio stations in one pack.

The point here is, that the codec used is MP1 Layer 2 @ 320k.

And to preserve that quality, you have to use higher bitrate, as we have one decode - encode cycle.

Edited by jadeclaw
Link to comment
Share on other sites

Forget it.

The internal Software must be changed. And I don't think, that Sony will do something about that.

Hmm, I'm wondering, what will be presented at the CES.

For Satellite Radio.

The german public broadcasters have joined forces and have rented a transponder on the Astra-Satellite,

using it to distribute almost all public radio stations available throughout germany, about 80 radio stations in one pack.

The point here is, that the codec used is MP1 Layer 2 @ 320k.

And to preserve that quality, you have to use higher bitrate, as we have one decode - encode cycle.

If its in MP1 then to preserve quality you will be looking at recording in PCM. The recorders will still only offer Hi-SP because that is all they were designed to record upto (as has been mentioned)

If you record in PCM, import into SS and convert to SP352 then that will retain as much quality as possible.

Personally I think you will be hard pushed to hear a difference between Hi-SP and MP1 320K as MP1 is not as good at removing unused audio data as Atrac is, its an older encoding method.

Link to comment
Share on other sites

Hi guys...I know everyone is talking about the lossless bit rate on SS.3.3 but I have a question about SS.3.3.I´ve just downloaded it and have installed it in my computer but it is in Spanish.My Windows is in Spanish but I´ve been able to have previous SS´s in English.Is it possible to change the language setting within SS.3.3.Thanks guys....amazing Forum....Keep up the Great work

Link to comment
Share on other sites

Ok I've ripped a CD to wav, I don't get any higher/lower quality option. Then I transfer this as 352kps to the HiMD, I still don't get any higher/lower quality options. Do you only get them when ripping directly to Lossless and ATRAC? I wonder how high quality HiSP (256) compares to 352kps?

Edited by Sparky191
Link to comment
Share on other sites

Guest
This topic is now closed to further replies.

×
×
  • Create New...