I plan on ripping my collection to my PC and need to choose a format. I was thinking MP3 at 320Kbps but also considered lossless. Is there going to be a big difference between 320Kbps and lossless?

I worked for for a satellite tv provider and we always had our audio bitrates fairly low in comparison. But then again it didn't sound the best either. But who really needs spectacular audio on their news channel :)

It's highly unlikely you'd be able to tell the difference in sound quality between 320 and lossless. Lossless arguments are largely for archive purposes (perfect backup copy), although some swear by lossless for playback too.

Here are some basic graphs showing some differences at different compression levels:

http://www.bradjudy.com/audioblog/2005/09/11/affects-of-data-compression-on-frequency-content/

I ripped my collection to 320k MP3 and they sound very good, but not as good as lossless. I've done direct comparison of the 320k .MP3's against .WAV and .FLAC and you can definitely hear the difference. You can notice the difference in cymbals especially.

I chose .MP3 because of the disc space savings, but now that disc prices are so low...I would use a lossless format.

I ripped my collection to 320k MP3 and they sound very good, but not as good as lossless. I've done direct comparison of the 320k .MP3's against .WAV and .FLAC and you can definitely hear the difference. You can notice the difference in cymbals especially.

I chose .MP3 because of the disc space savings, but now that disc prices are so low...I would use a lossless format.

There is a difference, i would go lossless or just leave it ripped in wav. You are talking about rip and encode. Ripping is simply extracting wavs off of the cd.

The differences mainly occur in dynamic range and high frequncies (16khz+)

I have compared lossless and 320kbps and I really can't tell a difference. Not that it isn't there, but I don't think my ears are good enough to detect it. Plus I listed to mostly Hard Rock and some Hip-Hop, which are two of the more forgiving genre's. If you listed to a lot of classical and Jazz, you may notice the compression more.

I also think the play back equipment makes a huge difference. If you have a $10K stereo setup you will probably want to ensure the best rip as possible to get the most of your system. If you have a lower or mid range setup the differences may not be as noticable.

I rip mine to wave and then convert to FLAC. Best combination I've heard. Granted, not everything will play back FLAC.

Bryan

I agree with Bryan for a few reasons.

- Ripping is tedious, and I only want to do it once. By ripping to WAV, you have a lossless copy.

- Lossless will produce the best sound for home listening, and some cool music servers are out with more coming.

- Once you have lossless Rips, you can have the computer run over night to produce FLAC, MP3, or whatever new standard comes out for you car, Portable, etc.

- If we get to a point where WAVs are not supported on computers (several years from now), you can convert the WAVs to whatever the new standard is with little work and no loss in quality.

Unless you can't afford the disk space to go with WAV files, I would Rip to lossless/WAV, then encode to whatever lossy format you need for your portable/mobile devices.

And if you use EAC to do the WAV's, you can set it up to automatically do the FLAC convert while you're ripping the next WAV.

Bryan

Rip it to lossless, rip it once, and be done with it.

WAVs are a waste of space, not to mention there's no tagging with them, and there's no benefit (quality) over lossless formats.

And FWIW, J River Media Center offers secure mode ripping (similar to EAC), and will compress on the fly to FLAC, Monkey's Audio (APE), or WMA Lossless. It can be configured to rip/download info automatically when you insert a disc as well.

Guys maybe a bit OT but what is the easiest way/application to convert my FLACs to CDA directly burnt on a CD so that I can use it on a CDP. I guess what I am looking for an inverse logic to the J River Media Center. Thanks

JRMC ought to do it. Otherwise I think there's a plugin for Nero.

I already have Nero on my PC so I guess now I need to look for the right plugin.

I already have Nero on my PC so I guess now I need to look for the right plugin.

Try googling for neroplugins, all one word.

found one at http://www.bitburners.com/Nero_Audio_Plugins_by_MauSau/

BTW any good sites other than www.allofmp3.com for downloading FLACs?

I ripped my collection to 320k MP3 and they sound very good, but not as good as lossless. I've done direct comparison of the 320k .MP3's against .WAV and .FLAC and you can definitely hear the difference. You can notice the difference in cymbals especially.

Well, that depends on what you use to encode the mp3s. The recommended LAME encoder and settings should produce virtually 'transparent' mp3s at that bitrate, though there may be *some* 'killer' tracks that will show artifacts when compared to source in an ABX. It should not be a general problem, unless you have absolutely extraordinary hearing and do all your listening on headphones. FWIW when I was using LAME and VBR (variable bitrate) encoding with --alt preset standard settings, I could not ABX any resulting mp3s from their sources, which included rock, jazz, and classical.

I chose .MP3 because of the disc space savings, but now that disc prices are so low...I would use a lossless format.

Ditto. I only convert my archival flacs to mp3 for my wife's Ipod.

Well, that depends on what you use to encode the mp3s.
no matter what mp3 encoder and what bitrate is choosed,you always lose because mp3 is not lossless and maybe never will be.
if you can't hear the differences,open the mp3 and the source wav in your audio editor,compare the waveforms ...the differences and what you lose are there.
...or convert the encoded mp3 as wave again....now your .wav from mp3 sounds worse.

regards.

No one EVER claimed mp3 was lossless, least of all me. Guess what: CD is 'lossy' too: you 'lose' frequencies above 22 kHz in the original signal. The thing is, you can't *hear* the loss. Now reread what I said about mp3. See the parallel?

I DID agree that lossless formats were preferred for archiving (because new artifacts are introduced in transcoding from lossy to other lossy formats) , so I'm not sure why you even brought up the transcoding issue. But curiously, you cited the one kind of transcoding that ISN'T likely to sound worse! Mp3 to *.wav* or other 'lossless' format should not introduce any *new* artifacts, unless you've got a very bad transcoder indeed; it's transcoding from m3 to other *lossy* formats, that's likely to produce an audible hit to quality (or going from lossy-->lossless-->lossy).

Mp3 encoders have been highly tweaked over the last decade or so, based on sound psychoacoustic principles and controlled listening tests, so to imply that all mp3s are created equal is to ignore tons of work e.g., at www.hydrogenaudio.org, where the LAME recommendation comes from. It's easy enought to encode *your own* mp3s and set up *your own* ABX tests of mp3s to .wavs, as I did, the software links are also at ha.org. *Those*, and not spectral views of mp3s vs wavs, will give you a better idea of what is and is not audibly transparent to you. (The misleading nature of spectral analysis for mp3 quality evaluation, is discussed at HA).

if you can't hear the differences....

Then it doesn't matter if they exist.

We're talking about ripping for listening here, not ripping for archive purposes.

No one EVER claimed mp3 was lossless, least of all me.then what means "depends on what you use to encode the mp3s" in your answer when you quoted Rupert?

Guess what: CD is 'lossy' too: you 'lose' frequencies above 22 kHz..means that all formats are lossy,only depend if the source have for example 1.000.000Khz but
The thing is, you can't *hear* the loss. as we are talking about audio and we can't hear 22Khz then....cd is not lossy...

But curiously, you cited the one kind of transcoding that ISN'T likely to sound worse!...It's easy enought to set up *your own* ABX tests of mp3s to .wavs, as I did, the software links are also at ha.org. *Those*, and not spectral views of mp3s vs wavs, will give you a better idea of what is and is not audibly transparent to you. i have one idea.
host 3 files:
1- source from cd 44.1-16bit,
2- the source encoded as mp3(best bitrate/encoder of your taste)
3- the mp3 back to any format that you chose.

the whole forum will apriciate to hear,i'm sure beacuse don't need any ABX test.

...the resume is:
no matter in what format,if wav or other from mp3,you always lose and schapman43 posted:
Is there going to be a big difference between 320Kbps and lossless?

at www.hydrogenaudio.org, where the LAME recommendation comes from. ...(In fact the misleading nature of spectral analysis for mp3 quality evaluation, is discussed at HA).in HA have tests for low bitrates and seems that they are bored about size and not for quality....read his tests in this thread,you will find lots about 80k,170k,etc: http://www.hydrogenaudio.org/forums/index.php?showforum=40
...i can't find nothing really serious about quality!
anyone found?

We're talking about ripping for listening here, not ripping for archive purposes.(sorry,i can't imagine if the target is for listening or archive, it's not clear in the first post :
I plan on ripping my collection to my PC and need to choose a format.

regards

then what means "depends on what you use to encode the mp3s" in your answer when you quoted Rupert?

It means that if you use a Xing encoder instead of LAME at the same bitrate, you may well get an mp3 that sounds worse than the original. It means that if you use 128 kbps instead of 320 kbps, you may well get an mp3 that sounds worse than the original. It doesn't mean mp3 is EVER lossless, in terms of data.


means that all formats are lossy,only depend if the source have for example 1.000.000Khz but
as we are talking about audio and we can't hear 22Khz then....cd is not lossy...


And if the listener can't hear the data losses in an mp3, then it's not *perceptually* lossy to the listener. This is the whole point of perceptual encoding.



i have one idea.
host 3 files:
1- source from cd 44.1-16bit,
2- the source encoded as mp3(best bitrate/encoder of your taste)
3- the mp3 back to any format that you chose.


Well, what format you choose *is* going to make a difference. Mp3 transcoded to *another* lossy format (even another flavor of mp3) runs a very good chance of pushing the artifacts above audibility threshold. Mp3 to a 'lossless' format (e.g., wav or flac) doesn't. When you convert mp3 to .wav, you are sampling the mp3 to 'CD' standard -- 44 kHz/16bit. That is more than enough to capture all the information in the mp3, with zero loss.


the whole forum will apriciate to hear,i'm sure beacuse don't need any ABX test.

I'll be happy to give it a try if you like (I've already done them in the past). But *I* get to pick the MP3 encoding, because which encoder is used, and what settings are used, *matters*.

...the resume is:
no matter in what format,if wav or other from mp3,you always lose and

Except, *data* loss does not NECESSARILY result in *audible* loss.


schapman43 posted:
Is there going to be a big difference between 320Kbps and lossless?

No, there isn't, in terms of *audibility*, assuming the encoding to 320 was doen witha good encoder. Differences are more likely in the *portability* of the file, and in whether it can be transcoded without quality drop.

in HA have tests for low bitrates and seems that they are bored about size and not for quality....read his tests in this thread,you will find lots about 80k,170k,etc: http://www.hydrogenaudio.org/forums/index.php?showforum=40
...i can't find nothing really serious about quality!
anyone found?


Yes, lots. Look in the 'Listening Tests' subforum, for example, on this page:

http://www.hydrogenaudio.org/forums/index.php?act=idx

Or read the 'pinned' threads at the top of the MP3-General subforum. Note that the primary focus is on optimizing low bitrate performance quality *because* they give the smallest file size -- in the best of both worlds, one would have small files that sound excellent. This is increasingly less of an issue for personal use because prices of storage are going down so much, but it's still desirable for transmitting files over the internet. The number of people who have ever reported being able to ABX *high* bitrate mp3s is *tiny*. Miniscule. It's considered a nonissue by all but self-appointed audiophiles who probably haven't ever tried identifying them in an ABX. In fact, HA.org won't even accept posts of such claims unless htey are backed up with some blind comparison results. There's a reason for that. I suggest you spend a bit more time on HA.org. YOu will find plenty of discussion about the characteristics of various bitrates, encoders, lossy formats, and lossless formats. Or, do some ABX comparisons of your own, and report the results. No one should assume mp3 must sound bad because they are 'lossy'.

krabapple
first i want to thank you for your educated answers and trust me,i always read your posts and apriciate your oppinions. :)

And if the listener can't hear the data losses in an mp3, then it's not *perceptually* lossy to the listener. This is the whole point of perceptual encoding. good explanations,you got the whole "feeling".

Well, what format you choose *is* going to make a difference.sorry,i was unclear.
i mean that when you encode the mp3 back to wav(for example) and compare with the wav source,you will listen the differences between the waves(original source vs wav from mp3) without any ABX.

When you convert mp3 to .wav, you are sampling the mp3 to 'CD' standard -- 44 kHz/16bit.only one clarification about cd standard:
wav format contains uncompressed audio in the pulse-code modulation (PCM) format.WAV is a data file format for computer use.
Audio CDs do not use WAV as their sound format, instead using Red Book audio.
The commonality is that both audio CDs and WAV files have the audio data encoded in PCM.
from wikipedia: http://en.wikipedia.org/wiki/Wav#Audio_CDs

Differences are more likely in the *portability* of the file, and in whether it can be transcoded without quality drop.nothing against but this is exactly what HA take care and not for real quality.

Yes, lots. Look in the 'Listening Tests' subforum..seems cool but not all members are really serious and have lots of kids,if you know what i mean.
as you know and recommend,ABX is for personal tests because you never will know how much another person can hear(to tell the minimum) and we can't use his results as base for anything.
is not a question of believe in their results,is how,where,what amplifier,what pc,what soundboard,what headphone and thousands more details.
then i quote you againAnd if the listener can't hear the data losses in an mp3, then it's not *perceptually* lossy to the listener. This is the whole point of perceptual encoding. do you understand my position of his tests now?


No one should assume mp3 must sound bad because they are 'lossy'. i agree with you....and don't lol....i like AC3 and it's lossy. :p
download the penteosurround sample to feel that the source is very important.
for me the source have first importance...bad sources,bad results.

my sincere thanks and best regards!
:)

edited: thousands typos(and have more)

sorry,i was unclear.
i mean that when you encode the mp3 back to wav(for example) and compare with the wav source,you will listen the differences between the waves(original source vs wav from mp3) without any ABX.


And I repeat : *possibly*. It depends primarily on how well the initial mp3 encoding was done. It;s also possible you will NOT be able to pass an ABX of the two to waves.

only one clarification about cd standard:
wav format contains uncompressed audio in the pulse-code modulation (PCM) format.WAV is a data file format for computer use.
Audio CDs do not use WAV as their sound format, instead using Red Book audio.
The commonality is that both audio CDs and WAV files have the audio data encoded in PCM.
from wikipedia: http://en.wikipedia.org/wiki/Wav#Audio_CDs

That was not in dispute. WAV files tend to be sampled at CD rates.
Also from the wikipedia entry for wav: "Though a WAV file can hold compressed audio, the most common WAV format contains uncompressed audio in the pulse-code modulation (PCM) format. PCM audio is the standard audio file format for CDs at 44,100 samples per second" So, one would not expect that converting an mp3 to .wav would introduce audible problems. Therefore, one would again expect that its the INITIAL ENCODING TO MP3 that is the crucial variable. Do that well, and you stand a good chance of producing an mp3 that sounds identical to its source, in an ABX.



nothing against but this is exactly what HA take care and not for real quality.

You base this one *WHAT* research exactly? I';ve been reading HA for several years now, quality is *certainly* a real concern.

seems cool but not all members are really serious and have lots of kids,if you know what i mean.
as you know and recommend,ABX is for personal tests because you never will know how much another person can hear(to tell the minimum) and we can't use his results as base for anything.
is not a question of believe in their results,is how,where,what amplifier,what pc,what soundboard,what headphone and thousands more details.
then i quote you again do you understand my position of his tests now?

Do your own then. Describe how you encoded the MP3s, all your ancillary equipment, and your ABX results. Post them the HA.org. You might learn something.

I ripped my collection to 320k MP3 and they sound very good, but not as good as lossless. I've done direct comparison of the 320k .MP3's against .WAV and .FLAC and you can definitely hear the difference. You can notice the difference in cymbals especially.

I chose .MP3 because of the disc space savings, but now that disc prices are so low...I would use a lossless format.

I feel the same. There is no debate that high bitrate (& optionally, vbr) MP3/AC3/OGG sound very good but there's that slight difference that only shows up during certain passages (like you mentioned: cymbals), and very rarely, noticable noise artifacts.

Disk is cheap. There's no reason to not rip to a lossless format like flac. Use a redundant disk array and never have to rerip your cd's. IMO the only reason to use a lossy format is for portable device use (limited storage) or transmission (limited bandwidth)

If you're pinching pennies building your media server I would jsut spend that money on music or other upgrades in your audio system until disk storage gets even cheaper.

You might learn something.
let me tell you what i learn:

Describe how you encoded the MP3s i don't do that.
i learn that mp3 is for internet,divx and now for mobiles where space is the target and quality is in the second place.
mp3 don't sounds good for my taste.

...
wow (i'm reading the "preview post" when was writing this)
here is the real good taste for who want qualityDisk is cheap. There's no reason to not rip to a lossless format like flac. Use a redundant disk array and never have to rerip your cd's. IMO the only reason to use a lossy format is for portable device use (limited storage) or transmission (limited bandwidth)
wonderful post ganto

;)
regards for all

I ripped my first CD yesterday from FLACs. So the process I followed was to download FLAC file from www.allofmp3.com and burn them directly to CD using NERO with FLAC -> CDA plugin (so no need to convert to WAV before burning to a CD). When I put that CD on my system it somehow did not sound as good as CD. Maybe psycho acoustics are at play here but I wanted to confirm that this process theoretically should give me the exact performance of a CD right? The other problem is that I don't have any of the downloaded songs on a CD previously to compare.

Am I better off downloading the PCM WAV in some way compared to FLAC? I know the answer technically is "no" but would definitely appreciate a confirmation from someone who has played with this stuff longer than me.

Sincerely,
-dollarman

I ripped my first CD yesterday from FLACs. So the process I followed was to download FLAC file from www.allofmp3.com and burn them directly to CD using NERO with FLAC -> CDA plugin (so no need to convert to WAV before burning to a CD). When I put that CD on my system it somehow did not sound as good as CD. Maybe psycho acoustics are at play here but I wanted to confirm that this process theoretically should give me the exact performance of a CD right? The other problem is that I don't have any of the downloaded songs on a CD previously to compare.

FLAC is mathematically lossless. You get out, exactly what you put in, like zip.

Am I better off downloading the PCM WAV in some way compared to FLAC?

No.

I know the answer technically is "no" but would definitely appreciate a confirmation from someone who has played with this stuff longer than me.

The answer is unconditionally no, FLAC is mathematically (not psychoacoustically, or perceptually), it is actually lossless.

was to download FLAC file from www.allofmp3.com and burn them directly to CD using NERO with FLAC -> CDA plugin ... When I put that CD on my system it somehow did not sound as good as CD.like stanger89 posted: FLAC is mathematically (not psychoacoustically, or perceptually), it is actually lossless.is the right answer but how allofm3.com did this flacs(have some informations/logs from the site about this flacs?) and how good is the nero plugin(flac-->cda)?

if you can download PCM WAV maybe will sounds better because you'll need less "steps"(convertions) to burn.

regards.

edit:
from the faqs in allofmp3,seems one cool option:
"There is also an advanced "Online Encoding Exclusive" (OEEX) feature for retaining the original quality of a recording in the following formats: Monkey's Audio, WMA 9 Lossless, FLAC, and OptimFrog. OEEX enables you to encode music using the original audio source — a compact disc."

... and what was the intermediate step to get them into FLAC to start with? Again, something like EAC (Exact Audio Copy) will make a bit for bit copy of the CD, then FLAC does it's lossless thing. If a plugin or another format was the intermediate step prior to FLAC then it's quality is likely the culprit as was previously described.

Bryan

let me tell you what i learn:

i don't do that. i learn that mp3 is for internet,divx and now for mobiles where space is the target and quality is in the second place.
mp3 don't sounds good for my taste.


Then I have to wonder, again, whether you have ever compared a well-made mp3 fairly to its .wav source in a blind test.

I have encoded all my wife's CDs to VBR (~198 kbps avg) mp3, with LAME, after ripping them with EAC, and then dumped them on her Ipod. I would bet good money that most if not all of the tracks would be audibly indistinguishable from their wav source, in ABX comparisons. A couple of dozen CDs worth and she's still got over 20 Gb left on her 30 Gb toy. of course, if I went with 128 kbps constant bitrate, I could fit far more tracks on there. And guess what, if I kept the same production pipeline (EAC-->LAME) the hit to quality would be relatively small, though more likely to be audible more often.

That's what *I* learned from HA...and from my own comparisons.

I ripped my first CD yesterday from FLACs. So the process I followed was to download FLAC file from www.allofmp3.com and burn them directly to CD using NERO with FLAC -> CDA plugin (so no need to convert to WAV before burning to a CD). When I put that CD on my system it somehow did not sound as good as CD. Maybe psycho acoustics are at play here but I wanted to confirm that this process theoretically should give me the exact performance of a CD right? The other problem is that I don't have any of the downloaded songs on a CD previously to compare.


[edit]

I see now that you were starting with a flac made by *someone else*. So there's all sorts of reasons it might not 'sound like CD'. You can test someof them by downloading a flac of a song you already have on one of your CDs. Decod ethe flownloaded flac to a .wav file. Then compare the .wavs of the decoded flac, to the wav of the same track, ripped from your CD. Use EAC's wav comparator to see if they're digitally identical.
- If they are, then the 'difference' you heard was probably 'psychoacoustic'.

- If they aren't, try telling them apart in a listening test with ABX software.
----If you can't, then the heard 'difference' was probably 'psychoacoustic'.
----If you can, it was probably due to the real differences between the .wavs.

I ripped my first CD yesterday from FLACs. So the process I followed was to download FLAC file from www.allofmp3.com and burn them directly to CD using NERO with FLAC -> CDA plugin (so no need to convert to WAV before burning to a CD). When I put that CD on my system it somehow did not sound as good as CD. Maybe psycho acoustics are at play here but I wanted to confirm that this process theoretically should give me the exact performance of a CD right? The other problem is that I don't have any of the downloaded songs on a CD previously to compare.Ummm, I always thought that allofmp3 was a grey market store in Russia. I've heard various complaints about poor quality rips (before encoding) and low bit rates on mp3s (regardless of what they say the bitrate is).

I would strongly suggest that you rip (with EAC) and then encode to FLAC and then use your FLAC->CD process for a comparison.

@ Krabapple

I have encoded all my wife's CDs to VBR (~198 kbps avg) mp3....on her Ipodseems cool for Ipod,is your target but i hear only in stereo or 5.1 system.

of course, if I went with 128 kbps constant bitrate, I could fit far more tracks on there.of course much more tracks but with worse quality(horrible) than vbr or 320cbr!
this "temptation" is to do exactly what you learned in HA as i wrote : space is the target and quality is in the second place.

mp3 don't sounds good for my taste.
...That's what *I* learned from HA :p

regards. :)

edit:
ok,maybe you can change my mind(not kiddin)...a long time i don't encode mp3.
(host)post one .wav and the mp3 encoded with the best bitrate/encoder.
i download and test the quality....can you?
;)

@ Krabapple

seems cool for Ipod,is your target but i hear only in stereo or 5.1 system.

of course much more but with worse quality(horrible) than vbr or 320cbr!

No, not *horrible*...actually more like *possibly* worse-sounding. Have you ABX-compared *well-encoded* 128 kbps mp3s to any made at higher bitrates, or to .wav? Have tyou read the test results for 128 kbps from various encoders, on HA? 'Horrible' seems a tad overwrought.


this "temptation" is to do exactly what you learn in HA as i wrote :

I'm afraid I don't believe you've really read much there, sorry. If you did , you wouldn't be yakking about 'horrible' sounding 128 kbps mp3s.




That's what *I* learned from HA :p

regards. :)

And to lump all mp3s together like that shows you haven't.

sorry,i edit my last post,can you read only the final again please?
:)

post one .wav and the mp3 encoded with the best bitrate/encoder.
i download and test the quality....can you?

To be clear -- you want me to upload a 320 kbps LAME-encoded mp3 (the theoretical best quality), and the wav it came from? Or are you asking to hear a good 128 kbps mp3?

(One problem with this approach fior blind comparison is that you can tell the files apart simply by their size. Would you mind if I converted the mp3s to .wavs again?)

you want me to upload a 320 kbps LAME-encoded mp3 (the theoretical best quality), and the wav it came from?yes,you know how to do that in best way.as i told you,a long time i don't encode mp3 and if the sound is cool as you tell (i can bet,i can trust in your good taste),i will use that too.why not? ;)

Would you mind if I converted the mp3s to .wavs again?will be cool!!!

thanks in advance.
:)

edit:
oh God....typos!

Jorge, could you name some popular or well-known CDs that you are familiar with, that I might have? I would prefer to encode a song you know, to give you the best chance possible. I've got lots of rock, jazz, and classical stuff so we should be able to come up with something we both have.

i like lots:
King Crimson,Beatles(solo too),Stones,yes,l.zep,megadeth,hendrix,elo,the who,santana,dream theater,...your choice.
i have round 400 cdas,round 300 lp (and round 52.000 mp3/320k very outdated(older encoders)).

...and thanks,you're very kind!
:)

OK, King Crimson. Pick a CD and a track and I'll set it up.

from "the compact King Crimson" - "Frame by Frame". re-mastered by Robert Fripp & Tony Arnold,1989(but not the digital remastered version)

if you don't have this cd,i can extract and host the .wav 44.1k-16bit and pm for you....or any other option,your taste!

regards!
:)

I no longer have the 1989 remaster of 'Frame by Frame', sorry (it's definitely a digital remaster so I'm not sure what you meant by "but not the digital remastered version'). I only have the most recent remaster. So please host the .wav (or convert it to flac, for faster downloading), and let me know where to dl it.

i mean digital remaster as "gold edition from master source".

hosting the wave as .rar.....is ok?
i post the url from rapidshare here or pm?

post here

54.4Mb wav inside 35.4Mb rar

Download-Link #1: http://rapidshare.de/files/31486668/KK-FBF_Track_06.rar.html


thank you krabapple.
:)

This may be useful in comparing the two:
http://www.kikeg.arrakis.es/winabx/

I'm the author of Kexis.. which was a precursor to FLAC.. they share no code but FLAC uses many of the same principals that kexis did. I think Kexis is still listed on the FLAC site as a comparison for bitrate, compression, ect.

Numerous listening tests by I and others show that on a reference system you can tell the difference between raw pcm and vbr encoded bitstreams, even at 320kbits/sec. The degree depends on the sourcematerial and how large the bitbucket is that the VBR algorithim is using.

Bigband is one of the "hardest" recordings to compress losslessly because of the range of hz that is present at the same time. Subsequently VBR also has a hard time. On a good reference setup you'll hear this on transitions where decay is relevant. For instance someone mentioned Cymbals previously. You'll hear cymbals "smear" and not ring clear. They will smear into the background. Its slight and it happens fast but its there. The lower the bitrate and the smaller the bitbucket, the longer the smear.

You ahve to know what you are listening for and most people don't, but its there. You also need to be able to produce a full range of sound and most setups to one degree or another can't.

Personally I rip everything to FLAC and forget about it. I have a parallel directory tree that is generated by a script that convers the FLAC into Ogg Vorbis for my portable. In house I stream to Squeezeboxes direct from flac... the Slimserver decodes the FLAC and tosses the pcm directly to the squeezebox.

Best of both worlds. Easy, fast, and space effecient.

This may be useful in comparing the two:
http://www.kikeg.arrakis.es/winabx/

Thanks. That's the tool I was going to suggest Jorge use fr the comparisons :) .


Some suggestions for running one's own ABX trials:

"what is a blind ABX test?"
http://www.hydrogenaudio.org/forums/index.php?showtopic=16295

I'm the author of Kexis.. which was a precursor to FLAC.. they share no code but FLAC uses many of the same principals that kexis did. I think Kexis is still listed on the FLAC site as a comparison for bitrate, compression, ect.

Numerous listening tests by I and others show that on a reference system you can tell the difference between raw pcm and vbr encoded bitstreams, even at 320kbits/sec. The degree depends on the sourcematerial and how large the bitbucket is that the VBR algorithim is using.

Never said it *couldn't* be done. In fact, I noted that a miniscule number of testers have reported ABX success at the 'best' bitrates. I have never heard of *numerous* people doing so at high bitrates (VBR or CBR) using a recommended LAME build. Can you point me to some such test results?

Bigband is one of the "hardest" recordings to compress losslessly because of the range of hz that is present at the same time. Subsequently VBR also has a hard time. On a good reference setup you'll hear this on transitions where decay is relevant. For instance someone mentioned Cymbals previously. You'll hear cymbals "smear" and not ring clear. They will smear into the background. Its slight and it happens fast but its there. The lower the bitrate and the smaller the bitbucket, the longer the smear.

A big band would have no more of a frequency range going on at the same time than a symphony orchestra -- probably less. And if we're talking about big band recordings from the Big Band era, those old discs and tapes aren't likely to be full range. I don't recall big band recordings being on the 'hard to encode' file list, used to weak LAME, but I haven't heardall of them. There's some harpsichord
tracks, and a killer called 'fatboy', though.

I have said that even with the best encoding, there is fraction of situations where one might still pass an ABX -- usually source-related (e.g., using a 'killer' track). We shall see what happend with King Crimson. To my knowledge, only one or two people on HA have ever reported successful ABXing of mp3s encoded usig late-model LAME at 'insane' (320 kbps constant bitrate) presets. They used headphones , IIRC, which is what I suggest others do as well to increase their chances of success, no matter how 'highly resolving' they think their loudspeaker system and room are is.


You have to know what you are listening for and most people don't, but its there.

Do you think such a trained person would successfully ABX a 320 kpbs mp3 of *most* tracks, or just a relative handful of difficult-to-encode tracks?

You also need to be able to produce a full range of sound and most setups to one degree or another can't.

Most people aren't hearing that full range anyway . Few people beyond the age of 20 have much in the way of hearing near 20 kHz. Audibility of decay tails have to do with resolution and noise of the system, more than frequency response.

Personally I rip everything to FLAC and forget about it. I have a parallel directory tree that is generated by a script that convers the FLAC into Ogg Vorbis for my portable. In house I stream to Squeezeboxes direct from flac... the Slimserver decodes the FLAC and tosses the pcm directly to the squeezebox.

Best of both worlds. Easy, fast, and space effecient.

Well, I do not question this strategy AT ALL. As I've said, it is what *I* do too, except that for portable players I convert flacs to mp3 copies. The OP's question has been answered: everyone advises him to go lossless. What I question is *Jorge's* blanket dismissal of mp3 as bad-sounding . I hear this a lot from audiophiles, few of whom have followed recommended ripping/encoding practice, and none of whom (IME) had ever tried abx'ing a well-made mp3 to its source. I contend that not only are well-made mp3s NOT bad sounding, but most of the time, most listeners, including self-proclaimed audiophiles, won't be able to tell them apart from source in a fair test.

Never said it *couldn't* be done. In fact, I noted that a miniscule number of testers have reported ABX success at the 'best' bitrates. I have never heard of *numerous* people doing so at high bitrates (VBR or CBR) using a recommended LAME build. Can you point me to some such test results?

Ok I guess let me rephrase. Numerous being the discusion on the kexis and later the flac developer list. Any time there is a hard transition its noticeable to one degree or another.



A big band would have no more of a frequency range going on at the same time than a symphony orchestra -- probably less. And if we're talking about big band recordings from the Big Band era, those old discs and tapes aren't likely to be full range. I don't recall big band recordings being on the 'hard to encode' file list, used to weak LAME, but I haven't heardall of them. There's some harpsichord
tracks, and a killer called 'fatboy', though.

ITs not the frequency range thats the problem. Its the digital representation of the PCM wave. By big band I mean something like the Brian Setzer Orchestra on a modern recording. Compared to a modern Symphony Brian is ALOT harder to compress loslesss. I'll go into alittle technical detail here as to why and if you are interested I can expand on this.

I'm sure you know a pcm stream for each channel is a wave form that has a resolution of 16 bits at a frequenzy of 44.1 khz. Ok so you ahve your wave and the question is how to compress the thing without throwing anythign away. The easiest thing to do is to RLE or run length encoding. Each bit is compared to the last and if they are the same save the bit, and the number of times its the same. since each chunk of time is 6 bits, and if you have that for 100 chunks of time, you end up with a 16 bit number followed by another 16 bit number that wil ltell you how many times to expand to. this is a great savings over 100 16 bit numbers in a row.

In real life RLE for sound is mostly useless since the bits are never the same. Even dead silence has some sort of noise floor with the bits moving up and down a wee bit.

So the next step is to record the first chunk of time, then record every chunk of time thereafter is the delta, or difference between the current chunk and the last chunk. How does this help? with a little cleverness you dont have to save each 16 bit number as a 16 bit number. You only have to save it in such a way that the number has enough bits to reprsent it.

Example. Lets say your starting number is 9. Your next chunk of time is 12. in PCM thats 2 16 bit numbers for a totaly of 32 bits. If you use the abov methid you store the first 16 bit number, then you notice the next number is only 3 "more" than 9. well 3 can be represented as 0000000000000011. If you can throw out those leading 0's and jsut store the 11 that saves you 14 bits. See where this is going?

Its gets more complicated than this and there are more tricks but thats the general idea of lessless encoding. The enxt step is to PREDICT the next number based on the past wave form and save the error in the prediction of the delta. If you ahve a good predictor you can get the bits WAY down. You also ahve to have a good encoder that will encode the predicted deltas with the minimum amount of bits... which is harder than it first sounds. Then you can do neat things like prediction amount stereo channels and all other sorts of cool stuff.

Ok the point is teh wave form for soemthing like big band jumps all over the the place. It looks like a speed freak on acid. ITs hard to predict there is so much going on. A symphony is much easier to predict because the the wave form is much smoother.

You can verify this yourself. Download flac and play with it. Try big band compared to Syphony or Chamber.


I have said that even with the best encoding, there is fraction of situations where one might still pass an ABX -- usually source-related (e.g., using a 'killer' track). We shall see what happend with King Crimson. To my knowledge, only one or two people on HA have ever reported successful ABXing of mp3s encoded usig late-model LAME at 'insane' (320 kbps constant bitrate) presets. They used headphones , IIRC, which is what I suggest others do as well to increase their chances of success, no matter how 'highly resolving' they think their loudspeaker system and room are is.

First lets get something clear. All mp3's are not created equal. LAME mp3 is better than a plain old vbr mp3 implementation because its model is more intelligent. Even better yet is Ogg vorbis. Moving up the chain of encoders the become better and keeping more bits around as "spare" bits ready to encode diffcult transitions.

Can I blind ABX crappy sources? I've tried and at high bitrates its a tossup. Can I ABX clean sources with the standard vbr mp3 encoder against source? It depends on the inherant entropy of the original source. The more "energy" it has as a whole the easier it is to ABX. A slow symphony with very little transitions is impossible at this point to ABX. Rock with lots of stuff going on is easy for me if I pay attention to the transitions. It will be harder for others because they don't know what to lsiten for and therefore wont be able to distinguish... in which case that encoding is fine for them. Also keep in mind that at 44.1khz you are limited to 22.05 khz of headroom (nyquist blah blah). Mp3 encoders are in no way oblidged to keep that 22.05 khz of headroom. Some will wantonly throw away everything above, say 15khz. Thats why women have an easier time at abXing crappy encoders if you've ever wondered.

Lame and ogg as you progress do a better job allocating bits to the frequency range we hear and getting rid of masked bits that we cant. Ogg at 320kish is damn hard from me to ABX with any source material. I've tried and its damn good. I cant do it with any statistical certainty.


Do you think such a trained person would successfully ABX a 320 kpbs mp3 of *most* tracks, or just a relative handful of difficult-to-encode tracks?

As explained above depends on the quality of the ecoder. They most certainly are not created equal. Ogg vorbis is the best from my personal experience.


Most people aren't hearing that full range anyway . Few people beyond the age of 20 have much in the way of hearing near 20 kHz. Audibility of decay tails have to do with resolution and noise of the system, more than frequency response.

Bingo. Thats why alot of encoders allocate less bits to the high end, or simply filter everything over a certain point.


Well, I do not question this strategy AT ALL. As I've said, it is what *I* do too, except that for portable players I convert flacs to mp3 copies. The OP's question has been answered: everyone advises him to go lossless. What I question is *Jorge's* blanket dismissal of mp3 as bad-sounding . I hear this a lot from audiophiles, few of whom have followed recommended ripping/encoding practice, and none of whom (IME) had ever tried abx'ing a well-made mp3 to its source. I contend that not only are well-made mp3s NOT bad sounding, but most of the time, most listeners, including self-proclaimed audiophiles, won't be able to tell them apart from source in a fair test.

mp3 i not bad sounding. How good it sounds compared to the source depends on many thigns. It depends on teh quality of the source, the resolution of the play back system, the quality and model of the encoder, ect.

I encourage people to ABX test at 128k, 192k, 320k a standard vbr encoder, LAME, and Ogg. Make your own decisions.

However if you are encoding for archiving, archive in FLAC. since its already there, play back from anything attached to the archive might as well be in flac also.

I have said that even with the best encoding, there is fraction of situations where one might still pass an ABX -- usually source-related (e.g., using a 'killer' track). We shall see what happend with King Crimson. To my knowledge, only one or two people on HA have ever reported successful ABXing of mp3s encoded usig late-model LAME at 'insane' (320 kbps constant bitrate) presets. They used headphones , IIRC, which is what I suggest others do as well to increase their chances of success, no matter how 'highly resolving' they think their loudspeaker system and room are is.


I dont want to derail this thread with a headphones vs speaker argument but I just want to point out that headphones have a different set of strengths & weaknesses and their design tries to make the best comprise, just like speakers, to get the best sound quality they can.


Do you think such a trained person would successfully ABX a 320 kpbs mp3 of *most* tracks, or just a relative handful of difficult-to-encode tracks?


I personally think well encoded can sound very good and the bitrate&vbr level sets the ante at where 'problematic' passages become apparent. I know what the differences sound like, and I admit at higher bitrates without a source to compare to (no AB) it can be difficult to tell with certainty whether something has been compressed. It really depends on the complexity of the source material and whether such a problematic passage is encountered.

As already mentioned, like others at the end of the day I just go lossless, it makes sense logistics wise and avoids having to worry how good is good enough.


Most people aren't hearing that full range anyway . Few people beyond the age of 20 have much in the way of hearing near 20 kHz. Audibility of decay tails have to do with resolution and noise of the system, more than frequency response.


As far as hearing abilities go, without scientific evidence/references, here's a hypothesis: if one is interested in audio would one lose their higher end hearing ability slower than someone that didnt really spend that much 'brain' time on it, or one is interested in audio because of thier better hearing? Without proof, I lean towards the use it or lose it question... NOTE dont confuse -use it- with -ABuse it-, as that answer is fairly obvious.

@ stranger89 and krabapple
thanks for the links boys!
;)

@ wmilas
Even better yet is Ogg vorbis.i feel(hear) the same. :)

@ all
Ogg vorbis is the best from my personal experience.
bingo... lol ...when i posted the same in HA i was.......banned!!! :eek:
it's somewhere in his "recycle bin",is not hard to find. :rolleyes:

wmilas,can you post links to your homepage or about Kexis?


regards for all!

Kexis still has all its old stuff on sourceforge. You can find it there. Like I said its a dead project. I made it because I didn't like Shorten which was the only real option at the time. It was as much a proof of concept as anything else. Flac currently does everything Kexis does, only better. Flac has some really slick options :)

I dont want to derail this thread with a headphones vs speaker argument but I just want to point out that headphones have a different set of strengths & weaknesses and their design tries to make the best comprise, just like speakers, to get the best sound quality they can.


Sure. But, IMO, even a moderately priced set of cans like Sennheiser HD-580 has clarity and detail/resolution that would cost several $K to get via speakers.

mp3 is not bad sounding. How good it sounds compared to the source depends on many thigns. It depends on teh quality of the source, the resolution of the play back system, the quality and model of the encoder, ect.


...which is basically what I have been telling Jorge for several days now.

(the rest of your post was great too...and thansk for the info on modern Big Band. How about *modernist* symphony music, e.g. atonal, arrhtyhmic stuff using a full orchestra?)

@ stranger89 and krabapple
thanks for the links boys!
;)

@ wmilas
i feel(hear) the same. :)

@ all

bingo... lol ...when i posted the same in HA i was.......banned!!! :eek:
it's somewhere in his "recycle bin",is not hard to find. :rolleyes:



'jeo':

Perhaps you should have read the HA Terms of Service before posting. :rolleyes:

Your tendency to make poorly-supported blanket statements worked against you. Offering up a variant of the tired (not to mention contradictory) 'I don't care what ABX says, in the end I trust my ears' defence didn't help either.

Btw, jorge, the download is no longer valid for me, can you repost the Crimson .rar. file again? (And is tehre a reason you used rar instead of FLAC?)

'jeo':

Perhaps you should have read the HA Terms of Service before posting. :rolleyes:

Your tendency to make poorly-supported blanket statements worked against you. Offering up a variant of the tired (not to mention contradictory) 'I don't care what ABX says, in the end I trust my ears' defence didn't help either.
as you really want to talk about it...(no problem)
i don't wrote it there: "'I don't care what ABX says, in the end I trust my ears'"

here is the thread and my last post:
http://www.hydrogenaudio.org/forums/index.php?showtopic=44146&pid=388125&st=0&#entry388125
and in the post #11 i wrote: "vorbis sounds better for my taste,is really champion"

how can i have to proove my --->taste<--- for everyone and using ABX?
do you like red cars? ok but i like blue cars,you know what i mean?i break any rule if my taste is blue and still have to proove it?
then...lol...

William that feels about himself that is very wise,posted http://www.hydrogenaudio.org/forums/index.php?showtopic=40134 and dissapear. lots of people posting good and bad explanations and he is absent because he will back there only after one final solution or when he think to call admin for show TOS in the thread and apply the "appropriate actions" as he wrote.
is nothing more than "one smart(middle stup)" in HA forum.

sorry and a pity because have lots of great persons there but i'm complete incompetent to proove my taste and was banned.
have another post there when someone call me to talk about the tos but was deleted.i don't remember who was and the date, i really forgot.what i remeber is that i call the admin to be removed of the forum but they,cowardly,delete the post and ban me.
from here, i feel that the tos is flexible because only was used where and when they want and delete the whole thread that i don't create but call to be removed is one proove of total incompetence in friendship and human relations.

better forget,right?
:)

edit: if you want,later in another thread i can tell why i don't use ABX test.
i have valid reasons,one need that you start a new thread with this topic ok?

i'm saving that thread from HA in my pc because they can delete it too using the "appropriate actions" and,if needed,i host the thread somewhere.

regards.

Btw, jorge, the download is no longer valid for me, can you repost the Crimson .rar. file again?
not valid?..strange.
i have flac inside audition and it's outdated.
comparing the size have only 2mb different from .rar.

(And is tehre a reason you used rar instead of FLAC?)

:p my taste (lol) is for .rar.

regards.

edit:
the links still running:
http://rapidshare.de/files/31486668/KK-FBF_Track_06.rar.html

if have something wrong and you still can't download it,show me another place to host please.

as you really want to talk about it...(no problem)
i don't wrote it there: "'I don't care what ABX says, in the end I trust my ears'"

here is the thread and my last post:
http://www.hydrogenaudio.org/forums/index.php?showtopic=44146&pid=388125&st=0&#entry388125
and in the post #11 i wrote: "vorbis sounds better for my taste,is really champion"

I'd already seen that,and the rest of the thread, before I posted, thanks.

the mp3 can be downloaded from the link below; i've decided not to convert it to .wav for now, as that introduces yet another variable.

www.m-ideas.com/sullivan/mp3/KK-FBF Track 06.mp3

please use an ABX comparator to compare it to the original .wav file. Here again is a good ABX app:

http://www.kikeg.arrakis.es/winabx/

and here is the readme telling how to perform an ABX test with WinABX

http://www.kikeg.arrakis.es/winabx/readme.txt

and here is some more info about good ABX practice

http://www.hydrogenaudio.org/forums/index.php?showtopic=16295

I suggest doing 20 trials (or some other predetermined number greater than 16).

krabapple,
don't deserve any ABX test because the sound is... :rolleyes:

W-O-N-D-E-R-F-U-L :eek: :) :cool:

my sincere congratulations because few people have dedication to do one good encode like you did!!!

--->now tell my what need and how i have to do to encode my mp3,please.(i told you,a long time i don't do that)

...and thank you so much because you could change my mind about good mp3 codifications!
:)

I haven't blazed any new trails, but thanks just the same. This is all standard stuff from hydrogenaudio.org. Do you already use EAC to rip CDs 'perfectly'? If so, you can create an EAC configuration that will automatically create nice mp3s whenever you specify 'mp3' as output.

Go here to grab the recommended LAME encoder (currently 3.97beta3, although I think I'm still using beta2)

http://wiki.hydrogenaudio.org/index.php?title=Lame_Compiles

Unpack it in the same directory as your EAC.exe (or install it anywhere convenient, just remember where you put it)

Then go here

http://www.hydrogenaudio.org/forums/index.php?showtopic=28124

and scroll down to where it says "# Setting up EAC for LAME.exe with tagging:"

Follow the directions showing you how to set the EAC options. At step 3 use the location where you put LAME.exe.

At step 5 cut and paste this commandline option into the EAC config page

-b 320 --add-id3v2 --pad-id3v2 --ta "%a" --tt "%t" --tl "%g" --ty "%y" --tn "%n" %s %d

this is equivalent to '--preset insane' ( that's the -b 320 part, the rest is all just smart tagging to add title, album , etc)

At 320 kpbs constant rate (-b 320), it's really overkill for most files. I still use

-V 2 --vbr-new --add-id3v2 --pad-id3v2 --ta "%a" --tt "%t" --tl "%g" --ty "%y" --tn "%n" %s %d

[EDIT : added --vbr-new, which according to HA.org is faster encoding, and supposedly even more likely to be transparent -SS]

It's a variable bitrate setting, which makes smaller files with a nominal average bitrate of 190 kbps, but can go as high as 320 kbps at any given moment (or as low as 128). It sounds very good (I couldn't tell them apart from .wav when I tried).

Step 6 is optional, you can either keep or delete the original .wav

When you're done, save the EAC config, in case you ever want to use other configs (e.g., one that creates FLAC files instead of MP3s), then afterwards need to switch back to making mp3.

If you already have .wavs that you want to convert, just use EAC's 'compress files' menu.

wow,thanks.
is one complete guide,this page is bookmarked! :cool:

Do you already use EAC to rip CDs 'perfectly'? If so, you can create an EAC configuration that will automatically create nice mp3s whenever you specify 'mp3' as output.i had problems with EAC as posted here(if you want to read) :
http://www.hydrogenaudio.org/forums/index.php?showtopic=44131&hl=

i have LG cdrw gce8527b and LG dvdram gsa4081b and with EAC i have "clicks" using both but not extracting with audition and easycdda.(don't know the reason)
i read the EAC guides and did hard tests.better is do all tests again.
i'm downloading everything needed now and i tell you if the clicks with EAC will remain.

i forgot this important details in my last post:

thank you so much because the best of all is your atention.
:)

later i back!
regards.

everything working fine krabapple,
"nobody" can call lossy encoding MP3-320K following your guide,
is really good,wonderfull! :cool:

thanks so much for your (extreme)atention,(extreme)patience and (extreme)friendship.
stay around!

:)

It is been a pleasure reading this thread and thus I decided to join AVS forum.

I would like to add to this thread that you should consider some other formats than mp3 if quality matter to you. For example AAC, OGG and WMA can produce excellent quality at lower bitrates than mp3.

About 200GB, probably a little less. Lossless compression gets you about 50% savings in space, so 600 * 50% * 650MB ~ 200GB. Though most CDs are probably not a full 650MB so it's probably going to be a bit less than that.

After reading this thread I am really confused. I have always downloaded my files in mp3 at 320kbps but now I feel I may be missing something.

I looked at a typical song and to download in FLAC will take up 3x more space which is not a problem but I really would like to keep the process, within sound quality reason, as simple as possible. So am I just missing some highs, that I probably at my age can't hear anyway, if I continue to download in mp3 at 320kbps?

Thanks,
Yeto

At 320 kbps you are missing data, but it's unlikely you are taking an audible hit. However, you have compromised your ability to convert to other lossy compressed formats in the future (if you transcode, you could reach the point where you've thrown away enough data for it to be audible). You avoid that problem with lossless. And creating flacs is as easy as creating mp3s.

At 320 kbps you are missing data, but it's unlikely you are taking an audible hit. However, you have compromised your ability to convert to other lossy compressed formats in the future (if you transcode, you could reach the point where you've thrown away enough data for it to be audible). You avoid that problem with lossless. And creating flacs is as easy as creating mp3s.

I am not sure that I fully understand the transcode process. Would you have time to explain some detail on the transcode process?

I have also noticed that some older DVD players will not read the CDs that I have burned using the mp3 format. When I burn my CDs, using Nero, should I use Audio CD instead of mp3?

Thanks,
Yeto

For DVD player to be able to play CD with mp3 it must have mp3 decoder built in. Old DVD players don't have that. It also might be possible that some new DVD player won't be able to play mp3. Then maybe there are DVD player that also might not play Audio CD.

Transcoding means decoding mp3 then encoding it again to mp3. When you decode mp3 to PCM (.wav) then there is no data loss (lets put aside quantization and dithering because that has very little impact). When you encode PCM to mp3 then there is always data loss.
Encoder "removes" data that it expect you will not be able to hear. But it is not perfect and it may reach boundary between "you will not be able to hear difference" and "you will be able to hear difference". You encode once from original and probably you will not hear the problems at boundary. If you transcode (encode, decode and again decode) then the chance that encoder wrongly assumed what you will not be able hear is doubled. And transcoding usually means that first you encode to high bitrate (320 kbps) and later decode and encode to lower bitrate (128 kbps). In scenario when transcoding from 320 kbps to 128 kbps, encoder is already lacking lot of information from original. That missing information you will not be able to hear, but that information might be crucial for encoder to make correct assumption how to recompress to 128 kbps in a such way that you can not distinguish from original.
But IMO if you transcode from 320 kbps MP3 to 128 kbps AAC there is little chance that you will notice difference (unless you are younger than 25 and among those 1 % with great hearing abilities). And you probably will transcode to 128 kbps in order to play music on some portable device while you are out and you will not pay attention to some very small details in music that have changed.

IMO it is theoretically possible to create encoder which would not suffer much from transcoding quality reduction. But probably nobody wants to invest time and money into it since there is not much from it and transcoding from 320 kbps to 128 kbps probably will be transparent to 90% of population that listens music.

And there is also scalable lossless which will be soon publicly available. With it you can store audio to hard drive without any loss and then simply and fast reduce it's size to any desired bitrate using lossy encoding. With it you can also store audio in a lossy format and later transcode it with less loss than when using traditional transcoding.

For DVD player to be able to play CD with mp3 it must have mp3 decoder built in. Old DVD players don't have that. It also might be possible that some new DVD player won't be able to play mp3. Then maybe there are DVD player that also might not play Audio CD.

Transcoding means decoding mp3 then encoding it again to mp3. When you decode mp3 to PCM (.wav) then there is no data loss (lets put aside quantization and dithering because that has very little impact). When you encode PCM to mp3 then there is always data loss.
Encoder "removes" data that it expect you will not be able to hear. But it is not perfect and it may reach boundary between "you will not be able to hear difference" and "you will be able to hear difference". You encode once from original and probably you will not hear the problems at boundary. If you transcode (encode, decode and again decode) then the chance that encoder wrongly assumed what you will not be able hear is doubled. And transcoding usually means that first you encode to high bitrate (320 kbps) and later decode and encode to lower bitrate (128 kbps). In scenario when transcoding from 320 kbps to 128 kbps, encoder is already lacking lot of information from original. That missing information you will not be able to hear, but that information might be crucial for encoder to make correct assumption how to recompress to 128 kbps in a such way that you can not distinguish from original.
But IMO if you transcode from 320 kbps MP3 to 128 kbps AAC there is little chance that you will notice difference (unless you are younger than 25 and among those 1 % with great hearing abilities). And you probably will transcode to 128 kbps in order to play music on some portable device while you are out and you will not pay attention to some very small details in music that have changed.

IMO it is theoretically possible to create encoder which would not suffer much from transcoding quality reduction. But probably nobody wants to invest time and money into it since there is not much from it and transcoding from 320 kbps to 128 kbps probably will be transparent to 90% of population that listens music.

And there is also scalable lossless which will be soon publicly available. With it you can store audio to hard drive without any loss and then simply and fast reduce it's size to any desired bitrate using lossy encoding. With it you can also store audio in a lossy format and later transcode it with less loss than when using traditional transcoding.

Thanks for taking the time to reply.

So if I understand correctly, as long as I leave my 384 kbps lame downloads as is I should be okay? Also, it won't matter which format (mp3 or audioCD) I burn to disc with as long as the DVD player that I intend to play it on will read it?

Is there a standard bitrate, that the music industry uses, to make regular store bought CDs?

Thanks for all your help,
Yeto

Regular Audio CDs use PCM that is 16 bit stereo at 44100 Hz. You can find more about this at wikipedia.

Is there a standard bitrate, that the music industry uses, to make regular store bought CDs?

You should use "Audio CD" in Nero if you want to burn a CD that will play in a normal CD player.

I am not sure that I fully understand the transcode process. Would you have time to explain some detail on the transcode process?

Transcoding just means changing from one format to another. So, suppose you have all your files as 320 mp3s now, but at some future time, you need to copy them to another lossy format (perhaps to make even smaller files). If you start from something that has already lost data , like mp3, and you put it through another process that discards data, you can reach the point where the effect is actually audible, more quickly than if you had started with a lossless file, with all its original data intact.


I have also noticed that some older DVD players will not read the CDs that I have burned using the mp3 format. When I burn my CDs, using Nero, should I use Audio CD instead of mp3?


OLder DVD and CD players cannot decode mp3 files. You have to use 'audio CD'

For DVD player to be able to play CD with mp3 it must have mp3 decoder built in. Old DVD players don't have that. It also might be possible that some new DVD player won't be able to play mp3. Then maybe there are DVD player that also might not play Audio CD.

Audio CD play is part of DVD players spec, I believe. So you won't see a DVD player that can't play CDs too.


Transcoding means decoding mp3 then encoding it again to mp3.

That's one of many possible transcodes.

I thought some of the HD DVD/Blu-ray players could play DVD but not CD. Or am I imagining things?

Ah, high-def DVD players are a new breed, with its own spec. I think they are required to be back compatible with DVD discs, but I don't know that CD play is also required. It would be incredibly foolish if not.

It seems that each new DVD player or recorder model be it HD Blu-ray or SD.
Has some sort of restriction or limitation. Some have this type of audio or video connector, but not another (HDMI/component, 5.1 / digital audio). Some are limited in video output formats 720P, 1080i, 1080p. Still others can only play some types of discs. (DVD-r, DVD+r, SACD, DVD-A and others)

For example AAC, OGG and WMA can produce excellent quality at lower bitrates than mp3.

While it is true that some of those codecs will sound better than MP3 for a given low bit-rate, extensive listening tests show that they're all about the same at higher bit rates. There was an article in the JAES about this a while back, but it didn't even extend up to 200 kbps; it focused on the relative performance of low bit-rate coding.

Also, everybody suggests Lame for mp3, and while that's "free" and all, it's not the best mp3 encoder out there. I believe Fraunhofer themselves has a pretty good encoder of their own, that can create virtually lossless encodings at 192 kbps, and do pretty darn well at 128; this is probably a better encoder than Lame.

Anyway, spending 320 kbps on passages that are mostly silence is useless. I suggest using VBR, which will give you the bit rate where it's needed, but still encode silence and quieter passages more efficiently. VBR 0 is a good "lossless" setting if you're using Lame.


Last: for anyone putting all your music on a magnetic platter with an expected lifespan of 3 years: for God's sake, make sure you mirror that drive! RAID-1 has saved my bacon several times in this decade. When one drive goes, you have the data on the other, and can buy another drive to replace the first. The next step up in data protection should probably be an off-site back-up, although putting all your CDs in storage somewhere probably counts (modulo the long effort of re-ripping everything, should your computer burn up or something).

While it is true that some of those codecs will sound better than MP3 for a given low bit-rate, extensive listening tests show that they're all about the same at higher bit rates. There was an article in the JAES about this a while back, but it didn't even extend up to 200 kbps; it focused on the relative performance of low bit-rate coding.


Can you please tell me the name of that article? I am researching objective measurement and that article would be helpful. Though only public listening tests with available samples are of the real use.

You can see the test at soundexpert.info (sorry I can not post links)
for comparison of encoders at 192 kbps.
I wouldn't use average grade there as a good measure (grading is different than standard SDG). But what has meaning in that test is that lowest grade for Nero AAC at 192 kbps is 5, which means that nobody could hear a difference between original and compressed at any sample.
And Nero AAC is also available for free. And of course AAC is also capable of VBR so that you don't vast bits when it is not needed.

I don't remember the title, and my journals are all in storage. It was probably two years ago, and it was in the issue that focused on modern audio broadcast/distribution IAICR. The title probably had "low bit rate" in it. It was based on a standard A/B listening test with defined signals, for sure (i e, real research, not just hand-waving).

IMHO many CD tracks can be considered lossy. By lossy I simply mean they lose information. The information is lost at whatever point analog was converted to digital. To convert analog to digital, you gave to quantize the source. Admittedly you can reconstruct the analog signal to a very high degree, but it still loses information.

People seem to get pretty "religious" about this whole lossy vs. lossless stuff. At the end of the day, if Joe can't hear the difference, Joe should not care that Frank does hear it.

I encourage people to try it for themselves and come to their own conclusion.

p.s. I could not tell the difference between a CD track and 128 AAC; does that mean there's no difference? No. It means I can't hear it, so 128 AAC is acceptable to me.

While it is true that some of those codecs will sound better than MP3 for a given low bit-rate, extensive listening tests show that they're all about the same at higher bit rates. There was an article in the JAES about this a while back, but it didn't even extend up to 200 kbps; it focused on the relative performance of low bit-rate coding.

There's also extensive testing and comparison of codecs at www.hydrogenaudio.org (which support what you say here too)

Also, everybody suggests Lame for mp3, and while that's "free" and all, it's not the best mp3 encoder out there. I believe Fraunhofer themselves has a pretty good encoder of their own, that can create virtually lossless encodings at 192 kbps, and do pretty darn well at 128; this is probably a better encoder than Lame.


Well, LAME , which has been under constant development for the past half-decade at least, is pretty darn good. (Which doesn't mean the Fraunhofer is worse, though the FhG codec that came with Audition 1.0 did no better, and sometimes worse, than LAME in a 128 kbps VBR multicodec listening test conducted at hydrogenaudio in 2004

http://www.rjamorim.com/test/mp3-128/results.html

it was however used in VBR mode,which might not have been optimal for this particular implementaion of FhG)

At 192 (average) VBR, I certainly haven't been able to tell any LAME encoded mp3s from source yet , in ABX comparisons, using LAME versions from the last two years. Others might, especially if they used known 'difficult to encode' samples, which I haven't tested myself with. But I would not expect LAME to be performing significantly worse at this point than the Fraunhofer codec.



Last: for anyone putting all your music on a magnetic platter with an expected lifespan of 3 years: for God's sake, make sure you mirror that drive! RAID-1 has saved my bacon several times in this decade. When one drive goes, you have the data on the other, and can buy another drive to replace the first. The next step up in data protection should probably be an off-site back-up, although putting all your CDs in storage somewhere probably counts (modulo the long effort of re-ripping everything, should your computer burn up or something).


Hear, hear.

IMHO many CD tracks can be considered lossy. By lossy I simply mean they lose information. The information is lost at whatever point analog was converted to digital.

Leaving aside that one has to be careful defining 'information'....

The most information lost -- spatial information primarily -- by far, happens at the recording microphone. Another huge loss point is at the speaker...by far the most 'distorting' component in most home playback systems (leaving aside the room itself).

To convert analog to digital, you gave to quantize the source. Admittedly you can reconstruct the analog signal to a very high degree, but it still loses information.

Not nearly as much as is lost when recording to magnetic tape (or transferring to vinyl). Really, theoretically digital captures ALL the amplitude and frequency information within the Nyquist bandwidth (a claim that cannot be made by any analog format), and in practice it can do so to a degree in the high 90% (losses are due to physical limits on technology).

But ALL recording is lossy by a strict definition. So is hearing. Ears are filters. They don't capture and transduce 'all' possible sounds.

All good points :)

IMHO many CD tracks can be considered lossy. By lossy I simply mean they lose information. The information is lost at whatever point analog was converted to digital. To convert analog to digital, you gave to quantize the source. Admittedly you can reconstruct the analog signal to a very high degree, but it still loses information.

People seem to get pretty "religious" about this whole lossy vs. lossless stuff. At the end of the day, if Joe can't hear the difference, Joe should not care that Frank does hear it.

There are really two separate issues that are often combined when discussing topics like this. Those are, the storage of music, and the use/playback of music.

In this age of TB size HDDs, there's absolutely no reason to rip music to a lossy format for storage. For example ripping music to your HDD on your computer, which is often used in various other ways (via "extenders", transfered to PMPs, etc). Ripping to a PC is basically archiving it, and there are very real advantages to using lossless encoding for archives (avoiding generational losses being the big one).

The other issues is playback, eg streaming to a remote extender, across the internet, via a portable player. In situations like this, where bandwidth and/or storage are often limited, there are very real advantages to using lossy compression, those being things like streaming over the internet (bandwidth limited), copying to a portable player (capacity limited), etc. Further these are often applications where quality is not the primary concern.

There's nothing wrong with lossy compression in general, it has it's applications. However what still baffles me, is why some seem to advocate the use of lossy compression in situations where it has essentially no benefit over lossless compression.

Not everyone has the money, time, energy or whatever to upgrade their HD's to a larger size. I spent $1800 on a PC in May, and only have 70GB* of HD space total (not counting an external drive I use from time to time.) My macbook only has 60GB, I think. And my iPod used to only have 10GB (I just got the newer crappier video one, but it does have more storage.)

I guess I seem to no reason to archive music I own on CD.

I guess if I was building a dedicated HTPC I might rip to something lossless...maybe. If you were also storing DVDs on it, you could eat up space pretty quick, even was a TB NAS server or something like that.

I have been told HD space is cheap for years...and maybe it's getting truer now. But try to get a few extra GB for use at work - every company I have worked for requires you to prove you need it. Off topic maybe, but I personally would not want to losslessly rip music on the off chance I MIGHT be able to tell the difference. I still contend that people should find out for themselves whether they need bigger files to achieve the best (audible to them) sound quality.

* Because dual WD Raptors are not cheap, and are not large drives

Not everyone has the money, time, energy or whatever to upgrade their HD's to a larger size. I spent $1800 on a PC in May, and only have 70GB* of HD space total (not counting an external drive I use from time to time.) My macbook only has 60GB, I think. And my iPod used to only have 10GB (I just got the newer crappier video one, but it does have more storage.)

You can get 500GB of storage for ~$140, maybe a bit more if you want/need an external USB drive. That's enough for >1500 CDs losslessly compressed.

Heck, within the next month or two, you'll be able to buy 1TB drives for $400 (http://www.pcworld.com/article/id,128400-page,1/article.html)

I guess I seem to no reason to archive music I own on CD.

Guess you haven't experience the joy/convenience of having all your music at your fingertips, instant, random access to everything. I don't think I could ever go back to listenning to "CDs".

I guess if I was building a dedicated HTPC I might rip to something lossless...maybe. If you were also storing DVDs on it, you could eat up space pretty quick, even was a TB NAS server or something like that.

Up to 3.5TB here :D But that's kind of beside the point because DVDs are an order of magnitude greater in storage requirements.

I have been told HD space is cheap for years...and maybe it's getting truer now.

~$0.14/CD doesn't seem bad to me.

But try to get a few extra GB for use at work - every company I have worked for requires you to prove you need it.

I hear ya there, I often joke with my co-workers that I have more HDD space than the company :D

Off topic maybe, but I personally would not want to losslessly rip music on the off chance I MIGHT be able to tell the difference.

I tried to frame my argument specifically to avoid that argument. Like I said, there are real, practical advantages to using lossless for the PC side of music. For example being able to convert to any other format without suffering generational losses.

Say you switch from an iPod to a Zune, now you have to re-rip all your music to avoid suffering generational losses, with lossless compression, you can transcode on the fly to smaller formats for your portable players. lossless to AAC today for the iPod, lossless to WMA tomorrow for the Zune. Something new comes out next year, new format, no prob. Your favorite audio format changes next month, you can batch up your collection and convert it without loss.

You can't claim that with any lossy format. I don't know now many times, but even with 320k MP3s, a couple conversions to other formats and you're at clearly audible issues.

I still contend that people should find out for themselves whether they need bigger files to achieve the best (audible to them) sound quality.

And I don't disagree that for applications where size/bandwidth are concerns that people should listen and figure out what they need. I did that for my portable player (only 1GB), and I've done it for streaming music to work (I can't quite stream lossless).

But if you've got CDs (lossless), I just don't see why you'd want to use any lossy format for the archives on your PC.

even with 320k MP3s, a couple conversions to other formats and you're at clearly audible issues

But you won't go more than one generation after that. The theory is: 320k MP3 is the archival format. Any other format is re-generated from this archive. I do not think you will hear any artifacts in any program material if you transcode from 320k MP3 to some other high quality codec, at least nothing you wouldn't hear in the 320k MP3 (and I highly doubt you'd hear anything different in ABX tests with lossless there).