I'm in school right now and the other day the teacher went on a half hour long rant that copying/ripping CDs was ALWAYS lossly and you should not do it for that reason. Direct CD copying = lossy; WMA-Lossless = Lossy; ALAC = Lossy; FLAC = Lossy; etc. He didn't know most of those, but his stance was quite firm that any ripping/copying was lossy...

Now, I can almost understand the idea that applying compression loses some data (despite all the evidence I've heard to the contrary, I could be talked into that...); but from my understanding WAV is the codec used on the CD itself and a direct cd copy is damn near bit to bit copying....

I really don't know, it kinda threw me for a loop. Can anybody clarify at all?

I think the point that he was really trying to make was that PCM is a lossy process.

FYI, PCM is the "codec" used for cd's. Wav is the container that we usually use it in.

The best you can hope for a bit for bit copy from cd, which is very possible.


edit: added links

ask him if you shouldn't play CDs either "for that reason".

I would've thought that as well; because he did make it quite clear that encoding onto a cd was a lossy process as well (which is readily known); but he was rather adament about the idea of copying CDs, not stamping them but burning them, as being lossy no matter what. Made me wonder...
*shrugs*

He is misinformed.

Some audio lunatics on the net might really have daytime jobs like teachers!

I picture the class -

Lesson 1: Lossless is lossy. Everthing is lossy.

Lesson 2: Joint stereo is evil

Lesson 3: You can always tell 320k from the original

Lesson 4: Use this to get acceptable mp3's - b320 ms X0 -Y -k -lowpass22000

So if I make a copy of a CD-ROM, the data on that isn't the same as the original? I don't think so...would make the idea kinda useless, don't you think.

Maybe he needs to be pointed to this (link from FAQ - "Why are audiophile not always trusted when they say that a copy doesn't sound as good as an original CD ?")?

What kind of course is this, anyway?

Audio screwing up.

Why is "encoding" onto a cd lossy? If you burn bits 100010101 onto a cd, that's exactly what will be on the cd after the burn (if no errors occured): 100010101.

Ask your teacher if he ever copies a CD-ROM. Or if saving a Word document on your hard drive is also lossy

I do hope there's been a miscommunication somewhere...

he/she is clearly one of those teachers that tells alot about stuff they know jack shit about.

i had one of these teachers trying to tell me that every kind of compression removed quality on the pictures.

even when i stod and drew a simpel RLE encoding for him, he refused to think otherwise.

The reasong i always hate teacher, to stupid and to proud.

Your teacher certainly has a point. There is a difference between data and audio cd's. When you rip an audio cd there is no proof that the data you rip is identical to the data on the original cd no matter what tool you use.
Lossless compression is a completely different issue because you can verify that the compressed data is identical to the uncompressed.

Just burn your own CD. When you rip it, you can check that it is the same as the original, since you've yourself burned it from this original.

You are wrong. The CD has various level of error detection correction C1/C2/CU. I am mot very seasoned in the details, but I know that if I do not have C2 error then I can sleep peacfully that I could rip the CD without error. I even remember thet C2 error alone does not mean error. Anyway my knowledge on the details are patchy, but your statement is clearly wrong.

Triza

You can't actually ascertain that it's bit identical unless you prove that the CD drive reads it perfectly... if you burn it, having read it wrongly, then any subsquent checks could also read it wrongly.

I think that's getting just a little too paranoid though

ask him what else in life isn't perfect

I think he is trying to make the point that nothing is a true copy, it's just no physicaly/numericaly/whatever possible.

Just to add a little grain of salt to the discussion, I would say that ripping a CD is a lossy process. An audio CD that is.

Audio CDs were not meant to be read exactly (which is bit-exactly) but were meant to be read real time. While there are a lot of error correction mechanisms written onto CDs, very few hardware is equipped to read it and even less software.

As someone that has tried (and finally succeeded on the last attempt) to rip his entire collection many times, I can tell you that I succeeded in being confident in my rips with one config: One old TEAC burner that detects C2 error codes and EAC.

And even then, I have had CDs that reported no errors but had clicks or audible distortions.

So your professor was right: Ripping is in no way a lossless process.

Now once you have a WAV on your HDD, applying any lossless codec will effectively be lossless - at least in terms of quality.

Well, either something went wrong in your ripping process, or these distorsions were present in the original recording as well.

If your CD drive is buggy, then it can perfectly output the same *wrong* audio sequence twice. Of course, a lossless copy implies that your hardware actually works.

If your hardware doesn't work, then data bits can be corrupted at any time (when in RAM, on HDD, while passing through the IDE cable.. anywhere).

To be perfectly safe, you should generate a hash (MD5 etc) as soon as your audio is extracted, then check it periodically.

CD players certainly fully use the error correction mechanisms in Red Book through actual data correction and interpolation. The difference between playing an audio CD in a player versus ripping an audio CD with EAC is that EAC can make up for the shortcomings in non-correctable error handling (interpolation) in CD players by applying a more rigourous non real time algorithm when errors do occur.

To say that ripping is lossy is no more than an admission that the user is either using the wrong tool to extract the audio data, or is using a heavily damaged disc in the first place. A CD in half decent shape with a good quality ripping process like EAC Secure mode and AccurateRip will certainly be lossless given proper due diligence. The same CD ripped with WMP may not be a lossless process.

It all depends on the tool. The potential for a known bit-perfect copy is always there.

It may just be quibbling over the definition of "lossy" and "lossless", but I think that ripping a CD is lossless no matter what. Errors during the rip process are accidents, not some fundamental flaw that can't be avoided. To me lossy means that you are guaranteed to lose information, which even in hypothetical mathematics-land you can't get back. While lossless means that in math-land where things are always perfect you get the identical copy. Errors from the ripping process mostly come from physical flaws like scratches and such.

I want to know what course this loon is teaching. I hope it's compsci where his ignorance of how the hardware works can be excused, not something like e engineering. And sven Bent, where did you find that teacher who didn't believe in lossless compression of pictures? Did he somehow think that the zip math somehow worked on any form of data except images? As the texas aggie said, "how do it know?"

All you have to do is an MD5 (or similar) checksum comparing the orginal file to the copy. When they match, the debate is over.

Just like there are "audiophiles" out there who claim to be able to hear the effects of bird wings flapping 100 miles away, there are "visiophiles" who claim to be able to see differences in various picture formats. Even between, say, PNG and GIF (both of which are, in fact, lossless).

I know what you're saying but it makes it sound like PNG > GIF is lossless...

For 8 bit images, it is.

Well yeah but... ahhh never mind, it's irrelevant anyway

I have run into tons of this sort of thing in the audio engineering realm. I had my album mastered at one of the top three mastering houses in the world and sat in on the session to learn a few things. I was told that by the mastering engineer (who is prolific but will remain nameless) that he can hear the difference between a file on one hard drive and a file on another hard drive. He requested that I present him with my source files on the first drive I recorded them to and that I shouldn't transfer them from my hard drive to his because they will start to sound harsh. Fortunately it was an external firewire drive, so it was possible to satisfy his request. I also brought backup copies encoded in Monkey's Audio... but after hearing that, I didn't even mention them

That said, he did a killer job mastering the cuts and I thoroughly enjoyed the experience.

Perhaps he's hearing the different sounds that the hard drives themselves make?

*enter audiophile - plagued with subjectivism and i'll-just-make-absurd-claims*
*enter xmixahlx - jackass objectivist*

xmixahlx - really? i'd like to see an ABX on that! it would be fascinating

audiophile - i don't need an ABX, i'm a professional!!! i 0wn j00!!!!!

xmixahlx - no? how about i just do an md5 checksum, then?

*end of argument*


sounds like he suffers from 3rd degree audiophilism


later

Hey, I'm the best audiophile! I can hear a difference between a CD!

Ha, that's nothing! I can hear the difference between a mp3 file tagged with ID3 and the same mp3 tagged with APE. I'm currently working on training myself to ABX between Intel and AMD processors.

If you have an old Pentium I (or was it II), the one with the rounding error, that might even be possible!

In theory

Pshaw! I can ABX between two different presses of the same CD! Even if they were pressed one after the other!!!!

I think these guys are FoS - but if someone KNOWS otherwise I'd love to be educated.

That's nothing! I'm pretty sure EVERYONE WITH GOOD EARS™ can ABX a given CD track to itself.

ha! better watch out they don't bust out their $2,000 cables out on you to proove a point . Back to the original topic though lossless is lossless. There are no and's, or's, but's about it. If you are still not convinced I recommend a nice read of linear prediction coding. I don't know about the whole DAE, but technically wouldn't the digital-analog-conversion play a large role if there was any loss? as oppose to just ripping to raw PCM file or something of that nature. I know the paranoia libraries do just that . There is no analog step in between.

Yes that's what I said:



Beat that

daPyr0x, have you spoken with your teacher yet?

What was he supposed to be teaching?

That the Earth is flat ?

It's a mobile audio/etc install course; and no, I haven't spoken to him yet because he's on vacation now. I'm not sure if I'll even bother bringing it up, because he was getting somewhat upset when myself and a few other people were all "um...what about lossless encoding? what about 1 to 1 copies? what about ripping direct to wav?" etc etc....he was rather heated for God knows what reason. If he brings it up again I'll ask him about it.

In my mind, a fully digital rip to FLAC using EAC's most paranoid rip settings would actually (in real life, not neccesarily under unworldly perfect circumstances) give you a better representation of the audio than a consumer level cd player because it not only has the error correction, but it also has the paranoia settings to do things like checking it all 3 times over and such. Of course that doesn't take into account how you get the audio out of the computer, as most common sound cards aren't exactly the cleanest audio output.

His whole point was that when testing a system you should be using stamped cds, not burned. While I agree with that statement, because burned cds aren't always the greatest quality, and different components have different results reading them, thus using a stamped cd would be the most fair testing tool; that wasn't even part of any of the "discussion." Maybe I should pull my shitty sounding car in one day and ask him to tell the difference between a cd I own and the FLAC output from my Karma (which is piped into the deck).....

*shrugs*
At least I know the real backing behind it all.....

Thanks, all.

Do it in a blind test with several trials.
Otherwise he'll tell you that the difference is so obvious that you urgently need to consult a physician if you don't hear it.

if it's a car/audio/install guy, try this: record some virtual silence to wav file, make an audio cd using the wav file. encode the original wav file to an mp3 (APS, API, APX, or something with high quality). burn the mp3 to a data cd. Play both in his car (he probably has a mp3/cd player), ask him which is which.

And I got you all beat, I can ABX with my brain between different sets of ears, and I dont even know what ABX stands for.

ABX stands for "A-B-X", commonly referring to the method employed by a double-blind ABX comparator...a system and/or process used to test and observe a scientific hypothesis with a finite number of criteria in order to generate an objective conclusion.

In the testing of psychoacoustic (or any other type of) audio compression...

A = sample A
B = sample B
X = A or B

The source reference may be A or B, and whichever it is the test sample will be the other. The tester is provided the random test case, X, chooses in each case whether X=A or X=B, enters the choice in the system, then moves to the next randomly generated test case. The final results are tallied after a number of cases are completed which is high enough to provide a meaningful conclusion. (I.e., can the tester REALLY tell the difference between A and B.)

I think QHOBBES was being fecitious.

If you read a CD in multiple different drives on different machines and get bit identical output the chances of you having a bad rip are astronomically low. If your teacher is worried about this I'd like to hear his thoughts on getting hit by an aeroplane.

On the point of trying to convince an "audiophile" of the value of objective analysis, what would the goal EVER be of even trying? Ever? I'd sooner try to convince a retarded person to not be retarded.

When in an A/V store populated with "audio-" and "videophiles", I just smile and nod most of the time, and only correct statements addressed directly to me (i.e., when they're trying to sell me something).

Trying to convince a teacher of a scientific principle would be even less fruitful, as a grade is on the line. Smile, nod, agree when he says that there's no such thing as lossless CD data extraction, and get your top grade. His lack of desire to expand or correct his knowledge will keep him in a lowly teaching profession forever, and you can move on to far exceed his abilities in life.

Everyone should print out this document: The Ten Biggest Lies in Audio and keep it with you all the time, to show to "problem" audiophiles

By the way, I can hear the difference between two different ABX software packages!

Yes. This is 100% complete and pure nonsense.
As to direct CD copying, in addition to preserving all the *correct* bits, it fixes questinable bits (like, doing ripping with EAC), so if the original CD is not 100% perfect, your copy may be in fact better than the original.

You said 'Teacher in school'. What was the subject? What does he teach?

Oi! My parents both used to teach

As I said earlier in this thread, I have ripped 300 CDs so far and I can tell you I have found no combination (hardware/software) that works flawlessly, which is, that would rip without any glitch OR would notify such glitches to me. I tried over 7 CD drives (burners, readers, DVD drives etc...) on 3 different machines.

My best combo was a very rare and old burner (TEAC) that support C2 error correction, and EAC. But even then, I had a few CDs that came out with cracks and pops even though EAC reported no errors. Same CDs in a cd player just played fine and did rip fine in another CDROM drive.

Granted I noticed this for only 2 of the ripped CDs, but then I haven't listened to all of my rips yet. It still remains a pretty low percentage, but still, lossy rips are there. I am not even mentionning the old days where one did not have EAC to help. I can tell you these rips were far from being lossless.

Or course, CD-Rs are much more prone to this than pressed CDs.

Data on an Audio CD has never been meant to be extracted exactly, and as a result, extracting it exactly is an approximate science. And saying that "it's digital so it's lossless" is utter nonsense. The first rule of lossless is: Can you read the data? Unfortunately, with audio CDs the answer is not a big yes... So ripping a CD is a lossless process only in theory.