Excessive Copy Protection?, Mp3 related Options

[Dub45]

This may be old news to some of you but its the first time that I have come across it.

I have just bought a copy of 'Round about Roma' by Stefano Di Battista on Blue Note.
This disc does not seem to have been released in the US yet.

I put it into my pc to have a listen while I browsed and suddenly a player that I had never seen before appeared! Anyways the cd provides its own player but apparently will not allow me to use my usual player to rip it as I usually do for use on my ipod.

When I opened the cd through explorer it does not display the audio tracks separately.

This is infuriating (I dont share MP3s on the net and I don't download them) but I can't listen to an album that I have bought as I would wish!

I presume the introduction of the player is to get over the complaints of cds not being playable on pcs.

Any other incidents of such cds? What are other people's views?

Incidentally I tried to copy it on a 'normal hi-fi type copier and it wont copy either!

[tigre]

@Pio2001:
Thanks a lot for your detailed information. I think now I understand much better what happens when drives try to rip copy protected CDs with corrupt error protection info. I'll do the test you linked to soon (atm no drive with 1.8 gig free available).

Unfortunately I don't have a CD player with digital out and only VIA on board sound, so I had to find another way to get an "as close to secure as possible" rip. I've got some ideas how to use different drives', programs' and modes' (EAC) rips to improve security, but I wrote my last program 8 years ago (does anyone here remember DOS ), and now I'm too lazy to find out how it works again. I don't know where to get matlab (AFAIK it's quite expensive, there's no free trial version) neither how it works. So I (ab)used Cool Edit Pro.

First I had a closer look on differences between the rips (comparing number, position, value of different samples, comparing spectral views) and got the impression that my plex121032A did the best job. Lowering extraction speed to 7.9 avoided the "exception (weird)" error (see previous post). All extraction modes I tried (secure, paranoid, burst, even manually offset corrected CDmage extraction from CloneCD image) gave identical results. The ripped songs sounded fine to me without knowing about error/suspiscious sample positions.
The results of my LG DRD 8120B DVD also sounded decent, my no-name 48x CD-ROM drive produced some audible clicks (both drives EAC secure mode, C2 off, copy/test CRC match).
Having a look at a lot of positions where drives give different samples I found that for most samples where plex and LG-DVD give different values plex and no-name give the same. On the other hand on samples where plex and no-name give different values, plex and LG-DVD are identical. This makes me guess that plex is most correct. BTW there are very few samples where LG-DVD and no-name give the same value but plex gives a different.
My goal now was to get a mixed file of each CD track containing as much correct (= identical values by >= 2 Drives) samples as possible. Here's my way (ab)using Cool Edit Pro:

ripped files
p plextor secure noC2 7,9speed test/copy CRC OK "gold standard"
q 2nd drive
r 3rd drive ...
n new file, containing of best possible samples of p,q,r

algorithm
if p=q then n=p [# I]
else if p=r then n=p [# II]
else if q=r then n=q [# III]
else n=p [# IV; no values match, so trust "gold standard"]

simplified:
if q=r then n=q
else n=p

To "program" this in Cool edit I had to do following steps (note: dithering must be disabled):

1. Difference

a=q-r b=p-q

problem: "clipping": a and b can have every value between -65534 and 65534, while in 16 bit audio there are only values between -32767 and 32767 possible. Solution: amplify to 50% before substracting, add twice if needed, correction of introduced errors see 4.

a=q-r is done in CEP by a="mixpaste_overlap(invert®;q)" (for a clipping is not an issue, because it'll be "booleanized", see 2.)
b=0,5*p-0,5*q : b="mixpaste_overlap(invert(amplify(q;50%));amplify(q;50%))"

2. Goal: 0 sample stay 0 sample, convert everything else to 32767 (boolean 1)
function: "booleanize"

a_b=booleanize(a) : a_b="mixpaste_modulate(amplify(a;+91dB);amplify(a;+91dB))"

Mixpaste_modulate problem: CEP mixpaste_modulate (edit>mixpaste check modulate in dialog) is should multiplicate 2 values and divide by 32767. Checking the results I found out that it lowers the value by 1. So "Booleanize" produces 0 on zero samples, on all other samples sometimes 32766, instead of 32767.
To avoid this and get only 0 or 32767 samples I used a_b_new="amplify(a_b;+1dB)"

3. Goal: if for a sample a_b_new=32767, then sample value shall be n=p=2*b+q ; if a_b_new=0, sample value n=q
n="mixpaste_overlap(mixpaste_overlap(q;mixpaste_modulate(a_b_new;b));mixpaste_modulate(a_b_new;b))"

Here the "mixpaste_modulate problem" occured again. It produced differences of the value 1 (eg. 23765 instead of 23766).

4. Correction of "clipping" and "mixpaste_modulate" inaccuracies:

c=p-n : c="mixpaste_overlap(invert(n);p)" (clipping no issue, because further needed samples have small values)

c needs correction to make the following mixpaste_overlap operation exact. 0 samples have to stay 0, negative have to be decreased by 1, positive increased by 1 (see "mixpaste_modulation problem")

c_corrected="mixpaste_overlap(c;amplify(amplify(c;+91db);-90,3db))"

n_corrected="mixpaste_overlap(mixpaste_modulate(a_b_new;c_corrected);n)"

That's it. In a 4:30 minutes song there are about 200 "insecure" samples left (before there were thousands, comparing any 2 of the drives), at the questionable positions everything sounds fine, I can't find anything strange looking in spectral view either (AFAIR in spectral view you can see "bad" things you can only hear AFTER seeing ). Probably I could even improve the result by using more than 3 drives/rips. But I'm satisfied with what I got (and have spent enough time on this). Ok, could be interesting how long it took: After figuring out and "debugging" (too long) it took about 3 minutes to perform this manipulation on a 4:30 song. It should be quite easy to be coded in a regular programming language. Someone interested? Or does something like this exist already?

Thanks for your patience reading this