Bug report: paranoia mode reads incorrect track, -Z reads correct track

[email protected] Wed, 10 Aug 2016 18:43:44 -0400
Newsgroups gmane.comp.audio.cd-paranoia.devel
Message-ID <[email protected]>
--===============6520827782863924262==
Content-Type: multipart/signed; micalg=pgp-sha256;
	protocol="application/pgp-signature"; boundary="y0ulUmNC+osPPQO6"
Content-Disposition: inline


--y0ulUmNC+osPPQO6
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

Greetings,

While working on some software based on cdparanoia-III-10.2[0], I have=20
discovered what I think is a bug in cdparanoia:

Writing a wav[1] with md5sum 3b265ba06de8cd27f200e79447c58035 to a disc=20
and then ripping in default full-paranoia mode will result in a track=20
with incorrect md5sum 063fa2d2df295f7dbfad8d75ce0aeafa -- however,=20
ripping with paranoia disabled (-Z) will yield the correct track rip.[2]

This is confirmed to happen on at least the two CD drives I have=20
available. Programs like cdrdao and EAC rip the track as it was written=20
without complaining.

Ripping the entire track yields this bug, but if you try to directly=20
rip a range containing the problematic samples, depending on which=20
parts of the track you include the bug may or may not occur:

ranges between  5[00:00:00.11]-5[00:04:15:05] rip correctly in both modes
ranges starting 5[00:00:00.00]-5[00:00:00.10]
AND ending      5[00:04:15:06]-5[00:04:15.69] rip incorrectly in paranoid m=
ode
ranges merely starting or ending in the 'problem' areas but not in both=20
are fine.

The differences in the tracks are reported as corrected 'jitter' (if to=20
exactly 04:15.07) or 'stage 1' (if >.07), and silently different if=20
just to .06.

The track is #5 of Julie Roberts' eponymous LP.[3]

I have a diff of a hexdump -C of the differences.[4]

Other testers have reported a different incorrect track (which I have=20
not looked at), but correct results ripping in -Z mode.

The CD drive for the above result is the LG Electronics - DVDRAM GT33N=20
LT20. (My read offset is +103 sample-pairs.)

Please let me know if I can provide any more information.

[0] https://github.com/JoeLametta/whipper
[1] http://hylaea.logik.li/05Z.flac
[2] http://hylaea.logik.li/05-paranoid.flac
[3] https://musicbrainz.org/release/0af88456-7c0b-4f9d-9610-cd62520d2fef
[4] https://gist.github.com/RecursiveForest/0ec337c448122d33032f8fb864e3f941
--=20
Samantha Baldwin - logik.li

--y0ulUmNC+osPPQO6
Content-Type: application/pgp-signature; name="signature.asc"

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
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=fKpP
-----END PGP SIGNATURE-----

--y0ulUmNC+osPPQO6--

--===============6520827782863924262==
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: base64
Content-Disposition: inline

X19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX18KUGFyYW5vaWEt
ZGV2IG1haWxpbmcgbGlzdApQYXJhbm9pYS1kZXZAeGlwaC5vcmcKaHR0cDovL2xpc3RzLnhpcGgu
b3JnL21haWxtYW4vbGlzdGluZm8vcGFyYW5vaWEtZGV2Cg==

--===============6520827782863924262==--