Back to:
Summary - technical details for high quality extractions
MPC Tutorial
Lossless Tutorial
rev 1.00
Why ?
No usage of c2 information ?
Test & Copy ?
is the lossless Shorten (shn) format not recommended ?
single tracks, no image ?
No Normalizing ?
Replaygain ?
EAC, but no Plextools, no CDex, no Linux/cdparanoia ?
not extracting from CD-R ?
Why No usage of C2 ?
Because some drives are not able to read/interpret/use the C2 information correctly. Despite it was recently (in 2004) found out, that probably more drives than assumed in past (because a special developer tool was flawed), are able to use c2 correctly, you can not be sure.
Additionally, there are non-compliant red book standard CDs (no CDDA), which contain wrong or bad mastered C2 info. So it is better, not to use C2.
The only advantage of c2 usage would be speed, but as this guide should offer you a secure and simple way, (simplicity produces security, too !), this guide has previously not recommended any C2 usage.
But as written above, the idea, that c2 error detection can today be considered safe, as a developer tool was faulty in past and gave wrong falses, the reasons against c2 usage for error detection are no longer valid. So, it might be, that I don't recommend anymore in near future, that no c2 is the only way for secure extractions. Today I would say, that no c2 or the c2 setting in EAC's drive features aren't so important anymore.
But: If you should consider to use c2, then carry out test & copy in any case, to get matching checksums !
Without matching checksums you don't have safety for a perfect extraction. This is valid for no c2 setting in EAC, too. The doubled extraction by Test & Copy will show you with very high statistical safety, whether your extraction was good or not.
Why Test & Copy ?
Test & Copy mean 2 separate independent extractions of each track by EAC. EAC computes the checksum crc for each track and compares the crcs. If the crcs do not match (match = they are the same), then the read data of that track was not consistent. The reason will be a scratch on the surface. Errors contained in a track can be characterized by 2 different kinds: 1. repeatable = consistent errors, 2. non-repeatable = inconsistent errors.
The consistent errors cannot be detected by EAC's no-c2 way of extracting, which sounds dangerous. But: The inconsistent errors are detected by EAC's no-c2 way of extracting. So, why is this described (in this guide) way of extraction so safe, that all errors are found ?
Because, if there is a scratch, which changes original data on that CD-track, there will be always not only consistent errors, but also inconsistent errors neighboured. So, all suspicious positions or tracks are shown in EAC logfile and you are warned.
EAC's secure mode setup by this guide is so safe, that you can be 100.00% sure, that the extraction was perfect, if no errors appeared and the crcs are matching.
Why is the lossless Shorten (shn) format not recommended ?
Shorten (SHN) cannot be recommended as Lossless format anymore, so it isn't supported by this Tutorial. SHN is somewhat outdated and old, so technically all other mentioned Lossless formats have advantages compared to SHN. Advantages like achieved compression ratio (smaller filesizes !), speed for en- and decoding etc.
Even the popular FLAC format seems to be outdated (regarding compression sizes & speed !) to my opinion these days, since there is WavPack 4.1.
Why extraction to single tracks, no big single image file ?
There are several reasons: Convenience, simplicity, success:
If you browse through your directories, you see your albums nicely as 1 folder, sorted by artist, year, album, if you follow the recommendations for naming scheme.
If you are interested in single songs, tracks, you can browse into the album directory, and you have a quick overview, what is in your album, e.g. song names, pictures from booklet scans etc. If you would create single files per album, e.g. Matroska container mka format, or a single file in rar or zip format, you lose this advantage.
Another technical reason, maybe more important: If a checksum crc mismatch happens during test & copy by EAC's extraction for a single track, as there is one bad scratch on the CD surface, you need to re-extract again only this specific 1 track. If you get the mismatch during extraction of the complete CD image, you have to extract again the complete CD. And this would be a big waste of time. Besides, for this reason EAC does not offer Test & Copy for extractions of images. So, an image extraction cannot be considered as High Quality, like I understand high quality, because of missing Test & Copy. See the chapter about importance of Test & Copy on this site above.
No Normalizing ?
If EAC or any other program would carry out 'normalizing' (e.g. from 100% to 98%), it would introduce rounding errors, which would be a kind of noise. And as result, the backup of the music wouldn't be perfect anymore, destroyed forever.
There is a way better kind-of-lossless solution for adjusting volume gain, perceived loudness. ReplayGain !
If you don't want the change ReplayGain applies during playback of the music, you can switch it off simply, e.g. in Foobar settings or Winamp. Then you have always the choice, if you want to benefit from the advantages of Replaygain, or if you just want the raw unaltered music. Read here some background info to Replaygain.
Replaygain ?
As written above in chapter 'No Normalizing', replaygaining offers only advantages:
E.g. clip protection for MPC, and for all formats, MPC & Lossless: selectable modes during decoding or playback: no replaygain at all; album gain (every album is played by same averaged volume of 89 dB, this keeps relative volume differences between songs of same album); or track gain (all songs are played/decoded with same averaged volume per song of 89 dB.). Read here some information.
EAC, but no Plextools, no CDex, no Linux/cdparanoia ?
EAC is still today the top-of-the-art program for secure extractions, unfortunately not simply running and working on Linux, though I have heard, that even EAC is possible to work on Linux.
Plextools work only safe with Plextor drives, and they are limited to windows only. They don't have a built-in test & copy mode like EAC. Though, a Plextool extraction with a Plextor drive can be considered as secure high quality, if certain settings, conditions were carried out or given. The knowledge of this guide grows every week, and so it might be extended by a Plextools tutorial for secure high quality extractions.
Same reasons for Linux based extraction programs. CDparanoia seems to offer a possibility and a way for secure extractions. This guide might be extended with a tutorial for this.
CDex: As there is EAC as freeware available, and unbeaten in reliability, safety, there is no room left for CDex.
Regarding Linux, it is technical no problem, to install a small windows partition, eg. simple but stable windows2000sp3 or sp4, on same HD like you have Linux, eg. different partition. This has the advantage, that you can run EAC simply, besides some other windows-only programs, which you might be delighted about to run them easily in future occasions. But I understand fully, why Linux users ban windows completely from their systems. You have my full support !
not extracting from CD-R ?
Because you cannot be sure, that the CD-R copy was made perfectly. So, it does not make any sense, to extract a CD-R copy by this High Quality guide. Consider, what it needs for a perfect CD-R copy: 1. a perfect extraction like carried out by this tutorial, 2. a perfect burning of the CD-R, 3. the CD-R needs to be in young and perfect condition, be careful, CD-Rs consist of organic materials, which can suffer by aging, especially, if it is a cheap brand, no-name or such. So my recommendations, to add par2 data to your archive backups on DVD+-R or CD-R. EAC and no other extraction program would be able to detect read errors of the 1st generation copy on the CD-R.
So, simple and short: No copy from a copy !
That is, as CD-Audio is completely different from data-CD. A 700 MB data-CD contains additional 100 MB hidden redundant data, to repair missing/corrupted data by scratches. When the CDDA, CD-Audio, was created, they thought, that Beethoven's 9th symphony has to fit to 1 disc. As that is a long symphony, they made this by sacrifizing the additional hidden background data. So, if you look carefully to Nero as burning program eg., you will find out, that 800 MB wave (ca. max. 80 minutes music) files fit to a 700 MB data-CD-R disc. Well, so we have the problem to make secure archive backups of CD-Audio.
Back to:
Summary - technical details for high quality extractions
MPC Tutorial
Lossless Tutorial
