best way to back up alot of tunes?

[redhooks]

We have a ton of cds both studio albums and bootlegs. I'd like to back them up either on itunes or rip to an external hard drive to either make mp3s or get an ipod. My wife is an audiobook junky who dowloads to mp3 but we are thinking of getting iphones. Any suggestions which is best way to go?

Thanks

[Kyhl]

Volumes have been written on audio forums about this. You might want to search audio forums for more details.

Here is the short version that I know.

The first rule, rip a lossless copy first and store it somewhere and store it somewhere else. Whatever format you choose is fine, as long as it is lossless. I'm not an apple guy so I chose .flac as my storage method. At the time, Apple did not support flac's. They might now, rather, they have their own version of flac that isn't compatable with the rest of the industry.

Use a good ripper that tests the drive and the rip for errors. I, most others rely on Exact Audio Copy, often referred to as EAC. It's a free download that will rip, add metadata, test the data, create logs, and compress to a or many lossy formats.

It also recognizes early discs (late '80's CD were pressed with pre-emphesis that needs to be removed or accounted for) and adjusts for them.

Next store them multiple places. How crazy you go with this is up to you. You definitely do not want to do this process twice, as I and countless others have.

I use a simple setup of an internal drive and a separate backup external drive that is disconnected from the computer. Others go as far as having another backup copy stored at another location. Then there is the whole Raid process that would be best when combined with backups.

So in my case, I have multiple versions, original CD's and LP's, .flac's X2, and mp3. The mp3's are streamed to my phone, the .flacs are used when listening at home.

[redhooks]

Thanks, I appreciate the info. Getting a new laptop so ill be working on backing my old stuff up.

[Whoaru99]

Then there is the whole Raid process that would be best when combined with backups.

Soapbox on...

Yup, RAID can be a great thing for uptime (hardware redundancy) and some other things but I always remind that RAID in and of itself is not backup.

Soapbox off...

[Kyhl]

Then there is the whole Raid process that would be best when combined with backups.

[Whoaru99]

In some regards I'm not unhappy that I didn't have a lot of money to spend in the heyday of vinyl.

[SkunkedAgain]

IMO - some things are better left to vinyl and some to CDs. I don't understand new bands releasing on vinyl. Music like jazz and blues really feel right on vinyl but I'll take the Beatles on CD any day over my original LPs

[Kyhl]

If I were a Beatles fan I'd say, sell me those LP's if you don't want them.

Die hard Beatles fans say that the CD reissues's have never been mastered correctly to match the vinyl versions. All of the CD's have problems with wrong generation of masters, or noise reduction added, or stereo where they should be mono, or bad eq choices, etc, etc.

Even the recent reissued vinyl has problems like being digitally sourced with all the problems listed above.

I appreciate the Beatles and what they did but honestly, I don't care about the Beatles music and might listen to a, as in one, Beatles song per year by accident on random play.

I get what you mean though. Spun a reissed Nick Drake Pink Moon last night and it was awesome. The beauty is that I also have a 24/96 digital version to listen in other places.

Hardcore classical guys will say the same thing about classical too. The digital era brought in multiple mic setups which blew up the spacial imaging of classical recordings causing individual instruments to become lost in the mix.

[Kyhl]

Another easy example of CD reissue problems is Led Zeppelin IV.

The original LZ IV pressing from the 80's - 90's were fine. Sometime (late 90's?) the channels were reversed.

Not really a big deal sound wise but when the guitar starts on Stairway To Heaven and it's on the wrong side of the room it is weird for those that grew up hearing it differenctly when it became popular. The reissues accidently reversed the right and left channels. Not to mention that Jimmy Page compressed the life out the album so the remastered versions sound like dump imo.

[Kyhl]

If anyone is interested, I uploaded a sample of the Albert King/Stevie Ray Vaughan - In Session recording that I dropped from vinyl.

The file is lossless .flac in 24/96 (high res). This has some surface noise here and there but is an example of how vinyl can sound.

PS, IMO, it sounds better played directly from vinyl.

FYI mods, this is a sample of songs and does not violate copyright rules because the file does not contain a complete song.

https://www.dropbox.com/l/XysC85iQPiTEynX6UARvie

[LMITOUT]

These .flac files are apparently quite large!

Sounds pretty good, though. Thanks for sharing.

[LMITOUT]

I decided to experiment with EAC and archive some CD's to FLAC.

Do you have any tips or "must-do's" in setting up EAC?

Thanks.

[Kyhl]

I don't get too crazy with it and just used the automated settings.

Clicked on 'detect read features', 'Examine C2 features', and 'Auto detect read commands' on the drive setup menu and used whatever it recommended.

Extraction Method is Secure Mode

On the Offset / Speed tab I check the Allow Speed Reduction and Use AccurateRip

Compression Options are:

Use external program

User Defined Encoder

File Extention: .flac

Use the browse button to point it to flac.exe. You may or may not need to download it.

Here's my command line. You might want to mess around with different options.

-8 -V -T "ARTIST=%artist%" -T "TITLE=%title%" -T "ALBUM=%albumtitle%" -T "DATE=%year%" -T "TRACKNUMBER=%tracknr%" -T "GENRE=%genre%" -T "COMMENT=%comment%" -T "BAND=%albuminterpret%" -T "ALBUMARTIST=%albuminterpret%" -T "COMPOSER=%composer%" %haslyrics%--tag-from-file=LYRICS="%lyricsfile%"%haslyrics% -T "DISCNUMBER=%cdnumber%" -T "TOTALDISCS=%totalcds%" -T "TOTALTRACKS=%numtracks%" %hascover%--picture="%coverfile%"%hascover% %source% -o %dest%

Check the boxes.

Delete WAV after compression

Use CRC check

Check for external programs return code.

High Quality

On the Comment tab I only chose Write freedb ID into ID3 tag comment field

On the ID3 Tag tab I have nothing selected.

When I rip the disks I only do a "copy selected tracks". The testing option seems redundant.

Then check the log at the end to verify that there were no errors. If there are errors then try the rip again. If errors continue then it's your choice on how to proceed. You can up the security and let it run all night, lower the security so it skips the error, or replace the disk.

I don't know what others do but I created a file structure where each artist has his own directory, then sub-directories by album.

So something like F:/Music/Flac/Artist_Name/Album_Name

and G:/Music/MP3/Artist/Album.

[Kyhl]

These .flac files are apparently quite large!

Sounds pretty good, though. Thanks for sharing.

Just for kicks I unpacked it into a .wav and it almost doubled.

Went from 155meg to 249meg.

This was at 24/96 also. CD's are 16/44.1 as a comparison. So there are twice the number of samples per second and an extra 8-bits of headroom.

[LMITOUT]

Thanks for the info. Seems like I have most of the same settings and didn't mess with it too much.

I was also wondering about the Test and Copy option and wasn't sure if the Test was critical or not.

What does the Command Line do exactly? Is this the information that is embedded into the resultant file? Reading through the switches in your line it appears that may be just that. I was wondering how to incorporate the album art and I see that is part of the command line.

The time to rip a CD is a lot more than I had thought using this method, but I suppose the time is required to get a perfect copy. Years ago I had ripped my entire CD collection using iTunes after purchasing an iPod. I've since transitioned to Google Music which imported those files into the "cloud". I understand they don't have the original CD quality but it's a compromise for the ability to listen on my phone wherever I am. It also makes it easy to listen at home via the PC. When I started to think about it I realized that I really don't have a true copy of my music collection and have some CD's that I would like to preserve. With the time it takes to rip to FLAC I might have to pick and choose which CD's I really want to keep a copy of, as I have too many CD's to even think about doing all of them.

Time to go shopping for an external HD!

Thanks again for the feedback.

[LMITOUT]

Hmmm. I guess my command line is similar so that eliminates my theory of why the album art isn't showing up. Maybe it's just Windows Media Player that I'm using to play the FLAC files.

[delcecchi]

What is the point in using more bits and a higher sampling rate than is on the original CD? I can conceive of such an argument coming from Vinyl which is analog (but whose dynamic range and S/N is hardly demanding 24 bits) (and whose bandwidth hardly requires 96k samples (48kHz BW)

I bet neither of you can hear even 20kHz, but it isn't even there coming from a CD. Well, 20 is but not much more or the sampling would alias it and be bad)

However storage bits are cheap, so if it feels good, do it. A couple TB is only around $100.

By the way, the dynamic range of a CD is at least 96dB, while vinyl is at best about 70 dB maximum.

[LMITOUT]

It seems that you're mixing bitrates (sample rate) and audible frequency?

Let me ask you this: If you want to replicate a 1 second waveform of varying frequencies, wouldn't your rather sample that segment as many times as practical to ensure it is as an exact copy as possible?

[delcecchi]

One might think that, but a guy by the name of Nyquist looked into it, many years ago...

Quote:

The Nyquist–Shannon sampling theorem, after Harry Nyquist and Claude Shannon,[1] in the literature more commonly referred to as the Nyquist sampling theorem or simply as the sampling theorem, is a fundamental result in the field of information theory, in particular telecommunications and signal processing. Sampling is the process of converting a signal (for example, a function of continuous time or space) into a numeric sequence (a function of discrete time or space). Shannon's version of the theorem states:[2]

If a function x(t) contains no frequencies higher than B hertz, it is completely determined by giving its ordinates at a series of points spaced 1/(2B) seconds apart.

In other words, a bandlimited function can be perfectly reconstructed from a countable sequence of samples if the bandlimit, B, is no greater than half the sampling rate (samples per second). The theorem also leads to a formula for reconstruction of the original function from its samples. When the bandlimit is too high (or there is no bandlimit), the reconstruction exhibits imperfections known as aliasing. The Poisson summation formula provides a graphic understanding of aliasing and an alternative derivation of the theorem, using the perspective of the function's Fourier transform.

So if the music has maximum frequencies of 20kHz you only need to sample at 40kHz. The 44kHz of CD is used to have a little margin. Anything over that is pretty much wasted.

Much of the lore in Audiophile circles is not true. Sort of like the 24 bit sampling. If the source has some noise in it, even a little, those low order bits will just be random anyway.

[Kyhl]

Hi Del,

The sample was from vinyl.

You are correct. There is no reason to add 8 bits to a CD file. It would just be 8 zeros added to the front of the volume number. It also doesn't make much sense to me to add sampling (up sampling) to 44khz recordings.

There is an argument to be made that 16 bits has plenty of dynamic range for vinyl. Although 24 bits allows for a louder recording, it also pushes the noise floor lower. From what I have found, the farther you can get from the noise floor of the DAC the better.

The 96khz sampling rate is more about noise and waveform shaping than it is about frequency of the playback. Nyquist does a great job replaying sine waves to 20khz but music is not necessarily in the shape of a sign wave. While 44khz can do shaped waves at 1khz, as you approach 20khz the waves become sine shaped.

So a 48khz sample will rebuild a square or sawtoothed wave at 1khz better than 20khz sample, and 96khz will reproduce it better still.

Noise is introduced when you process and convert the sound. Noise shaping allows that noise to be pushed up above 20khz when you use a higher sampling rate. Some say a good place to push that noise is above 30khz, others say higher. What I've read suggests that 48khz is the minimum top frequency that should be used.

And true, I can only hear to 17khz in my good ear. Only 14khz in my bad ear. The majority of music is below 12khz anyway. The treble region starts around 2khz depending on your definition. Some say more, some say less. I lean toward less. I count my mid-range, where the meat of the music is, as between 200hz and 1khz. Using that definition, high frequencies start above 1k. Male vocals dip below 200hz and I consider much of that as mid range.

[Kyhl]

Limit,

I use Foobar for my playback. It does gapless and plays about any file format. It's no frills free software that doesn't take over you computer or music files, a la iTunes. It allows you to manage tags and cover art, although you have to add the art manually. It also converts files for you. I use it to convert .wav's to .flac and MP3's.

And it has a free remote in the google play store so you can run it from the couch over a wi-fi connection.

I feel it's playlist GUI is clumsy and may eventually be the deal breaker.

I've thought about trying JR Music Center but I'm too cheap when Foobar does it for free, until I get tired of the playlist GUI.

[LMITOUT]

One might think that,

More than one "thinks that" and I'm well aware of Nyquist theorems. You may also note that I said "as many times as practical", with practical being the key word.

[delcecchi]

If the source is band limited to 20khz, then a 20khz wave form by definition looks like a sine wave. So, yes you are correct but also incorrect in that the components of a 20kHz triangle wave for example that make it look triangular are at multiples of 20kHz. So if you want to reproduce the 40 or 60 kHz components for some reason then you need a higher sampling frequency.

I would have to read up on all the noise shaping stuff in order to comment. I will say that much of what is published for "audiophiles" in areas that I am educated and experienced in is hooey, so I tend to presume what I read in other areas written by the same type of people is also hooey.

And the 24 bits wouldn't have 8 zeros on the front. That would be silly. It would be properly scaled in the input stage to use full range. There would however be an additional 8 bits of essentially random data on the low order end.

For example most of the stuff written about speaker cables is nonsense. I particularly thought the idea that cables need break in and are directional to be false. (to be polite about it).

[Whoaru99]

There are directional cables, inasmuch the intent of the cable designer regarding the connection of the shield floating on one end and whether the floating end should be connected to the source or destination.

However, since musical signals are AC in nature, certainly the music doesn't "flow" in just one direction of any cable.

[delcecchi]

So far as I know, few speaker cables are shielded. Speaker cables are notorious for the amount of bovine excrement used to justify charging high prices.

And I recall reading TAS or Stereophile or one of those where the guy was talking about "auditioning" some speaker cables and needing to let them "break in" for a couple weeks...