How to rip CDs for your Ipod

[Patrick Bateman]

Here's the best way to rip CDs for your Ipod. It takes 10x as long as
the normal method, but it's the best way to do it. Five years ago I
wrote a script that does it automatically, but I lost track of it a
few months ago. So I thought I'd post the method here, for my own
reference, and in case you guys are interested. In the car I mostly use an Ipod. I store some music uncompressed, but 90% of what I listen to doesn't suffer much from well-done compression.

There are three big problems with putting songs on your Ipod. The
first one is that the data compression causes audible artifacts in the
playback, particularly at low bit rates. You can hear the errors as a
"swishing" noise at high frequencies. The best way to deal with the
data compression is to use a high bit rate.

The second problem with putting music on your Ipod is that the
playback levels are all over the map. Doesn't it drive you crazy when
you listen to a playlist and one track is recorded at a level 10db
higher than the next? A popular method of adressing this is with
"replygain" in Foobar2000.

One of the biggest problems with playing music back on your Ipod is
that compact discs are recorded at a sampling frequency that doesn't
match Ipods, DVD, PS3, etc... Why should you care? Because changing
the sampling rate *incorrectly* leads to a litany of problems with the
sound of your tracks. It's like putting a square peg in a round hole.
Some of the people on the list have messed around with audio editors
I'm certain, and you've seen what happens to a wav file when it's
resampled poorly. Therefore, resampling your files properly maximizes
the playback quality.

If you follow these steps, you'll notice a number of improvements with
your music. The first thing you'll notice is that your tracks are
quieter. A *lot* quieter. The reason for this is that modern music
is recorded at a level which is way too high. This is done to make
music sound more "exciting" and to "stand out." It comes at a high
price however; it kills the dynamic range of the original track.
You'll address that by increasing the bit depth by 50%. The other
improvement you'll notice is that the treble, and the higher octaves
in general, don't sound as "hashy." This is thanks to using a
"modern" sample rate (48khz.) CD players use a sample rate of
44.1khz, but all modern gear uses a sample rate of 48khz or higher,
and improper resampling leads to that high frequency hash we've all
grown accustomed to. Last but not least, you'll find yourself playing
your music much LOUDER, because everything sounds better now :)

Enough of the hyperbole, here's how ya do it:

First, start with a lossless format, like wav, ape, or flac. Open up
foobar2000 and load up the files. Right click on the files and choose
"convert." You're going to convert them from wav to wav. (or ape to
wav, or flac to wav, etc...) During the conversion process, choose
and output depth of 24bit and a sample rate of 48khz. By doing this
we're creating a "container" which is more than twice as large. This
is comparable to taking a picture in photoshop, then doubling it's
size. These extra bits will come in handy later, when we're running
our filters.

We'll use two filters during the conversion process. The first is
"replaygain" which normalizes the volume of all your tracks. Select
"apply gain and prevent clipping according to peak." The second is
the DSP filter "pphs resampler." We're going to convert from 44.1khz
to 48khz. Last but not least, we need to select a 24bit wav file.
This step is the most important of all. A 16bit file is so 1982, it's
not even funny. It's 2008, I think we can afford the space for a
24bit file. A 24bit file gives us a lot more room to play with
filters while minimizing conversion artifacts. You can choose 24bit
output by clicking "more settings" in the converter setup, and
selecting "24 bit" under "output bit depth."

By the way, 24bit/48khz is the highest we can go. MP3 won't support
anything higher. The only way to go higher is to use a different
output format than mp3, such as DVD-Audio (which the Ipod won't play.)

The conversion process takes about 5-10 minutes for an album. When
it's done, you'll have a folder full of 24bit/48khz wav files. You
should double check the file properties to be sure you did it right.
I've attached a pic of what it should look like.

Now we have this big fat juicy wav file to convert. When it comes to
mp3 conversion, it's "garbage in / garbage out." The last few steps
insure we're feeding our converter the best file possible.

You can convert those wav files with foobar2000, but I do it on the
command line. I do it with lame using this command:

lame -s48000 -V0 --noreplaygain --ta [track artist] --tn [track
number] --tt [track title] --tl [album] [wav file] [mp3 file]

Here's an example:

lame -s48000 -V0 --noreplaygain --ta "Prince" --tn 2 --tt "The Future"
--tl "Batman OST" "1 The Future.wav" "Prince - The Future 24-48.mp3"

Normally I wrap a dos batch script around this to make it fast, and I
convert a few dozen at a time.

Hit me up if ya' got questions. Let me know what you think of the difference.

[newtitan]

WOW, I do this for the most part, but very nice writeup, and nice inform on the replay gain, never used that one before, and thx for the script too

[Et Cetera]

great writeup! Thanks!

[txbonds]

I'm tagging on so I can find it again. Want to give this a try myself. Thanks for taking the time to post it.

[Selkies]

Nice info! Another process to consider while I think about digitally archiving my music...

[snaimpally]

Interesting. But why spend time converting to 24 bit when the original files are 16 bit? The resolution of the files doesn't increase by converting them to a higher resolution. If you take an mp3 encoded at 64 bits and convert it to 44.1Khz wave, its not like you have actually increased the quality of the source material, you have just changed the sampling rate.

I am just trying to understand why you do the extra step. I have learned a lot on this forum so I am open to learning something new.

[robbyho]

Interesting. But why spend time converting to 24 bit when the original files are 16 bit? The resolution of the files doesn't increase by converting them to a higher resolution. If you take an mp3 encoded at 64 bits and convert it to 44.1Khz wave, its not like you have actually increased the quality of the source material, you have just changed the sampling rate.

I am just trying to understand why you do the extra step. I have learned a lot on this forum so I am open to learning something new.

x2, does "upconverting" actually do anything beneficial?

Also, what bit rate are your final files in your ipod? I've recently been putting in all of my cd's into apple lossless and they are at around 1000 kbps. I've got 10 albums with lossless, 30 at 320 kbps, and 1500 songs at 128 and 192. That takes up 15 gigs on my 60 gig ipod. Now sure how much room I'll have when I finally convert them all to lossless.

Robby

[t3sn4f2]

Here's the best way to rip CDs for your Ipod. It takes 10x as long as
the normal method, but it's the best way to do it. Five years ago I
wrote a script that does it automatically, but I lost track of it a
few months ago. So I thought I'd post the method here, for my own
reference, and in case you guys are interested. In the car I mostly use an Ipod. I store some music uncompressed, but 90% of what I listen to doesn't suffer much from well-done compression.

There are three big problems with putting songs on your Ipod. The
first one is that the data compression causes audible artifacts in the
playback, particularly at low bit rates. You can hear the errors as a
"swishing" noise at high frequencies. The best way to deal with the
data compression is to use a high bit rate.

The second problem with putting music on your Ipod is that the
playback levels are all over the map. Doesn't it drive you crazy when
you listen to a playlist and one track is recorded at a level 10db
higher than the next? A popular method of adressing this is with
"replygain" in Foobar2000.

One of the biggest problems with playing music back on your Ipod is
that compact discs are recorded at a sampling frequency that doesn't
match Ipods, DVD, PS3, etc... Why should you care? Because changing
the sampling rate *incorrectly* leads to a litany of problems with the
sound of your tracks. It's like putting a square peg in a round hole.
Some of the people on the list have messed around with audio editors
I'm certain, and you've seen what happens to a wav file when it's
resampled poorly. Therefore, resampling your files properly maximizes
the playback quality.

If you follow these steps, you'll notice a number of improvements with
your music. The first thing you'll notice is that your tracks are
quieter. A *lot* quieter. The reason for this is that modern music
is recorded at a level which is way too high. This is done to make
music sound more "exciting" and to "stand out." It comes at a high
price however; it kills the dynamic range of the original track.
You'll address that by increasing the bit depth by 50%. The other
improvement you'll notice is that the treble, and the higher octaves
in general, don't sound as "hashy." This is thanks to using a
"modern" sample rate (48khz.) CD players use a sample rate of
44.1khz, but all modern gear uses a sample rate of 48khz or higher,
and improper resampling leads to that high frequency hash we've all
grown accustomed to. Last but not least, you'll find yourself playing
your music much LOUDER, because everything sounds better now :)

Enough of the hyperbole, here's how ya do it:

First, start with a lossless format, like wav, ape, or flac. Open up
foobar2000 and load up the files. Right click on the files and choose
"convert." You're going to convert them from wav to wav. (or ape to
wav, or flac to wav, etc...) During the conversion process, choose
and output depth of 24bit and a sample rate of 48khz. By doing this
we're creating a "container" which is more than twice as large. This
is comparable to taking a picture in photoshop, then doubling it's
size. These extra bits will come in handy later, when we're running
our filters.

We'll use two filters during the conversion process. The first is
"replaygain" which normalizes the volume of all your tracks. Select
"apply gain and prevent clipping according to peak." The second is
the DSP filter "pphs resampler." We're going to convert from 44.1khz
to 48khz. Last but not least, we need to select a 24bit wav file.
This step is the most important of all. A 16bit file is so 1982, it's
not even funny. It's 2008, I think we can afford the space for a
24bit file. A 24bit file gives us a lot more room to play with
filters while minimizing conversion artifacts. You can choose 24bit
output by clicking "more settings" in the converter setup, and
selecting "24 bit" under "output bit depth."

By the way, 24bit/48khz is the highest we can go. MP3 won't support
anything higher. The only way to go higher is to use a different
output format than mp3, such as DVD-Audio (which the Ipod won't play.)

The conversion process takes about 5-10 minutes for an album. When
it's done, you'll have a folder full of 24bit/48khz wav files. You
should double check the file properties to be sure you did it right.
I've attached a pic of what it should look like.

Now we have this big fat juicy wav file to convert. When it comes to
mp3 conversion, it's "garbage in / garbage out." The last few steps
insure we're feeding our converter the best file possible.

You can convert those wav files with foobar2000, but I do it on the
command line. I do it with lame using this command:

lame -s48000 -V0 --noreplaygain --ta [track artist] --tn [track
number] --tt [track title] --tl [album] [wav file] [mp3 file]

Here's an example:

lame -s48000 -V0 --noreplaygain --ta "Prince" --tn 2 --tt "The Future"
--tl "Batman OST" "1 The Future.wav" "Prince - The Future 24-48.mp3"

Normally I wrap a dos batch script around this to make it fast, and I
convert a few dozen at a time.

Hit me up if ya' got questions. Let me know what you think of the difference.

I think he's saying that you want to convert your 16/44 files in foobar over letting the ipod up convert it to 48 (it has to do this regardless of the original file sample rate and bits in order to play) which is an inferior conversion process over foobars.

And while you are at it bump up the bits so that the files better suited for replay gain or mp3 conversion later if you like.

[ehiunno]

Now you just have to deal with the crappy output stage of the ipod and the HPF they through on there so you don't break your cheap earbuds.

Great tutorial, but your SQ is still choked by the output stage of the ipod even if you have a great mp3. This and poor listening devices are the key reason why so many people say they don't notice a difference between 128 and lossless. Put the same track of both on an ipod, listen to it out of the 1/8th inch jack or analog line out and the difference will still be minimal, even on a reference system.

[Autiophile]

Now you just have to deal with the crappy output stage of the ipod and the HPF they through on there so you don't break your cheap earbuds.

Great tutorial, but your SQ is still choked by the output stage of the ipod even if you have a great mp3. This and poor listening devices are the key reason why so many people say they don't notice a difference between 128 and lossless. Put the same track of both on an ipod, listen to it out of the 1/8th inch jack or analog line out and the difference will still be minimal, even on a reference system.

I believe the frequency of that HPF (due to the DC blocking caps) is in part dictated by the input impedance of the amp/headphone being used. Hook it up to a headphone amp with an impedance of 1K+ ohms and you drive the cutoff frequency lower. A reference system should have no problems if using a headphone amplifier (hell, Wilson Audio used an iPod as a source for a demo system). I certainly agree it can be even better if they use a digital output, which speaks to your point about the less than stellar analog outputs.

The shuffle has no capacitors in the output and thus has better measured bass performance.

Couple interesting sites below. iPods have been pretty thoroughly measured and it has exposed both flaws and some things apple did right.

http://members.chello.nl/~m.heijligers/ipod/Performance/measurements.html

http://homepage.mac.com/marc.heijligers/audio/ipod/engineering/engineering.html

And the obligatory quote from Wikipedia (understood to be a non-definitive resource)Bass response

The third generation iPod had a weak bass response, as shown in audio tests.[44] (http://en.wikipedia.org/wiki/IPod#cite_note-43)[45] (http://en.wikipedia.org/wiki/IPod#cite_note-44) The combination of the undersized DC-blocking capacitors (http://en.wikipedia.org/wiki/Capacitor) and the typical low-impedance (http://en.wikipedia.org/wiki/Electrical_impedance) of most consumer headphones form a high-pass filter (http://en.wikipedia.org/wiki/High-pass_filter), which attenuates the low-frequency bass output. Similar capacitors were used in the fourth generation iPods.[46] (http://en.wikipedia.org/wiki/IPod#cite_note-45) The problem is reduced when using high-impedance headphones and is completely masked when driving high-impedance (line level) loads, such as an external headphone amplifier (http://en.wikipedia.org/wiki/Headphone_amplifier). The first generation iPod shuffle uses a dual-transistor output stage (http://en.wikipedia.org/wiki/Bridged_and_paralleled_amplifiers#Bridged_amplifie r),[47] (http://en.wikipedia.org/wiki/IPod#cite_note-46) rather than a single capacitor-coupled output, and does not exhibit reduced bass response for any load.

[WRX/Z28]

I've noticed a difference from 320kbps and lossless. What are the gains in this method over lossless, do you still keep your albumn art and albumn/track names? Does gracenote still recognize everything and name it? I wouldn't want to go through and name my whole library, i'd be there for days.

[ehiunno]

Interesting point that I wasn't aware of. In my personal experience using the headphone jack of the ipod 5g into the aux input of my headunit, I noticed a significant decrease in bass, overall output, and total SQ as compared to the same CD track, or the same track from my laptop.

I am surprised by that statement about the shuffle though... when my gf plugged hers into the jack I usually plug mu ipod 5g into I noticed the output level was significantly weaker (weak enough to get me to change my DCX-730 sensitivity from 1v to .2v) and there was very, very little low frequency output. It was much worse than the regular ipod...

I've seen some really cool sites around detailing how to mod the ipod to overcome the output stage problem, apparently with fantastic results. Anyone using an ipod in their car as a main source should talk to the guys on headfi.com. They know their stuff.

[Autiophile]

I am surprised by that statement about the shuffle though... when my gf plugged hers into the jack I usually plug mu ipod 5g into I noticed the output level was significantly weaker (weak enough to get me to change my DCX-730 sensitivity from 1v to .2v) and there was very, very little low frequency output. It was much worse than the regular ipod...

I've seen some really cool sites around detailing how to mod the ipod to overcome the output stage problem, apparently with fantastic results. Anyone using an ipod in their car as a main source should talk to the guys on headfi.com. They know their stuff.

My experience has been good with the white shuffle than looks like a pack of gum. I've not used the more recent versions of the shuffle nor have I seen measurements for them. In my car I used the alpine full speed connection.

I hang out on head-fi a good bit. If you can dial your bullshit filter up to maximum you can get some great information. There are some guys on there who are in serious pursuit of high value audio, the length of the Zero 24/96 DAC thread shows just how much people want high value components.

[ehiunno]

I've not used the more recent versions of the shuffle nor have I seen measurements for them.

If you can dial your bullshit filter up to maximum you can get some great information.

Ah, that makes sense. Hers is one of the newer square clip on ones.

Good advice about the bullshit filter LOL. They do seem to have a significant number of *true* audiophiles :rolleyes:. I'll have to pick up some Eichmann $100 RCA ends to go with the $200/pc caps that I need to use in the output stage of the ipod. NthxKthx. Joking aside, still a great source for information about getting the best quality out of digital music with few compromises.

[HooRide]

Why not just use the actual line-out on the dock connector and avoid the headphone circuitry completely...? I believe that's how stereophile did their review.

[MIAaron]

No joke. The headphone out going to a stereo sucks, but the line out of the dock connector does great. I think I paid $12-$15 for mine and it made my ipod a tolerable music source for my home & car audio systems.

[ehiunno]

Meh, the output stage is still subpar even if you use the line out IMO, though its definitely better than the headphone out. The general consensus on head-fi is that neither of them are good sources for critical listening, and my experience tends to lie up with it.

[txbonds]

How hard would it have been for apple to have included a digital out on the ipods? I have a 5g video 80gig that is nice, but I sure wish it had better output. On the 5g video, you can get them converted for better output, but you have to use special cables from that point forward, as well as a headphone preamp, which doesn't work for car audio as far as I can tell.

[Autiophile]

Digital output would be very nice and I'd be surprised if we didn't see it on some future version (probably an expensive version unfortunately). Obviously Apple is aware of the desire, given the digital output on macs and airport express units.

Currently, you have to depend on the aftermarket to let you get bit perfect output via a device like Wadia's iTansport (http://blog.stereophile.com/ces2008/010708wadia/). It would appear that any manufacturer should be able to do something similar in a car audio head unit once they figure out how to get the output before the iPod's DAC.

All that said, I've got a 40gb 4th gen ipod that I use for all my lossless files. I use the headphone output to a Xin Supermacro and use an assortment of headphones and occasionally hook it up to my home system with the Supermacro ahead of my preamp. Honestly, it sounds pretty damned good. Maybe my ears are less discerning than most (I think that's unlikely) but I find it sacrifices very little when compared to my Rotel or Rega CD players. Using either the Mac Pro or Airport Express (both with optical outputs) are even closer to the CD players. For 99% of my listening the apple devices are completely sufficient. That other 1% of the time I consider to be "critical listening" and use the CD players, and even then I'm not sure there's a discernible difference.

[Autiophile]

On the 5g video, you can get them converted for better output, but you have to use special cables from that point forward, as well as a headphone preamp, which doesn't work for car audio as far as I can tell.

I wouldn't necessarily advocate modifying iPods (redwine or others) because I'm not sure the results justify the cost, but I don't see how they preclude you from using them in a car. All the headphone amp is doing is amplifying the signal. It's still analog. You could either use the headphone amp as the output to your car's system or you could just forget it and use the analog output from the modded ipod and take it directly to your system (albeit at a lower line level).

Special cables are bullshit. Analog is analog. A little splicing and you'll be back to the analog connection of your choice (1/8, rca, etc.).

[txbonds]

I wouldn't necessarily advocate modifying iPods (redwine or others) because I'm not sure the results justify the cost, but I don't see how they preclude you from using them in a car. All the headphone amp is doing is amplifying the signal. It's still analog. You could either use the headphone amp as the output to your car's system or you could just forget it and use the analog output from the modded ipod and take it directly to your system (albeit at a lower line level).

Special cables are bullshit. Analog is analog. A little splicing and you'll be back to the analog connection of your choice (1/8, rca, etc.).

The redwine special cable deal is because they claim to put some of the required capacitors into the cable due to space restrictions in the unit itself on the 5gen units.

[Autiophile]

The redwine special cable deal is because they claim to put some of the required capacitors into the cable due to space restrictions in the unit itself on the 5gen units.

Just cut the cable off after the capacitors and you're done.

Edit: Better yet, just find out what caps they use and make your own cable rather than paying audiolineout's crazy ass prices.

[txbonds]

Just cut the cable off after the capacitors and you're done.

Edit: Better yet, just find out what caps they use and make your own cable rather than paying audiolineout's crazy ass prices.

They make a dock adapter that lets you use your own cable, but even it is $85.

If I were going to do the Redwine mod, I'd probably attempt to do it myself using one of the many DIY threads or writeups out there. Cost is about $20 for doing it yourself, versus the $250 mod plus $150 cables that redwine/ALO wants.

Here is one example of DIY: http://ayl.nuwen.net/index.php/projects/diy-imod-ipod/

[Autiophile]

They make a dock adapter that lets you use your own cable, but even it is $85.

Hell, they show the caps in the picture of the $340 dock. Just need to track those caps down (or caps with the same values) and save probably 95%.http://aloaudio.com/store/catalog/images/imod-fidelity-dock.jpg

[txbonds]

Yeah, duplicating what they are doing doesn't look that hard as long as you can handle the soldering and desoldering part. My problem with hacking up the ipod is more the lack of use for any other use. In other words, once you do most of those mods, you can't use it for general purpose stuff any more and have to always have those caps in use.

I use mine on a small radio in my office to play random music throughout the day. It's one of those radios that is basically a speaker with some black knobs, and a dock on top. Works great, but definately wouldn't be of benefit or usable with a moded ipod.

If I were to go after a moded unit, I'd have to consider a second unit and that's just too much ipod for me. LOL

[Autiophile]

They are $22.75 a piece.http://www.audience-av.com/capacitors/a_prices.php. Though you could just get some Xicons (or any other 2.2uF 200V capacitor) for less than a buck a piece.

Grab a refurb iPod. There are some deals out there. My 4th gen refurb is going strong. I actually prefer it to my 5th gen.

[Patrick Bateman]

Interesting. But why spend time converting to 24 bit when the original files are 16 bit? The resolution of the files doesn't increase by converting them to a higher resolution. If you take an mp3 encoded at 64 bits and convert it to 44.1Khz wave, its not like you have actually increased the quality of the source material, you have just changed the sampling rate.

I am just trying to understand why you do the extra step. I have learned a lot on this forum so I am open to learning something new.

I convert to 48khz because that frequency is more common these days than 44.1khz. I was doing some recording on my PC, using the motherboard sound card, and discovered that a crummy Windows driver was mangling the output because it was doing a terrible job converting from 44.1khz to 48khz. I wish I had pictures of the waveforms - they were terrible! All kinds of artifacts, the output was a complete mess.

So basically I can't rewrite the software, but I can transcode my MP3s so they're "aligned" with the most popular bitrate in the world now: 48khz.

As far as going with 24bit, that's handy because it gives us more headroom for replaygain.

To picture it visually, imagine if you were going to touch up a photo in Photoshop. Before you tweak the brightness or the contrast, it's a good idea to use more bits, because it gives you more flexibility when you change it. If you've ever tweaked the brightness of a photo and it ended up washed out, you know what I mean.

[Patrick Bateman]

x2, does "upconverting" actually do anything beneficial?

Also, what bit rate are your final files in your ipod? I've recently been putting in all of my cd's into apple lossless and they are at around 1000 kbps. I've got 10 albums with lossless, 30 at 320 kbps, and 1500 songs at 128 and 192. That takes up 15 gigs on my 60 gig ipod. Now sure how much room I'll have when I finally convert them all to lossless.

Robby

I use VBR at the most aggressive settings possible (biggest filesize.)

[Patrick Bateman]

I think he's saying that you want to convert your 16/44 files in foobar over letting the ipod up convert it to 48 (it has to do this regardless of the original file sample rate and bits in order to play) which is an inferior conversion process over foobars.

And while you are at it bump up the bits so that the files better suited for replay gain or mp3 conversion later if you like.

Yes! The whole idea is to do the transcoding in software, rather than letting the Ipod / PC / whatever do the transcoding in real time.

Real time transcoding = FAIL

[Patrick Bateman]

I've noticed a difference from 320kbps and lossless. What are the gains in this method over lossless, do you still keep your albumn art and albumn/track names? Does gracenote still recognize everything and name it? I wouldn't want to go through and name my whole library, i'd be there for days.

One of the reasons I like mp3 over mp4 is that I use the BPM tag for playlists. So it's important to me to retain all the album art and all the tags. So it's a regular ol' mp3. Once you create the file, you can add tags, artwork, etc...

[Patrick Bateman]

I wouldn't necessarily advocate modifying iPods (redwine or others) because I'm not sure the results justify the cost, but I don't see how they preclude you from using them in a car. All the headphone amp is doing is amplifying the signal. It's still analog. You could either use the headphone amp as the output to your car's system or you could just forget it and use the analog output from the modded ipod and take it directly to your system (albeit at a lower line level).

Special cables are bullshit. Analog is analog. A little splicing and you'll be back to the analog connection of your choice (1/8, rca, etc.).

And one greaaaaaat thing about the Ipod is NO NOISE. In my last car I had a blingy head unit, and I could NEVER GET RID OF THE F'ING ALTERNATOR WHINE.

[Patrick Bateman]

I wrote a script that automates the wav to mp3 conversion process.
Basically run it in a folder of wav files and it does all the heavy
lifting for you. The cool part is it preserves the tags. The only
prereq is that the wav files are 24bit/48khz. My previous post
explains how to do that in foobar2000. Also, they must be named a
specific way.

As long as you follow these guidelines, creating 24bit 48khz mp3s is a
cakewalk now. I did it with cygwin on xp; change one line for linux
or osx.

Here's the script.

$ cat /cygdrive/c/windows/system32/wavtomp3.sh
#!/bin/bash
#
# converts 24/48wavs to 24/48mp3 with id3 tags
#
# wavtomp3.sh [filename.wav]
# The wav file must be in the following format:
# artist-album-track.wav
# note there are no spaces between the dashes
# for example "Rolling Stones-Tumbling Dice-Exile on Main St.wav"
#
for i in *.wav; do
FILENAME=$i
#
echo "FILENAME is $FILENAME"
ALBUM=`echo "$FILENAME" | cut -d - -f 2`
echo "ALBUM is $ALBUM"
ARTIST=`echo "$FILENAME" | cut -d - -f 1`
echo "ARTIST is $ARTIST"
TRACK=`echo "$FILENAME" | cut -d - -f 3`
TRACK=`echo "$TRACK" | cut -d . -f 1`
echo "TRACK is $TRACK"
# using a cygwin path here; change path if using linux or osx
/cygdrive/c/windows/system32/lame.exe -s48000 -V0 --noreplaygain --ta "$ARTIST"
--tt "$TRACK" --tl "$ALBUM" "$FILENAME" "$ARTIST - $ALBUM - $TRACK.mp3"
done

[gsr22]

foobar doesnt work for mac any other alternatives t

[Pseudonym]

I wrote a script that automates the wav to mp3 conversion process.
Basically run it in a folder of wav files and it does all the heavy
lifting for you. The cool part is it preserves the tags. The only
prereq is that the wav files are 24bit/48khz. My previous post
explains how to do that in foobar2000. Also, they must be named a
specific way.

As long as you follow these guidelines, creating 24bit 48khz mp3s is a
cakewalk now. I did it with cygwin on xp; change one line for linux
or osx.

Here's the script.

$ cat /cygdrive/c/windows/system32/wavtomp3.sh
#!/bin/bash
#
# converts 24/48wavs to 24/48mp3 with id3 tags
#
# wavtomp3.sh [filename.wav]
# The wav file must be in the following format:
# artist-album-track.wav
# note there are no spaces between the dashes
# for example "Rolling Stones-Tumbling Dice-Exile on Main St.wav"
#
for i in *.wav; do
FILENAME=$i
#
echo "FILENAME is $FILENAME"
ALBUM=`echo "$FILENAME" | cut -d - -f 2`
echo "ALBUM is $ALBUM"
ARTIST=`echo "$FILENAME" | cut -d - -f 1`
echo "ARTIST is $ARTIST"
TRACK=`echo "$FILENAME" | cut -d - -f 3`
TRACK=`echo "$TRACK" | cut -d . -f 1`
echo "TRACK is $TRACK"
# using a cygwin path here; change path if using linux or osx
/cygdrive/c/windows/system32/lame.exe -s48000 -V0 --noreplaygain --ta "$ARTIST"
--tt "$TRACK" --tl "$ALBUM" "$FILENAME" "$ARTIST - $ALBUM - $TRACK.mp3"
done

for the tards, such as myself, who only know how to hit play on foobar to make the magical sounds come from the speakers, could you explain in as easy a way possible how to do what ive quoted here? where does this info go and such?

[aboof]

So, what about the newer headunits that use USB connections to the iPod? Aren't they bypassing the iPod's DAC (and probably also the mp3 decoder)? My assumption was that they were transferring the mp3 data files via USB to the headunit, and that the headunit was doing the mp3 decoding and DAC. This also has to be the case with the headunits that accept mp3s on USB drives or SD cards, since those devices obviously don't have DACs or mp3 decoders.

As far as upsampling to 48kHz or going to 24bit, I've never seen this suggested. I assumed that 44.1kHz was the sample rate more commonly native to hardware other than sound blaster cards in PCs and the like, and that you'd be better off sticking with what's more common than letting your devices resample, which they may not be great at. If 48kHz really is the more common native sample rate in the various devices you intend to do playback on, then I guess it would be advantageous to use that? But I'd be surprised if that were the case, since CDs are 44.1kHz and I don't know of any popular ripping tools that upsample by default, meaning that the vast, vast majority of files out there in the wild are going to be 44.1kHz - if I were designing a hardware mp3 player and knew my resampling wasn't going to be great, I'd definitely use 44.1kHz, but maybe that's just me or maybe I'm misunderstanding something.

I'd be pretty surprised if the iPod didn't do at least as well playing backing 44.1kHz files as anything else, since iTunes' default is 44.1kHz when ripping a CD. Not sure on songs from the iTunes store, but I'd guess it's the same.

I'm a little dubious that unsampling would really do anything to improve sound, as some claim, unless your playback device is upsampling and doing a really shitty job of it. Can anyone confirm that the iPod's DAC is natively 48kHz and that its resampling process is any worse than a good software resampler? Because I'd find this very surprising.

I'm more of a computer guy than a sound guy, so I'm always suspicious of stuff like this - you cannot restore data that was never there to begin with, you can only create it. Which is what upsampling does - it looks at the data that you have, and makes guesses about what to fill in the blanks with. To carry on with the OP's Photoshop analogy, you can take a 640x480 image, and convert it to an 1600x1200 image, if you like, but nothing magic happened. Photoshop just guessed at what the values of all those new pixels should be, and you don't actually end up with any more meaningful data than you originally had in your 640x480 image, which is readily apparent visually - you may have an image that's now filling up your monitor's screen instead of just a fraction of it, but it looks pixelated and blurry, because there's no new meaningful data in the larger image, just new data interpolated from the content in the original. If you stand far enough away from the monitor that you can no longer see the pixelization and blurriness, then you're also far enough away now that the image looks as small as the original did.

The analogy with Photoshop for 24bit depth for editing may be good, though - at least in Photoshop, it is good to do your edits in a large color space (e.g., AdobeRGB or ProPhoto), as long as you have a color bit-depth that's also large enough - otherwise it can actually be worse to use a large gamut - editing in ProPhoto with a 24bit color depth is not recommend, for example - it would be analogous to a sound format that included a huge range of inaudible frequencies, say, .0001Hz - 1,000,000 Hz, and then tried mapping that to the same bit-depth, resulting in the entire audible spectrum now having much fewer of those possible values to represent it, since so many of the possible values are being 'wasted' on inaudible tones (that's the ProPhoto + 24bit color case). But I'm getting off track, because I don't think this is analogous to what's happening with audio data.

By the way, if you get an image that looks "washed out" after making edits in Photoshop, it's usually a case of a color profile mismatch, not a result of not having enough color depth or a large enough gamut. The smallest gamut anyone really uses is sRGB, which has enough range to cover any color a monitor can display. If a computer is misinterpreting an sRGB file as AdobeRGB, or vice-versa, you will probably get that famous washed-out look, but it's not because of any limitation in either gamut, it's just a case of the computer making the wrong assumption about what colors the number correlate to. The larger color spaces typically really only matter for print work, where there are more colors available than computer monitors are capable of reproducing. Using a color depth that's too small typically results in 'banding' or 'posterization', where what's supposed to be a smooth tonal gradient is instead marked by visible bands of adjacent colors. Okay really off track now, sorry!

I don't think it can hurt to use a larger bit depth for your audio edits (e.g. ReplyGain, etc), but I'm still not positive that it really gains you anything, if you're starting out and ending up at the lower bit depth. My general philosophy with ripping and encoding is to just get as exact a copy as I can of the original (limited by file size, which is why I use VBR mp3s), and then if I want to do any DSP, I'll do it at playback time.

By the way, this is the guide I'd recommend for ripping on Windows, starting with a CD and producing mp3s:

http://www.fryth.com/eacfaq/

Using this, you should end up with mp3s are that utterly indistinguishable by ear from the original CD, no matter how good your ears and/or equipment, except maybe in some extremely rare combinations of ears/equipment/songs. If you are paranoid, you can use -V0 instead of -V2. It also doesn't do any replaygain / normalization, which I've always been paranoid to do on my rips. I figure if I ever decide to do it, I can do it during playback, and then I won't have made a decision that requires reripping my collection if I regret it down the road.

Which is also probably my strongest argument against upsampling to 48kHz when ripping your collection. Even if software upsampling is the way to go, (which I'm dubious about), when you do this, you are using software algorithms that decide how best to interpolate the 'missing' data. If, next year, a new algorithm comes out that does a much better job, you will have to re-rip your collection from CD to take advantage of it, which takes forever. Since I ripped my collection years ago, storage wasn't really cheap enough to make lossless a possibility, but if I were doing it over today, I'd probably choose a lossless codec, and rip my entire collection with the goal of making as exact copies of the original CDs as possible (i.e., without doing any resampling or DSP edits such as replaygain). I'd keep these files as my 'archive masters' - then, if I wanted, I could make transcoded / lossy / upsampled / replaygained / device-specific / whatever versions at any point in the future without having to go through the process of getting the data off of the CDs again.

[ngsm13]

By the way, this is the guide I'd recommend for ripping on Windows, starting with a CD and producing mp3s:

http://www.fryth.com/eacfaq/

Using this, you should end up with mp3s are that utterly indistinguishable by ear from the original CD, no matter how good your ears and/or equipment, except maybe in some extremely rare combinations of ears/equipment/songs. If you are paranoid, you can use -V0 instead of -V2. It also doesn't do any replaygain / normalization, which I've always been paranoid to do on my rips. I figure if I ever decide to do it, I can do it during playback, and then I won't have made a decision that requires reripping my collection if I regret it down the road.

Which is also probably my strongest argument against upsampling to 48kHz when ripping your collection. Even if software upsampling is the way to go, (which I'm dubious about), when you do this, you are using software algorithms that decide how best to interpolate the 'missing' data. If, next year, a new algorithm comes out that does a much better job, you will have to re-rip your collection from CD to take advantage of it, which takes forever. Since I ripped my collection years ago, storage wasn't really cheap enough to make lossless a possibility, but if I were doing it over today, I'd probably choose a lossless codec, and rip my entire collection with the goal of making as exact copies of the original CDs as possible (i.e., without doing any resampling or DSP edits such as replaygain). I'd keep these files as my 'archive masters' - then, if I wanted, I could make transcoded / lossy / upsampled / replaygained / device-specific / whatever versions at any point in the future without having to go through the process of getting the data off of the CDs again.

I came in here to post that link too.

I think this is NOT the best tutorial. Ripping accurate from the source is the best way to go. WITHOUT resampling OR replaygain...

Also, EAC (ExactAudioCopy), the Internet standard for mp3 ripping is the best way to go as well. I always rip in v0 or flac.

nG

[aboof]

I came in here to post that link too.

I think this is NOT the best tutorial. Ripping accurate from the source is the best way to go. WITHOUT resampling OR replaygain...

Also, EAC (ExactAudioCopy), the Internet standard for mp3 ripping is the best way to go as well. I always rip in v0 or flac.

nG
Yeah, I agree. It's not that I can hear the difference between -V0 (or -V2) and FLAC, because I can't, but if I were doing it over again today, I'd rip to FLAC just so that I would be future-proofed. Because if you care about quality, one thing you definitely won't do is transcode from one lossy format to another. So, what if in 20 years, the new standard is some lossy codec other than mp3, and all the music-playing devices being sold are built for that new format? As it stands now, I'd have to re-rip my collection from CD, since it's a bad idea to convert from one lossy codec to another. If I'd ripped to FLAC though, then all I have to do is transcode from FLAC to the new format, which is perfectly fine. I doubt I'd even listen to the FLACs, either - I'd simply keep them in an archive and make -V2 or -V0 mp3s from them to actually listen to, for space and compatibility reasons.