Guide to converting your Audio CDs to high quality FLAC lossless format, followed by conversion to Ogg Vorbis format, for portability.
(Copyright KVCD.Net 2004 )


Goals: To be able to backup your music collection to a high quality, lossless format (FLAC), so you can preserve your original Audio CD tracks in almost half the space required if raw extraction was used.
This way, you can store your precious music on CDs or DVDs , and encode to any present or future format.
This is the method I use to archive my music, and I currently store all my music in either DVD-R or DVD-RAM.
In the future, when new Codecs appear, I can always encode to new formats, because FLAC stores the files at almost half the size as an extracted WAV file.
Also, FLAC is a multi-platform Codec, so your files can be read on Windows, Mac OS X, Linux, etc.


(1) Preliminary steps.

You will need the following software packages:

EAC (Exact Audio Copy)
http://www.exactaudiocopy.de/eac6.html

FLAC (Free Audio Lossless Codec)
http://flac.sourceforge.net/download.html

OggDropXP
http://www.kvcd.net/downloads/oggdropXPdV1.8.4-P3.zip (For Pentium III)
http://www.kvcd.net/downloads/oggdropXPdV1.8.4-P4.zip (For Pentium 4)
http://www.kvcd.net/downloads/oggdro....4-generic.zip (Generic CPU)
And always check for latest versions here: http://www.rarewares.org/ogg.html


(2) Software installation.

2a. EAC Installation and Configuration.

Extract the downloaded file into a directory of your choice. I prefer to install at C:\EAC

After this is done, run the program eac.exe, which should be in the directory you just decompressed. Cancel the wizard, as we will configure the program with the following screenhots.

Click on the “EAC” menu option (top left), followed by clicking on “EAC Optins”, or press F9.

You should see a similar screen on the “Extraction” tab, and make sure you set your options just as the screen below.




After you set all options on the screen, select the next tab “General”.

Follow the same procedure as above, making sure all options are set just like the next screen.




Select the next tab, “Tools”, and as above, verify and correct all options to be just like the next screen.




Select the next tab, “Normalize”, and repeat the same steps as above.





Select the next tab, “Filename”, and repeat the same steps, making sure all options are set the same.




Select the next tab, “Catalog”, and repeat the same procedure.



Select the next tab, “Directories”, and repeat the same procedure.
Note: I selected the directory “C:\EAC_MUSIC”.
You may select another directory. This is where all ripped tracks will be stored, but each album you rip, will be created in a separate sub-directory with the title of the Audio CD.






Select the next tab, “Write”, and repeat the same procedure.




Select the next tab, “Interface”. and repeat the same procedure.
Note: The screen reflects an external ASPI driver, which in this case, I'm using NERO ASPI driver.
If you don't have such driver, select the “Native Win32 Interface” driver option.



After you have finished this step, you may press the “OK” button.


Now you're ready to move to the next EAC menu option, the “Drive Options”
Before you do that, make sure you select your drive in the selection on the top left of EAC window. You should see all your available CD/DVD drives. Choose the one you will be using for ripping, and insert an audioCD on that drive now.
Click on the”EAC” menu, followed by the “Drive Options”.

Disregard the information window that pop-up, and click “OK”

You should see the following screen, but this is the setup on my machine!
Click on the “Detect Read Features”, and let EAC find the correct options for your drive. When EAC is done, press “Apply” to accept the values found.




Next, select the next tab, “Drive”, and you shouold see a screen similar to the one below. Select “Auto detect read commands now” from the combo box, and press the “Auto detect read commands now” button.
In my screen, EAC detected “Read Command MMC1”.
Your value might be different.




Select the next tab, “Offset/Speed”, and enter the values exactly as shown on the following screen.




Select the next tab, “Gap Detection”, and enter the values exactly as shown on the following screen.
Note: These values are the ones that worked best for my particular drive. Your options might be different. Please try setting the values as shown on the following screen. If you have problems (you shouldn't!), then you might want to try changing these options.



You're done with these options, so press the “OK” button.



Now we are going to set the compression options.
Click on the EAC menu, and select “Compression Options”
Select the “Waveform” tab.

All options should be grayed out.
Note: They might not be grayed out, if an external encoder is not selected in the next screen. So you can really disregard the options on this screen. The "External Compression" options will override the options on this screen.





Select the next tab, “External Compression”
Enter all information just as in the screen below.




Note: Where it says “Additional Command Line Options” above, enter the following text:
-T "artist=%a" -T "title=%t" -T "album=%g" -T "date=%y" -T "tracknumber=%n" -T "genre=%m" %s


Select the “Offset” tab.
Enter all information just as in the screen below.





Select the “ID3” tab.
Enter all information just as in the screen below.



You may now press “OK”, and you have succesfully configures EAC!


2b. FLAC Installation and Configuration.

FLAC comes packaged as a self installer. Run the program, and it will automatically install all necessary Codecs.


2c. OggDrop Installation and Configuration.


Unizip the file, and you will find a single executable named “oggdropXPd.exe”
You can put this file anywhere, as it doesn't need a special installation folder.
You may copy it to the same folder you installed EAC, such as C:\EAC

Now run the file, and you will be presented with the following screen:




Right click anywhere on the fish, and you will get the program settings.

Select “Encoding Options”, and set all options like in the screen below.
Press “Accept” when you are done.




Right click on the fish again.
Choose “Select Output Directory”, and set all options like in the screen below.
Press “Accept” when you are done.




Right click on the fish again.
Select “Tagging, File Renaming, Playlist Creation”, and set all options like in the screen below.



Press “Accept” when you are done.

You're done with OggDrop setup!

Now we're ready to encode!

Your EAC screen should look something like this:




Notice that I have highlighted one track, track#14 - Relax (New York Mix - The Original 12")

Now I want to extract this track, and encode it to FLAC. This is just one step away!
Press in the left icon, the one that says MP3, yes!, that's right, because we want to ripp and encode the track, but we have set the internal compression to FLAC. So this button is the one that will rip the track, and then launch a command line encoder to compress the temporary WAV file to FLAC.

When you click on the button, you should see a screen similar to the following:





After the track is ripped, the command line utility pops-up and you should see something like this, on top of the previous screen:




Note: Some times this command prompt runs behind your screens, so you might not notice that it's running, but it is.
If you have selected multiple tracks for ripping, the way the program has been configured, is that it can be ripping a track, and at the same time encoding to FLACK one track at a time. So the ripping process can actually finish before the internal queue is still processing temporary WAV files in the background, converting them to FLAC.

After the track is ripped and converted, you should have a directory and a file that looks like this:



Now we simply drag the file with the mouse, and drop it into the OggDrop program icon.
When you do that, you should see this:





Once OggDrop finishes encoding to Ogg Vorbis, it will save the converted file in the same directory from where you dragged the original FLAC file. That's the way we configured OggDrop. You may change that option in the OggDrop settings.

So you're done, and your directory now looks like this:





Final notes:

The quality setting I have set on this guide is a good choice for protable players. Q=3.
Most music sounds indistinguishable from the original Audio CD at that Q value, but there are some cases where some minimal differences will be heard.
In reality, and for practical purposes, a Q of 3 is equivalent to a MP3 of about 160Kbps (or more).
But the file size is smaller than an MP3 file encoded at 128Kbps.
All ears are different, so you may want to try different Q settings on the OggDrop encoder parameters. I believe that with the current version of Ogg libraries, a Q of 5 is transparent, with 99% of music encoded. There might be some cases with some complex music, that you may require a higher value. Some people say that a value of 6 or above is 100% transparent. Do your own tests
The settings presented, will carry the ID tags into the FLAC file, and thus, into the Ogg Vorbis file. So your files, when played on WinAmp or other ID3 capable player, will recognize the information embedded into the stream, like title, song, etc.


I hope you enjoy creating your Ogg Vorbis files as much as I do.
And please, report any bugs or corrections, typos, etc, on this thread.

Thanks,
-Karl

If I may
I didn't read carefuly, but I sugest Xiph.Org's Vorbis 1.1 library version (today is 15.XI.2004), or tool (oggdrop in this case) with this build. It's contains aotuv improvements that make miracles Is this is your link build, Karl
As for quality: for audiophiles I sugest quality ~4.5 which is close to the so called "transparent" quality. It means that it is indistinctive form lossless, but it isn't lossless. But for me I also use q3 or even q2 for "lite" music

PS: Great quide

__________________
Go for SECAM =)

Hi Hydeus,

Yes, that's the current build on the OggDrop link I supplied Yes. Thanks
I use Q=3 most of the time, because it's close enough to audio transparency. Actually, with the latest vodbis lib (the one on OggDrop), even some music at a Q=0 sounds just too good to be true
It's a good value to encode for portable devices.

-kwag

If I may say so, GREAT GUIDE Karl
U da man Kal, u da man

__________________
Rui

all i can say and write is a single THANKS (with caps lock)

but i can listen all with wonderfull quality.

great quality from a great guide Kwag!

Thank you all guys
Comments are always welcome.
Suggestions for future improvements too

-kwag

complemental information!

to listen use Winamp or FooBar 2000 .

download here the Foobar Special Installer

get the in_mp4.zip and unzip inside winamp plugins folder.

thanks to muaddib for that links in another thread here!

You shouldn't normalize when creating lossless copies. If you want to have something like that, use replaygain on playback (requires foobar2000).

But then, what happens if the original is poorly mastered, and you want to, say, create a new Audio CD from your FLAC files in the future whith better "normalized" volume
If you don't normalize, then your new (reproduced) Audio CD will just sound as bad as the original. Or am I wrong

-kwag

you're right is some situations and Latexxx too .....both are right!
creedence albuns for example have low volume, megadeth have high volume,some remastered albuns like jimi hendrix and led zep are normalized(shame) and sounds high but loose dynamics. a long time Prof.Farina posted some examples in flash(to help me) when i was showing the differences of dynamics in new songs where the engineers of studio(gadget as i call),normalise "everything" to get volume but the songs loose the impact in transitions cos when is playing an instrument alone in one long sequence and came more instruments together the sound breath like dnl(old Philips noise reductionfor tapes) and loose the impact..
seems that engineers don't like to see the red line from -3db to zero db without sound... this is the exact point when the sound get volume but loose the softness!
who use sound forge(and seamless) knows what i mean cos you can listen and see where the sound start to be agressive!.....if i find this old thread with flash samples showing the differences, i post the link!
in the end is hard to choose what to do........or choose normalize to get equal volume or don't normalize and adjust the volume for each album.......hard choice but all the new albums are normalized!
for where we will run? need hard tests for each case!
personal choice? 98% to "mix" lots of differents sources!

OT
this is why lots of people think that ac3 have low volume comparing with mp2 but they forget that to encode as mp2 the sound was normalized 100% and get high volume(and some people want to encrease MORE volume-boost 1 or 2 db ), then ac3 "seems" low comparing with mp2!
same taste: normalize mp2 at 98% max!
OT out!

OT #2
normalizing oggs(98%) sounds 2 times better than normalizing mp3.
oggs get good softness and mp3 get gain but noise some trebles!
OT #2 out!

Ok, thanks jeo
I'm going to do several tests, with and without normalization.
I guess the correct way would then be to store (backup) to FLAC in it's original form, un-normalized, so it's identical to the original Audio CD.

-kwag

Hi jeo,

There is a BIG difference between "normalizing" and "compressing dynamics".
Most of new albums (and about all remasters) pass thought a Dynamic Compressor <- that's a shame
In a very rough way, a dynamic compressor will increase the volume high above the 0db and will "cut" the dynamics to avoid sound clip. This way you will have a much louder volume without distortion, but as you said, you loose dynamics, impact, punch or whatever you call it.

You just described exactly what a dynamic compressor poorly used will cause... not normalizing!
You can even normalize a stream to a lower sound level then the source.
Normalize will NOT cause any loose of dynamics. The problem with the normalization used by EAC (and many others) is the method. They will just get the highest peak level and put it near 0db, applying the same amplification to the entire stream. This is a BAD implementation if your goal is to have all the songs with the same volume. Because the peaks will vary from song to song and you can have a very soft song with one very high peak... this will spoil a multiple file normalization. If you want to normalize all the songs with the same volume, then you can't get any better than the ReplayGain analysis. Read about it here -> http://www.replaygain.org/

As Latexxx said, foobar2000 has a beautiful implementation of the replay gain analysis, and it will NOT modify the file in any way. The replay gain information will be stored in the tag information and will be used just in the playback time. The only problem with this is that foobar is the only player (that I know) that will read the replay gain tag information and use it to properly normalize the playback.

There are other methods that will edit the file and apply the replay gain directly to it. Like MP3Gain, WAVGain and OGGgain (gotta check this last one… not so sure). These programs will modify the file that will be played correctly with any player.

I think that 98% still too high normalization, and if you check a 98% normalization with sound forge, you will notice that the wave still way above the "red line" (-3db). This will let you with very little possibility for equalization. A little adjust in the frequencies will make the wav fall out the 0db making the sound clip and distort. To get the wave below the red line you will have to go for about 75%... that’s what I use in HeadAC3he for any music video. (for audio only I use replay gain )

Yes... I guess you are right, but remember that many AC3 are also dynamic compressed. I just saw a Pink Floid DVD (live performance) that has an AC3 that was terrible compressed.... that's a big shame! Dynamic compress PinkFloid should be a crime!

To end this long post (sorry for that), I want to say that I’m not against dynamic compression. I’m against just the aggressive way it has been used by sound engineers without care… and look like it is getting worst and worst every day.

Hi kwag!

Well... I have to agree with that! Original is always...original!
You can normalize or anything else after that... but keep the backup as close as possible to the original.

I guess I'll be changing that picture screen in the guide later today to point this out .
Thanks

-kwag

You can delete your own post as long as noone posted something after it.

(now it's too late because I posted this )

ok.... i found one place about normalization...quoting ursamtl:
"As I understand it, all normalizing does is take the highest peak in an audio file and compare it to a target level. Then it takes the entire file and adjusts by the same amount. For example, if the highest peak in a file is at -3dB and you normalize to -1dB, then the entire file will be increased by by +2dB. In your case, de-amplification and normalization to the same level of 80% should have no difference.

True normalization is not supposed to change the dynamic range of the file. In other words, if the highest peak is at -3dB and you normalize to -1dB, if a given point in a file is at -10dB then it will be adjusted upwards to -8dB.

The flat tops of the peaks suggest that some kind of brick wall limiting or compression was used or else the file was overnormalized and clipped somehow. If normalization is not used properly or if the program is incorrectly written then it might cause this kind of problem, but normalizing to anything less than 0dB should not be a problem as it simply raises the loudest parts of your file to this level. I was reading an article in the latest issue of Sound On Sound magazine (excellent mag by the way) that suggested normalizing to -0.5dB just to allow a margin of error. Even -1dB should probably be fine. Notice that these are not percentages, but 100% on a decent program should be equivalent to 0dB.

Using something like MP3Gain should be ok as long as none of the files ever pass 0dB.

Having written all this, it's still a good idea to avoid using too much of this kind of thing. Some professional comments I've read on the subject consider normalization a complete no-no, whereas others use it regularly. As with all processing, a little goes a long way so don't overdo it.

Regards,
Steve."
http://forum.doom9.org/showthread.ph...rel#post546013
(searching Farina link)

edited
@ muaddib
about replay gain:
"The Replay Gain proposal sets out a simple way of calculating and representing the ideal replay gain for every track and album."
this is exact what i don't like cos change the "nuances"(attack,decay,(sometimes sustain)and release) of changes in volume(impact of a new note)in the songs!

in the other side, dynamic compression is the poor solution from engineers without taste and fast solution(too bad solution of course )
normalization finds the highs peak and increases all data points by the same amount(dbs). too bad but better than dynamic compression!

Screenshot updated, "Normalize" unchecked

Thanks everyone

-kwag

all right Kwag!
i did "blind" tests.
was choosed "black sabbath-war pigs/luke's wall" for one special reason:
only in the cd, this music start with low volume.after 2minutes-50seconds the volume grow and at round 3minutes-12seconds the volume is up(don't know why and who did it in the record cos in the vinyl the volume don't change ) and i think that is one very cool sample to see differences between normalizing(old way) to 98%(smaller than 85% or greater than 99%) and without normalizations(new way)!
atention please that after 7minutes i "feel" best trebles in one of this cases.

now the "blind" samples(Q=6):
sample1
sample2

ok is too big sample but see if you can feel the differences cos this record with issues in the volume seems to me one great example to choose with or without normalize!
can you tell what is with normalize and what is without normalize?
thanks boys, the idea is remove any doubts doing blind tests,
i need your welcoming opinions please to confirm if normalizing change the general performance and the trebles in the music(round in 7 minutes playing)!

Hi Jeo,

Well, Steve is right about some things but he missed some points...

Well, that’s not what all normalizing methods do. What he described is what we can call “max peak normalization” and is exactly what I told in my previous post. This method is the most used and unfortunately the worst. There is other method called RMS (Root Mean Square) normalization. This method uses an analysis of the wave to calculate the mean volume of a track and sets this to a given value. In my opinion the RMS is better than the “Max Peak” but it still far away from ReplayGain analysis.

ReplayGain also does an analysis to calculate the mean volume of a track, but it uses a much more complex algorithm. For example it worries about the frequencies perception and response of the human ears. RMS doesn’t take that into the analysis, it uses just the absolute values from the wave, and that’s why ReplayGain results are much more accurate than RMS.

Not just “not supposed”, but true normalization will NOT change the dynamics of the file. It could happen only (and just only) if it was bad implemented and will let you normalize above 100% or 0db.

Well, if you intend to never mess with the frequencies of this file than -1dB should be fine, but if some day you equalize this track -1dB is just too little for ANY equalization. Even a small and soft equalization will make the wave fall out the 0dB and will make the sound clip and distort. I was going to say that -3dB is the minimum, but if you like to do some equalization, will play this file on a good sound system (I know you have that ) and keep the sound crystal clear... -5dB (or less) is even better.

Yes that’s right. Even if the default normalization made by MP3Gain will be about -5..-8dB, it let you set any volume you want. And if you set the desired volume too high, it will make the sound clip. MP3Gain will warn you about that, but you know... some people just don’t pay attention to warnings . It even have a setting that will prevent any normalization that can cause sound clips, but I think that is better to verify that by hand, and adjust the desired volume the way you want.

I don’t get you here... exactly what you don’t like?

Well, I have to disagree with you. ReplayGain, if used correctly, will NOT change ANY nuances (attack, decay, sustain, release or dynamics in general) in the volume or instruments variation.

ReplayGain can even make your sound better. Nowadays music industry standard seems to compress the sound and normalize it near 0dB. This is a really bad thing, and ReplayGaing will give the sound some “breath” to play without distortion.

The only thing I don't like about this "ReplayGain" thing, is that it's only supported on a handfull of players
So I think, in my opinion, that it's best to just leave the audio extraction from the CD un-normalized, and then if needed, normalize it afterwards.
This way, the FLAC files are truly identical to the ripped WAV.

-kwag