Hello,

First it was CDA --> Mp3, then Mp3 --> Mp3Pro, then i started listening to Grateful Dead live concerts and i switched to Ape. Then i got crazy with EAC, spending way too much time to make sure i got correct extractions. Then i had everything setup with Ape, i used Normal Compression, and I used Media Jukebox to sort all of the tags.

Everything was pleasant...

Then i started doing research on .Flac, i decided to do a little format switching. Particuarly the support of Regain in Winamp won me over.

However, i'm tired of testing songs, and figuring out settings and whatnot. So i just want to post my settings, and be enlightened if my settings are correct for what i want to achieve. Which primarily is to store the songs in the format (now being .flac) as small as possible, play them in winamp without my CPU going "AHHHH", and when i return them to .wav, they are exactly the same (no added 0's or crazy other data to mess up the original .wav)

So i have FLAC frontend v1.7.1 Etree edition, i have .wav files extracted with EAC (all correct offsets used), and these are my settings in FLAC
Level 8 Compression(I am under the impression that max compression doesn't take up much CPU time)
Verify Enabled (just to be sure, lol)
Add Tags Enabled (Cause I got alot of music on here...)
Replaygain with Treat input files as one album (This is the main reason i switched to FLAC, and i am under the impression that the .wav files made from .flac --> .wav will not have the gain applied to them, hence keeping them the same as the original)
Align on sector boundaries Disabled (i have this off but i am not sure of it use, since i am using original .wav files, i believe it is not nessessary to use this, am i correct?)
Cuesheet Disabled (already have it from EAC)
Delete input files Disabled (I do my own housekeeping)
Ogg-FLAC and Dec. through errors Disabled (Not nessesary in my opinion)

I thank anyone and everyone in advance who will help enlighten me in this matter. Sorry for the long post but i'd rather be precise in my settings than have a miscomunication.

Thanks!

You will find that it is 1000% more timeconsuming and gives you 10% smaller files than level 0.



I don't trust flac to verify itself... which IMHO is a good rule of thumb.



What's "Add tags"?



Well, or better yet just set EAC to ripping the CD as one wav-file. That way you can be certain to get the correct gaps included.



I'm not sure what you mean, but above should eliminate the need for any alignment.



Yup, and it isn't really supported anyways.



Word.



Don't trust ogg-flac after what has been brought to light about it on this forum.

For all that, I wonder how good is the equipment you use to listen to all your wonderfully encoded lossless files.

Actually not that great... my headphones are not very great, and my stereo is the best i could get at a college guys pay, but my car IMHO has a excellent sound system. I just want to store these files as best i can so later in life hopefully i can get a great sound system and enjoy them at their fullest.



Well, time consuming isn't a important factor, i'll compress them overnight if need be, the smaller the better since i plan to compress alot of music. What i'm more concerned about is the processor useage during playback. I've tried to test it with Winamp 2x, but i can't get any results besides watching the usage in Windows Task Manager. Their did not seem to be any difference in processing between 0 and 8 compression. Does anyone know if their is a difference?



Well, i started trying to verify after compression instead of during, but then realized this is the same thing as having it check itself! lol. Does anyone have a way to verify outside of Flac? (besides just uncompressing it back to .wav and comparing, lol)



Add tags adds tags to the flac files after compression. I think i have this one down, the only thing it can't add automaticaly is Year and Genre.



Well... i think you missunderstood my question...

Replay Gain has nothing to do with gaps. It allows during playback to have the volume normalized for the track or for the whole album. For example each track of Pink Floyd's Animals has a different Gain setting, but the whole album has a gain of -1.86.

I was curious if when you decompress them to .wav if the gain would come back to it.

On the topic of sector boundries, i don't think i need this since i extracted the .wav's from the CD. But does anyone have a input on this?

And doing your own housekeeping is the bomb! "word" lol

Thanks again, particuarly DrDoogie

BTW, i was reading on the forums of a "Better" Winamp 2x input plug-in, can anyone tell me more information about that? Also do i have the latest version of FLAC? i have v1.7.1 Etree edition

The newest version of FLAC is available from http://flac.sf.net/

Using compression levels greater than 6 generally results in very slow encodes for very little gain. I really don't think that the tradeoff is worth it. I don't believe that decompression is significantly more CPU-intensive for higher compression levels, being part of the design and all.

ReplayGain will just add the tags to the files, not modify the flac stream itself, so you can decode without the gain information to get WAVs identical to your rips.

Ogg-FLAC is really neither here nor there. I don't use it myself, but the raw FLAC stream could be extracted from the ogg easily.

Align on sector boundaries wouldn't be necessary. Your rips should already be anyway.

Decode through errors doesn't matter, because you're not decoding.

After through examination, i see that their is really no difference in decompression playback CPU usage as well. I will do some test with the compression levels to determine which is optimal, cause 8 does take longgggg time. lol. The flac site recommends 5, i think i'll give that a whirl.

Thank you for the rest of the information Ardax.

Thanks!

I encode all my flac files at level 8. It is true that it's only a very small difference, but it is still smaller files...I don't really care how long it takes to encode since I only have to do it once. Since flac produces files larger than most other lossless compressors, every little bit counts.

My Lossless bible

FLAC level 8 should take no more than a second longer to decode than FLAC level 0, with a gain of ~3.5% more compression

Adding a gain tag to the tracks doesn't modify the bitstream, just makes it louder or quieter when being played (not decoded to a wav file) as far as I know

Excellent, EXCELLENT chart!

An equivalent Lossy (Standard profiles) thing would be almost as interesting (although not really as useful)

Cheers! B)

Certainly does seem to put FLAC in a very good light, thanks for the graphs, Mac. These 2 almost just clinched my decision on the lossless codec I would use if I ever went lossless, until processor clockspeeds become fast enough to make that really irrelevant.

I'd have thought it puts either APE or LA in a good light, depending on what you want

Well, one of my prerequisites is that the codec I choose must not take up more than 1% of CPU power as shown by the Task Manager. Heheh. Does LA do that with an Athlon XP 2000+ yet?
I'm not discounting the superior compression ratio of LA as compared to the other codecs though, it becomes significant with larger digital music collections.

the very generic way would be to generate a CRC32, md5 or similar hash on the wav.
Then you can uncompress and do the same on the resulting wave, it they're the same you can be relatively sure the files are bit-identical (md5 is is a lot more sure than CRC32 but less widespread and slower).
If you keep the CRC32 files (often extension .sfv) you can do the check anytime later (in 10 years ie), or you can generate a hash on the compressed file and later check if it has not been damaged. This is used in RAR and probably other achive formats to detect damage, and i do it for all my video (im a lot more picky about video than audio, since it takes me 30 min to encode a audio CD but 3 days to encode a DVD :)

FAQ: How can I be sure FLAC is lossless?
If you study how verify works it should put you at ease.



That doesn't always work like you expect.
FAQ: I compressed a WAVE file to FLAC, then decompressed to WAVE, and the two weren't identical. Why?

Josh

Huh? It's true that decompression times don't vary significantly, but compression times sure do. So level 8 vs. level 5 takes 6x as long to encode for maybe a 0.5% gain in compression.

And yes, as always, the raw numbers comparing codecs don't mean a whole lot unless you know what you're using it for - archiving, sharing cross platform, etc., and what your own tradeoffs are.

Verification - won't shntool do a checksum on just the music (data, no headers, chunks, etc.) part of a wav or a flac? If so, that would allow a comparison / verification. Vague recollection there ... It should at least solve the "I encoded a wav and decoded and got a different file" problem.

Not that it's necessary, but this being HA and all, people like to do that stuff.