Hi, I've came across to this flac command line:

-8 -A tukey(0.25) -A gauss(0.1875) -b 4096 -V -T "artist=%a" -T "title=%t" -T "album=%g" -T "date=%y" -T "tracknumber=%n" -T "genre=%m" %s --sector-align

What's the bold part doing? I've read the flac --explain help, but I don't really understand what it means.

Thanks

It's a way of doubling your encode time in order to get 0.2% better compression.

Yes, I'm exaggerating but probably not by all that much.

Well, can this be improved? I mean, can I get a little bit smaller file size without any change on the audio data?
I have a decent computer and I'd love to have lossless audio in less space, even if it means it will take 30x times to encode, it's a onetime thing anyway.

Those are the only switches I can use?

Thanks

The resulting file will be "smaller", but the question is was it worth the time and effort. If you're mainly listening on PC, you might want to check out other codecs like TAK and Wavpack which offer better compression rates without having to resort to such "brute force methods".

In that case FLAC probably isn't the format for you. If you look here:
http://www.synthetic-soul.co.uk/comparison...sion&Desc=0

You'll see that FLAC is beaten by other formats. The main advantage of FLAC is it's rapid decoding time.

If you wanted maximum compression then I'd probably go for TAK. I personally want a good all-rounder (and more mature) so I use WavPack whenever I rip to lossless.

And it's extremely wide support in both hardware and software.

Other than that, you are completely correct.

Thats true, I forgot about that...

That's exactly the reason I use FLAC, I'll be getting an Squeezebox Duet and a few giant speakers quite soon and I'd like to have lossless audio playing on them. FLAC is the only "real" solution in such case, cause, well, wave files are just yo darn big.

What switches should I use o get the smallest file possible using FLAC?

Thanks

What switches? Maybe some of those hard-to-remember ones.

Me, I'm sticking with -V8 or -V --best. Hey, it's --best.

I think it should support wavpack:

http://www.hydrogenaudio.org/forums/index....showtopic=58073

I'm checking out wavpack stats right now, and for what I'm reading, it's quite slow when decoding compared to FLAC, but still, I'll read about it.

By "switches", I meant what should I add to my -A command line options to get smaller flac files. Cause I've seen in the documentation that there are a lot of possible functions, but I've no clue of what they do, or ir they can be used effectively one with the other. In the worst case scenario, disk space ain't such big problem anyway, as part of my "audio equipment renovation project", I'm getting 2 1tb hard drives. But, still, I'd like to use the space properly.

If WavPack decoding is fast enough for playback why do you care that it's slower than flac?

Something about "apodization" you can read here: http://mathworld.wolfram.com/ApodizationFunction.html

flac, wav, and I think maybe apple lossless are the only lossless codecs the squeezebox can decode natively (on the box itself), the rest have to be transcoded on the server machine to another format. so to do wavpack->flac you need a wavpack decoder piped to flac endoder running realtime, which may be a problem on a NAS. transcoded formats cannot use some features like ffwd/rew.

the windowing functions are hit-and-miss, some work well with certain types of music, or sometimes tweaking the parameter itself makes a difference between genres. I did a lot of experimentation to come up with the defaults but if you really want to add more -A options you'll have to experiment with your own collection.

I think there's a thread here somewhere where I posted window results for different genres.

I tend to use this command to encode to FLAC:

-V -8 -T "artist=%a" -T "title=%t" -T "album=%g" -T "date=%y" -T "tracknumber=%n" -T "genre=%m" %s

I've decided not to use apodization functions. Its not worth all the trouble, the size difference is barely appealing.

Anyway, one last thing. Do you think I should keep a copy of the wav files also? or an image of the cd with a cue sheet?. I find it pointless but, somehow useful. What do you guys think?

Thanks

Though not the bolded part of the command line, I think it's also worth mentioning that the last command, "--sector-align" is pointless when encoding files extracted from a CD. The FLAC documentation says as much: "This option will have no effect if the files are already aligned (as is the normally the case with WAVE files ripped from a CD)."

No reason to keep a copy of wav files or image/cue if you plan to play your flacs on Squeezebox. But if you want some kind of backup of your audio discs, then you might want to encode and keep somewhere that image/cue files.

Yeah, I'll make images of all the cds and I'll burn them to new cds, just in case.

Thanks for the advice

Don't forget that the image file takes up more space on a data CD than the audio takes up on an audio CD. In some cases there might not be enough room.