I didn't see anything obvious in the help or detailed usage regarding whether this was possible.

We're quickly coming to a point in history where the use of compression for the storage of two channel audio on hard disk is becoming unnecessary. At some point in the evolution of disk drive capacities, not only do the minor compression differences betwen different codecs and between flac -0 and flac -8 become insignificant, the difference between some compression and no compression at all becomes nearly so.

A lot of audiophiles have expressed a desire to store audio in an uncompressed format for systems that play Flac (PC based systems, SqueezeCenter & Squeezebox, etc.). Whether their arguments against compression are valid or not is a separate issue, but it would be nice if they could store uncompressed audio and still take advantage of the Flac file format and the ability to easily store metadata in Flac tags.

flac --fast not fast enough?

Perhaps an option to add lots of filler to your FLAC files? Maybe make them even larger than uncompressed?

It's not a matter of speed for those people interested in such things.

If speed is not the issue, what's left? Don't tell me audio quality?
Edit: OK let's make this quick. You tell them tagging is a lossy process and induces white noise in the mids. You tell them PCM is lossy too, and that there is no substitute to raw DSD streams. Then you tell them there is currently no offering for that kind of stuff and that they must buy the $9,999 Marantz SACD deck or whatever. And if they reply they want the convenience of tagging and using their PC as a digital library, laugh at them and call them noobs.

Seriously though. There are tagging facilities for both WAV and AIFF. There's Matroska and Ogg as well (the containers).

Yep, that's the perception. As I said, IMO that's really nothing to do with the question. Someone wants to store uncompressed audio, yet they can't currently use Flac to do so, even if the only benefit they want from Flac is the metadata storage and compatiblity with what is becoming a lossless storage standard.

To get away from any quality arguments, look at it another way: In say 5 years from now, when you have perhaps 30 TB disk drives for $150 and your entire _UN_compressed audio library is 0.5 TB, is anyone going to be spending their free time tweaking the Flac algorithms to squeeze another 0.01% compression or 5% decoding speed out of it? Absolutely not. And certainly end users aren't going to care any longer whether their files are compressed or not. It would be like going through the bother of individually zipping up a couple MB of small text files today. Nobody does that with current disk capacities.

People are making that same argument already: "I have a $120 750GB external drive on which I keep my Flac music library. My entire uncompressesd library easily fits on that one inexpensive drive. What in the world do I need compression for?"

That would be a speed issue. You just said they didn't care about speed. And if you mean they don't want to bother going through the process of converting to FLAC, well, compression or no compression, if they're going to use FLAC, the hassle will be exactly the same.

Thus, back to my question. FLAC --fast not fast enough today?

No, it's not a speed issue. They'd go through the bother of using Flac to include metadata.

Then… again, what's the problem with already available solutions (flac --fast and all the software/hardware that supports it)? FYI, I can encode a full 74-minute CD to FLAC --best in 28 seconds on my cheapo quad-core CPU, 4 seconds to FLAC --fast. That's 200 megabytes per second, twice faster than any mainstream hard drive can handle.

Just be clear about it and say they think compression (even lossless) hurts quality, that their brainwashed minds can't be reasoned with, and you have a business venture looking to make profit in their niche.

I guess I am having a hard time understanding the whole purpose of wanting uncompressed FLAC when a compressed FLAC file offers the exact same audio quality. To me, that is like getting a nice new hybrid car that is supposed to get 80MPG on the highway and then taking out that engine and replacing it with that of a 1970 unleaded truck. Every reason to own that hybrid just went right out the window.

It is true that storage space is becoming a non issue these days. I have my whole lossless library (ALAC) and two lossy libraries (both lossy library contain the same data but one is AAC and one is Lame mp3) on one 1TB external hard drive. I could have gone with an uncompressed format such as AIFF or WAV but I chose not to. Why? In the big picture, that little bit of compression that ALAC offers comes out to be a big deal. Not only that but I wanted my lossless files to have tags that would be compatible with many other programs and hardware.

I guess I just don't see the point of having a non-compressed FLAC file whenever FLAC at its fastest compression setting is nearly uncompressed and more than fast enough with encoding speed. I have seen audiophiles argue left and right that FLAC, ALAC, WavPack, etc. don't have the same sound quality as WAV/PCM/AIFF. They don't come up with any hard evidence other than what they think they hear. Tell them to get off their high horses and accept FLAC for what it is, a lossless encoder. They can use WAV or AIFF if they must go with an uncompressed format but that would be rather pointless.

Also remember that library sizes tend to increase drastically. So your situation of 5 years from now and 30TB drives is probably right. However, if my library currently takes up 500GB (it took me one year to go from 10GB to 500GB) then my library could go up to 30TB in that period of time. What you have now will always increase over time and a person might not have a hard drive dedicated to just music storage.

Either way, tell them to get off their high horses and use WAV or AIFF if they think that FLAC isn't offering lossless quality.

I find that rather optimistic. My five year old drive has a capacity of 200GB (and it wasn't even the highest capacity I could find). Today's highest capacity is 1TB (and has been for a while), with 2TB coming next year. I doubt capacities will start growing exponentially again. Even if they do, I doubt it will be on a yearly basis.

Thanks for the ideas. OggPCM looks like it may address the need. I'll have to see what kind of support exists for it.

I found what you want: use flake -t 0. With that setting (disable the use of predictors), the resulting FLAC file is slightly larger than the WAV file, and is deemed valid by the reference encoder.

Not to get too far off-topic but a notebook with a 1TB drive is coming out later this year (it will be expensive) and external USB hard drives are already in the 3TB capacity. Those 3TB drives use more than one hard drive but still, it is in a single enclosure. So a 30TB drive (whether internal or external) is very possible in the next 5 years or even with some type of hard drive configuration with multiple drives. I remember back in 1996 when 10GB was a lot but hard drive capacities increased drastically over the period of a couple of years. Back in 2002, 80GB was quite a bit and that increased to 160GB within the year.

So 30TB might be possible in 5 years but who knows how much it will cost or what type of configuration it will be. Not only that but the market shift to SSD/NAND flash memory might bring capacities back down for a little while. Who knows?

Good grief. How many CDs can a body own? A 500 gig drive can hold over 600 80 minute CDs with no compression at all for $100. What's that, less than $0.17 per disc? 30 Tb would hold 47888 HOURS of audio so if you never went to sleep and listen to all of it, it would take 5.46 years to play it all. OK the math is off some but it's still nuts. Use FLAC and get 10 years.

Last week Fry's had a Samsung 1Tb drive for $170.

I would strongly consider retiring a 5 year old drive. While it's still working its easy to copy the whole thing to a larger, new drive. If it fails, it wil be nasty to rebuild as I had to do that last year. BTW GetDataBack for NTFS works very well.

G²

The reality is that there is little to no support in current software for these tagging standards. Also there is more than one 'standard' for WAV. BTW, what's the max stream length for FLAC? Can it be used in place of RIFF WAV to break the 2 GB barrier?

Is there a "need" at all? If you do not care about a big file size, why do you consider small file size bad? What kind of advantage do you see in a big file size?

If you do not understand technical details like what sound is (physically), how it is represented in the computer, and what "compression" means, then you should not argue about them. The only difference that you can see as a layman between wav and flac is file size (but you said you don't care about that because you can afford big hard drives), and that flac may contain metadata. If you need metadata, just use flac, and leave the technical details alone. Wav and flac are just different representations of the exact same data.


More precisely, wav and flac are different representations of the exact same data. Leave out "quality" from the discussion. It doesn't even matter that they are used to store sound data to understand that wav can be converted to flac and flac can be converted back to the exact same wav file that we started with.

If encoding speed is out, what about decoding speed? Is it any easier to decode/play an uncompressed file as opposed to a lossless compressed one?

Yes, it is faster. But I can listen to flac just fine on a 233 MHz computer. And I could decode flac at ~10 MB/s on an 1.7 GHz CPU. That's not even an order of magnitude slower than the transfer rate of a fast hard drive. So why does it matter?

For portable devices: If one supports decoding flac, it should support all compression levels, so it is the slowest decoding speed that matters, anyway. Adding a compression-less mode for flac would not give us anything.

Also, FLAC uses 'assymetric' algorithm, so there is almost no decoding speed difference between --fast and --best. Furthermore, it is better to use compressed format than uncompressed on portables, because by doing so disk spinning will be minimized and battery life will be extended.

not to go too far off topic and start flaming, but do these hypothetical people you speak of also feel they need to keep their shopping bags unfolded all the time while carrying them? or do they at least feel comfortable those folding up?

Building a FLAC encoder that uses nothing but independent verbatim (uncompressed) subframes is very easy. But once disk drives get to the point that the difference between compressed and uncompressed sizes is negligible, CPU speeds will also be at the point where performing the compression (even maximum compression) will also be negligible. And since any compliant FLAC decoder can handle compressed or uncompressed equally well, there's nothing to gain by not using it.

This is never going to happen! You will always be able to fit twice the content onto any sized HD if that content is compressed to half the size. We're dealing with real, finite numbers here.

Cheers, Slipstreem.

WavPack can be used for that .

Major LOL @ skamp (flake -t 0) and boombaard (unfolded shopping bags).

Meh, why are you laughing at me? I found a solution for him!
Flake -t 0 is the audiophile codec! It makes no compromise on quality! Lows are deeper, mids are richer, highs are clearer! Woot!

Now somebody shoot me please. kthxbai

You're aware that PDF and MS word docs are compressed right? A .docx file is actually just a zip.

You're right that people won't care, mostly because using uncompressed data is stupid and everything is compressed transparently to them.

No matter how many times I read this thread (four times so far), I still can't see the logic behind NOT using lossless compression for audio, especially when the decompression of the format in question is so quick that a one-armed blind monkey could almost do it on an abacus in real-time.

I'm confusled!

Cheers, Slipstreem.

there is none.
yes

I'm with skamp, there's no point unless you believe lossless != lossless.

There is also the issue of transporting the files, for example an online store selling FLAC files will use far less bandwidth than if they sold WAV files. The customer will get their music considerably quicker too, so they will be happier.

No matter how much space people have, they will expand to fill it. I've built dedicated machines for storage space, so I have a lot of it, but I've just reached the point again where I've run out and need more.

If you don't have the original discs for your music (for example if you bought online), then you may also want to backup your collection. To do so you'll have to use a storage medium far less dense than a hard drive. If you choose the biggest currently available, a dual layer Blu-Ray disc, you can only get 50GB on each. Using FLAC would allow you to save discs and money and TIME by only requiring about half as many of them.

I don't understand the point of having uncompressed FLACs at ALL.

As pointed out by skamp, the encoding process itself is already nearly "transparent". If you're encoding while ripping, even compressing at -8 isn't a problem. Decoding is definitely not a problem even on slower computers. So there is absolutely nothing to be lost by encoding to flac.

Just because hard drives may reach 30TBs or 150TBs or 1000TBs sometime in the future, is no excuse, IMO, to needlessly waste space.

The zip file format allows the contents to be either compressed or uncompressed, but how often does anyone select for the contents to not be compressed? If I were downloading a copy of War and Peace, would I download it uncompressed because it gave me a warm and fuzzy feeling?

A few years ago I knew a few people doing this with their MP3 albums, to keep an entire album as a single file. (ZIP up a bunch of MP3 files with zero compression, rename the *.zip to *.mp3, and you can play it in any standard MP3 player... regardless of whether or not that player understands compressed archives.)

While I agree that the perceived "necessity" for this is pure rubbish, it nevertheless presents an interesting puzzle/challenge: Is there be any way to re-write the standard 44-byte WAV header such that a native FLAC transport could understand the uncompressed PCM stream?

This would allow simply re-writing the file, with no delay other than that attributable to your drive and RAM speeds....

Of course, every FLAC player I know of also plays WAV files, so it still seems like a moot point.

- M.

Obviously you and I (and about 20 outraged commenters) don't believe this, but you've no doubt read opinions from those that think there's a difference. I'm NOT trying to make their argument and I'm NOT getting into an ABX debate. It has no place here.

I'm only saying that it would be nice to accomodate those that think this. And only for the reasons I stated previously: format compatibility and tagging. I would think it would require minimal effort in changes to the encoder. I don't know whether backward capability would be an issue, though.

if they already believe lossless != lossless, all logical bets are off. you could accommodate them and they'll still tell you it sounds different because it's not .wav

but anyway, flac supports 'verbatim' (uncompressed) subframes and there are undocumented options in flac.exe to generate such files (for testing)

War is peace. Freedom is slavery. Ignorance is strength.

Those who say that there is (even theoretical) difference between a losslessly compressed and uncompressed audio (or any other file) are absolute ignorants. Everyone with IQ at lease 80 must understand that there is no difference between a compressed flac and uncompressed PCM.

There's no need to use uncompressed data. Losslessly compressed data are used everywhere - RAW picture formats, PNG, PDF, ODF, OOXML etc. Using uncompressed data brings only drawbacks, no advantages at all.

So once, if you want uncompressed flac, please answer this question: What advantages will an uncompressed flac bring over a compressed one?

There is no answer to this question. I think this useless thread should be closed...

Please, do tell, just for fun, and so that JJZolx can have his solution.

Why would a person want to go out of their way to cater for a tiny and stupid minority who refuse to believe that lossless audio compression IS lossless? It defies logic! Even going to lossy encoding, give these people a LAME VBR MP3 at -V3 (~175Kbps) in a blind ABX test and the vast majority wouldn't be able to tell any difference.

Cheers, Slipstreem.

Slipstreem: money. Why make $499 cables when $9.99 will do just fine? Because people buy them, and a lot of profit is to be made.

That would be encouraging ignorance and backing unfounded myths. That is dangerous and IMO even unethical.

I for one can sort of see the allure in such a thing as uncompressed audio in a FLAC 'container', but not for music. It could be an interesting thing to allow tidy organization of my effects library without the need of expensive proprietary software for databasing BWFs and writing BWF metadata and the various application incompatibilities that poses. Right now, I use standard WAVs for my library and use only the filename to describe the type of file (for instance, "ANML bear grunt short sweetened01.wav"). It'd be nice to be able to define custom tags that virtually any application could read and interpret without having to be extended specifically for that. I'd be using Mp3tag and foobar to organize my effects library, and that sounds pretty damn good to me.

For music, I truly don't see the point, as a music file generally does not need to fully decompress to be able to be utilized (to play). For files that need to be fully decompressed when opened by an audio editor or DAW, the decompression phase is an additional and unwanted annoyance in addition to peak file/waveform calculation, so uncompressed files are "king" in terms of minimizing wait time, which can be significant enough as it is.

Still, this isn't exactly the 'right path' if you were to ask me.

What bothers me about threads like these is that the OP just opts out of the thread, either feeling that the answers are ignorant (remember, he is an audiophile) or that he opened up a can of worms that he wishes he hadnt.

I agree completely. Creating a "no compression" version of flac would add credence to the false claim that "compression == bad" (even when the compression is lossless).

It might've been worth considering if there were some other applications of this uncompressed flac. For example, if some portable was unable to decode even -0 compressed flacs due to the "complexity". Fortunately, this is not the case. So, I don't see why any effort should be expended to create this useless behavior.

So, if there is some genuine application of an uncompressed flac, other than pleasing people who believe in fairy tales, then maybe the feature should be added. Otherwise, we should let those people live with the consequences of their irrational beliefs.

Josh already answered the question. That's all I need. The rest is niether ignorant nor a can of worms.

If storage is going to make that sort of leap, imagine the leap processing power is going to make*.
Imagine how unnoticeable FLAC --fast will be then.


*at least with x86 hardware this is likely possible considering that, with Nehalem, Intel is replacing the longtime FSB bottleneck with QuickPath.

Surely pandering to an unproven perception that losslessly encoded audio data is in some way perceptibly different from uncompressed WAV data is in and of itself dangerous?

WAV data, after all, has to be "converted" before our ears can make any use of it....

flac decoding is so fast that decompressing a smaller flac file could be faster than reading the wave file due to extra i/o.

An archive with no compression is very useful. You can use it to quickly group many small files together for transmission over high latency mediums (like FTP whch is also prone to break down every n connection attempts), and to store verbatim files that are already compressed: such as FLAC, MP3, JPEG. Zip with store method is great for photo albums.

Why did the topic of this discussion turned to audiophiles?

I missed that post.

Boo to spreading FUD to sell products.

It is our duty to inform the world that lossless == lossless, not the other way around.