The differences between FLAC and MP3, Lossless vs Lossy

kooldeep

Member
Joined
Apr 19, 2008
Messages
47
Points
8
Location
Haryana
The differences between FLAC and MP3, Lossless vs Lossy

The time has come where you can get a 500GB or 1TB hard drive for pretty cheap. Space is no longer a issue for most people. Alot of people are wondering what music format they should use, and what's the difference in quality between each format is.

Lossy Format
This means when you rip a CD, you lose audio data and quality. The most common lossy file is MP3.

Lossless Format
This means when you rip a CD, it is an exact audio copy and you lose no quality. The most common type of lossless files are FLAC and WMA Lossless.

Will you be able to hear a difference between lossy and lossless?
Depending on what you use to play your music, what your speakers/headphones are, and how good your ears are, you may or may not notice a difference between them. Some people, however, cannot hear the difference between them even on the highest end of hardware.

Does my player support lossless?
Some portable music players (iPods, Zunes, etc) will not play certain lossless formats. You should check to see if your player can handle it, and what lossless format your player likes best. Foobar and Winamp should be able to play all type of lossless formats, however WMP11 will only play WMA Lossless.

Lossless is more future proof than Lossy
The best thing about the lossless format is that you can convert FLAC to WMA Lossless and back to FLAC as many times as you want without losing any quality. The same goes for any lossless format. With lossy format (MP3) each time you convert you lose quality and cannot get it back.
 
Hi,

Kooldeep very good thread, and the overall explanation is nice,Only one clarification or rather addition, If you check all over the net the next format to FLAC is APE (MONKEYS AUDIO) not WMA(I've very rarely seen one) so thats the one in famous list !!

Though this debate of Lossy Vs Lossless will never end ( some of the Audiophile's just won't accept the FLAC quality same their Audiophile CD ) for obvious reasons including some that you've mentioned !!

Regards.
 
I don't have that golden ears to judge the difference between flac & CD :(, so I'm pretty sure that I don't include in that class :)

Audiophile's just won't accept the FLAC quality same their Audiophile CD ) for obvious reasons including some that you've mentioned !!

Regards.
 
I don't have that golden ears to judge the difference between flac & CD :(, so I'm pretty sure that I don't include in that class :)

dont worry, mate. you can't tell because there is NO difference. FLAC and CD recordings are bit-perfect.

FLAC is analogous to using ZIP to compress files. there is no loss in data. only better compression.

the best thing about FLAC, IMHO, is that using the cue sheet, one can obtain perfect replicas of the ripped CD. Also, for portable use, copies of the FLAC file can be transcoded to your lossy format of choice. Personally i ditched the physical CD's a long time ago. I'm slowly replacing all my 320kbps mp3's with fresh FLAC copies.
 
hi guys
not too savvy with the intricacies of formats etc - my knowledge of computers is limited to ms office and a web browser

i dont understand that mp3s lose quality and its pretty clear to hear as well. when i rip cd's they are saved in wma format. not too familiar with flac etc. is there a better format for me to save it in other that wma and if so how do i go about it
 
Can people here actually tell a FLAC file from a wav? With all due respect, this is pure placebo.

Try this:

  1. Rip a CD to wave using your favourite CD ripper (try ExactAudioCopy or DbPowerAmp).
  2. Encode this into FLAC.
  3. Decode the FLAC to a .wav file.
  4. If you are on Linux, do a 'diff' on the two .wav files-the one that went into the encoder and the other that the decoder generated. If you are on Windows, use an MD5 hash checker. Either method will tell you the the two .wavs are bit-identical.

So how can people tell the difference between a FLAC and a .wav file? It is mathematically impossible for differences to exist between the two. To go one step further, you can perform a blind ABX test and judge for yourself.

If you can indeed distinguish a FLAC from a .wav on the same setup, then there is something wrong with the decoder. This, however is highly improbable.

If someone wants to see this, I can provide screenshots of the above process, but I recommend that you do this yourself.
 
not really sure what a wav or a flac file is but i call definitely tell the loss in quality in an mp3 file. highs becoming sketchy, thinner mids and bass and a lot of listening fatigue while listening
 
Sure, afj!

I too can tell differences between MP3s and the original CDs if the MP3s are not ripped properly. Since MP3 is a lossy format, I completely agree that people can differentiate between MP3s and CDs.

However, FLAC is a lossless format. This means that information is not lost when you rip your CDs into FLAC. You can technically re-construct your original CD from a bunch of FLAC files (I have done this on numerous occasions). Hence, I disagree with people who believe that FLACs sound different compared to their CDs. All this is assuming the CD was ripped properly in the first place, and that the FLAC decoder (software) is not broken.
 
Can people here actually tell a FLAC file from a wav? With all due respect, this is pure placebo.

Try this:

  1. Rip a CD to wave using your favourite CD ripper (try ExactAudioCopy or DbPowerAmp).
  2. Encode this into FLAC.
  3. Decode the FLAC to a .wav file.
  4. If you are on Linux, do a 'diff' on the two .wav files-the one that went into the encoder and the other that the decoder generated. If you are on Windows, use an MD5 hash checker. Either method will tell you the the two .wavs are bit-identical.

So how can people tell the difference between a FLAC and a .wav file? It is mathematically impossible for differences to exist between the two. To go one step further, you can perform a blind ABX test and judge for yourself.

If you can indeed distinguish a FLAC from a .wav on the same setup, then there is something wrong with the decoder. This, however is highly improbable.

If someone wants to see this, I can provide screenshots of the above process, but I recommend that you do this yourself.

UPDATE: While making the above statements, I assumed that the CD was ripped into FLAC properly using good software. If the CD ripper that you use somehow alters the audio before encoding into FLAC (for example, normalization), then it is definitely possible to differentiate between a FLAC and the original CD (or the .wav file). Hence, please use good CD ripping software such as ExactAudioCopy or DbPowerAmp to rip your CDs.
 
Hehe, while we are on the topic of John Atkinson and his article on Lossless vs Lossy, I came across this rather interesting discussion on the artile on Hydrogenaudio

Some of the comments seriously made me laugh out loud

They used JPEG on the spectrum charts. That about sums it up what they know about data compression.

and one that actually is pretty hard hitting:;

How disappointing - that the "golden ears" rely purely on measurement to make a case of audibility! The opposite error of "I can hear it but can't measure a difference," now we have "I can measure something but won't bother to check if I can hear the difference." How about measuring some spectra of cables while we're here.... Meter reading when its convenient and listening only test when it's convenient to a predetermined outcome. C'mon. I'm not against subjective evaluation per se, but this is weak.

Of all things audio, codecs are some of the simplest to do a listening test on (ABX, blind, etc.).

Stereophile Article: MP3 vs AAC vs FLAC vs CD - Hydrogenaudio Forums


Coming back to the original topic, I and a few friends did a serious ABX listening test over a few drinks by choosing some well mastered recordings (which all of us were also very familiar with)
- Private Investigations - Dire Straits
- Spanish Harlem - Rebeca Pidgeon
- Come Away with me - Norah Jones

We selected original pressings, and ripped them into four formats
a) untouched wav
b) FLAC
c) Lame (arguably the best mp3 codec) encoded to 320 kbps
d) Lame encoded to 192kbps
e) Lame encoded to 128kbps

We did what is possibly the most scientific ABX test, using an application called Foobar that allows for real time ABX testing
(You select two files for the same song, say 1 FLAC, the other 320kbps mp3
and you can switch between the two formats real time and at the end of it, you need to choose if A is X or Y and vice versa for B)

I can post the exact findings (the results are archived somewhere on my laptop) if anyone is interested, but here is what I roughtly remember

Between FLAC and 128 kbps MP3, all of us could spot the diff 10/10 times (99.9% degree of accuracy)

Flac and 192, all of us except 1 could tell the diff most of the time but it took very very careful listening ( I think i got the diff right 8/10 times)

Flac and 320, we failed miserably..
I got it right 4/10, another guy got it right 6/10 and in both cases which is basically the same as guessing
 
We did what is possibly the most scientific ABX test, using an application called Foobar that allows for real time ABX testing
(You select two files for the same song, say 1 FLAC, the other 320kbps mp3
and you can switch between the two formats real time and at the end of it, you need to choose if A is X or Y and vice versa for B)

I can post the exact findings (the results are archived somewhere on my laptop) if anyone is interested, but here is what I roughtly remember

Between FLAC and 128 kbps MP3, all of us could spot the diff 10/10 times (99.9% degree of accuracy)

Flac and 192, all of us except 1 could tell the diff most of the time but it took very very careful listening ( I think i got the diff right 8/10 times)

Flac and 320, we failed miserably..
I got it right 4/10, another guy got it right 6/10 and in both cases which is basically the same as guessing

Cool! Now we are talking. :p

This is exactly what I have done many times over, and my personal preference is to use VBR. The exact LAME 3.98.2 switches are "--no-replaygain -V3". For my not so "golden" ears, the resulting MP3s are not distinguished from the reference FLACs. I am not too worried about this as I have FLAC backups of all my CDs and can transcode to any lossy format of my choice any time I want.

In my opinion, Foobar2000-the same program that you mention, is the best audio playback software. I have never felt the need for anything else.

On a more objective note, it has been concluded many times at the HydrogenAudio.org forums, that LAME's real strength lies in VBR encoding - not CBR. The settings I mention above are known to produce VBR MP3s that sound identical to the source. Give it a shot, and tell us what you think.
 
^^ Thats what I use
Peak quality VBR at -v2
my ears can't hear a difference between these and the original pressing at what can be termed as mid-end audio reproduction equipment

(Mac Mini - M Audio Audiophile USB - Marantz PM-17 - Phase Tec V12/KEF iq5)

Perhaps some people have better ears or maybe more revealing equipment can potentially reveal a difference, who knows?

But on a side note, I found it really funny that Stereophile used:

a) Jpeg's to display the graphs
b) Oscilloscoes to measure the difference instead of using their ears (though they conveniently ignore osc. tests when measuring cable differences claiming ears can go beyond judging what equipment can)
 
Nice. I used -V2 myself all these years until just last week when I couldn't distinguish LAME 3.98.2 -V3 encoded files from their respective sources.
I therefore switched to -V2 and the savings in space have been astounding.
 
I second that.

Thanks Sandeep. But I am afraid that you might have misunderstood my statements.

EAC is far superior to FLAC. Also, try monkey lossless APE codec (with separate cue) files for maximum resolution.

FLAC is a lossless format too just like APE (Monkey's Audio) is. EAC on the other hand is a ripper that can convert your CDs to any format you want using a codec - be it MP3, FLAC or APE. Hence, EAC is a good ripper while FLAC is a good lossless format. I also don't understand what you mean by "Resolution". I think you intended to say "Compression".

APE may offer better compression, but FLAC enjoys a lot of advantages like:
  • Broad hardware support (there are players that can play FLAC files directly).
  • Excellent metadata system (you can embed any info you want into the FLAC file including CUESheets, album covers, etc) which means that you don't need to have FLAC files and CUE files and album cover JPG files - just embed all the info into the FLAC and get rid of the other files if you want to.
  • Fast decoding speeds - which means that any new hardware players can easily implement the FLAC playback feature.
  • Open Source - no DRM, no threat of patent infringments, no nonsense.

The list can go on, but these are the main features. I personally love FLAC for all the above reasons and have EAC and FLAC to backup all my CDs. This means that all my CDs have been played back exactly once and have been stored away carefully. I now listen to my FLAC files and convert them to LAME MP3s for usage in my car, on my iPod, on my wife's cell phone, etc.

For all recordings purposes, try the DSD format (unfortunately proprietary from SONY), or ELAC.

I don't know much about this. However, I would avoid all proprietary formats - especailly ANYTHING from Sony. There are plenty of open-source alternatives to choose from.

PS: any rip is a rip off

I completely disagree. I rip my CDs in order to make backup copies. I keep the original CDs away. I use the lossless FLAC files to burn, transcode and play them on a plethora of devices. If anything, ripping your music to a digital format saves money as you don't have to buy a new copy of a disk just because it was scratched due to constant handling.
 
hdgopala: know how to get itunes to recognise/ play FLAC?

If not, how to put FLAC into an OGG container? Apparently iTunes can now play this.
 
Apple Lossless

Where does apple lossles stand?
I am planning to use I pod video as mini server with initially giving output through headphones - --headphones - 2 RCA's

Hoe bad or how good it will be?

Hemant
 
A beautiful, well-constructed speaker with class-leading soundstage, imaging and bass that is fast, deep, and precise.
Back
Top