1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.

    Dismiss Notice

Poll: Audible Difference between FLAC and 320kbps MP3?

Discussion in 'Sound Science' started by el_doug, Aug 10, 2009.
First
 
Back
1 2 3 4 5 6 7 8
10 11 12 13 14 15 16 17
Next
 
Last
  1. ninjikiran
    I am not an audiophile but most of the time FLAC or any other lossless audio ripped directly or made as a lossless before being printed to CD sounds better.

    Usually it is the case imo, but there are exceptions. MP3's encoded properly can reach the point where the differences are barely noticeable. Problem is most mp3's are not encoded as such. When I was an encoder for fansubs(a long long time ago) groups used similar codecs (DivX 3.1 for instance) but the difference between a good encode and a poor encode were vast even if they were a similar size.

    Some recordings sound bad regardless of format though...
     
  2. PiSkyHiFi
    I feel these are separate goals:

    1. Achieve highest signal integrity.
    2. Achieve highest transducer integrity.
    3. Achieve most faithful/enjoyable musical experience.

    I think the problem with matching #1 and #3 is not just in bitrates, but also in the questions of how faithful is the signal to the original sound environment anyway - the purest signal is only representative of a single point in the room for each mic - this itself maybe a futile goal for truly achieving #3.
     
  3. TheRH
    I will have to admit that if you are not heavily involved in the music that you can not tell the difference between the two. But when it comes down to it I can tell the difference.
     
  4. PiSkyHiFi
    Quote:

    Originally Posted by TheRH /img/forum/go_quote.gif
    I will have to admit that if you are not heavily involved in the music that you can not tell the difference between the two. But when it comes down to it I can tell the difference.



    This is it. This is exactly why I posted my article about the physics of sound and the math of signals.

    I believe one may be heavily involved in linear accurate sound, where you can spot the difference between FLAC and high rate Mp3, but this is not necessarily any more or less accurate a reflection of the music.

    The problem is that its possible to be into the music enough to not be concerned about precise spatial sound positioning and more concerned with timbre, dynamics and full-spectrum frequency response. This could mean that someone heavily involved in the music may benefit from decoupling their ears from the forced locality of an ultra linear source and it may feel more like a live acoustic recording they have experienced where there was a lack of this precise spatial locality.

    Add in the uncertainty principle and ultimate linearity becomes an unachievable goal that is further away from a real sound space.

    I hope that makes sense to you, it is not easy to find the right words to express this - I get frustrated by those trying to knock high rate Mp3s when I feel it is only in very specific recording spaces that the difference is important one way or the other, I think for 90% of good quality recordings the differences are irrelevant and accurate signal and transducer integrity becomes the only factor in getting close to the original sound-space. In terms of spatial sense, most recordings are artificially linear and of those that are not, only a few require things to be stationary. For artificially created sound-spaces, if it is a requirement that everything stay perfectly still, FLAC is good, otherwise MP3 can give the fake environment a touch of uncertainty which many ears will find more pleasing.

    Mathematically, there is enough detail in a high rate Mp3 to achieve the same level of timbre, dynamics and full-spectrum frequency response as the FLAC to the same degree of uncertainty as a room filled with moving objects.

    Corley.
     
  5. TheRH
    Corley,

    I do not think I could have said it better. As an audiophile in the Houston Audio Society we do critical lsiting for certain details. Although for the most part just as with the 320 vs FLAC the difference is almost inaudible with regular listing. And when it comes down to it, is about enjoying the music not pin pointing every detail.
     
  6. analogsonar
    320 sounds good but a large part of the file is missing. ? It might depend on the equipment. I have A/B'd a few files using fubar, most recently Tool's undertow 320 and flac, the difference is mostly in space and air, more punchy, etc. It could also depend on the recording, alot of new stuff is recorded onto hard drives so that could reduce dynamic range or the difference may be more obvious on speakers than headphones.
     
  7. PiSkyHiFi
    Quote:

    Originally Posted by TheRH /img/forum/go_quote.gif
    Corley,

    I do not think I could have said it better. As an audiophile in the Houston Audio Society we do critical lsiting for certain details. Although for the most part just as with the 320 vs FLAC the difference is almost inaudible with regular listing. And when it comes down to it, is about enjoying the music not pin pointing every detail.




    You did go further than to state you can tell the difference between the 2, you implied that only those "heavily into the music" can do so.

    For many, being into sound can at times, go against being into music.

    I see it all too often in this thread, people making difficult to prove assertions and combining it with ego, I think much of it is done precisely because it can be difficult to prove especially with regards to what individuals hear. When I read these, I see this connection and this person loses a little credibility in my mind.

    I'm sorry if I am being blunt, I only hope that with your statement above, you are about enjoying the music not pin pointing every detail.

    The biggest problem I think we may agree on is that with great music and/or sound, words really fail to express the emotions.

    To quote Frank:

    Quote:

    "Talking about music is like dancing about architecture."



    Corley.
     
  8. TheRH
    Exactly, it is all about the music. When I am on my computer or listing to my stereo I do FLAC, main reason is hard drive space is almost irrelevant these days. So why not rip in FLAC. Now when I am on the go ( Sony Walkman) I want to preserve as much detail as possible now that I have some real IEM's ( Klispch S4's) and I can tell the difference in lower bit rates. And to address what you said, it is not blunt but rather the truth. 95% of the time it is about listing to and immersing yourself in the music and enjoying whatever it is that you have on.
     
  9. Yngwie
    I've never noticed a difference. But my only headphones are Sennheiser HD280's, so I'm not really much of an audiophile.
     
  10. PiSkyHiFi
    Quote:

    Originally Posted by TheRH /img/forum/go_quote.gif
    Exactly, it is all about the music. When I am on my computer or listing to my stereo I do FLAC, main reason is hard drive space is almost irrelevant these days. So why not rip in FLAC. Now when I am on the go ( Sony Walkman) I want to preserve as much detail as possible now that I have some real IEM's ( Klispch S4's) and I can tell the difference in lower bit rates. And to address what you said, it is not blunt but rather the truth. 95% of the time it is about listing to and immersing yourself in the music and enjoying whatever it is that you have on.



    That's all well and good, but it does sidetrack my point a little - my point being that the extra detail you store in FLAC may be sonically equivalent to a completely different dataset coming from a mic sitting 1 foot away from the original mic's location. Hence the extra detail you get puts you in exactly 1 spot in the room.

    In fact, the 2 datasets may look completely different in the time domain and remain sonically pretty spot on, only the frequency domain interpretation would show you how similar they really are. Without compression, the frequency domain interpretation is lossless.

    Adding more precision to the dataset by using a higher bandwidth ADC process will increase the high frequency cut-off point further beyond what we can hear and otherwise just localise the sound more, which in most cases can not be faithfully reproduced as the same location, just reproduced as some uncertain location very precisely.

    The extra precision in the frequency domain will affect the faithful reproduction of relative volume of different frequencies. By compressing this dataset slightly, you may lose consistency in both frequency response and phase locality. When I say frequency response, I mean its all there, just a calculable error in relative volume of different frequencies. The spatial locality loss is due to the loss of phase information, putting you in an uncertain location in the original soundspace.

    The loss of some "punch" in MP3 is probably due to the inability to accurately recreate the onset of sound from exactly 1 point in space.

    Its miniscule in high rate MP3 and although detectable with good equipment, the question of whether this is a less faithful reproduction of the original sound-space depends on other factors like how the sound-space is intended to be reproduced with respect to locality using either speakers or headphones.

    In the case of "punch" the original sound-space may have reflective surfaces that will color any impact and you may actually experience a greater sense of impact if their respective locations are uncertain - i.e. the error in position may create slightly less or more of an impact for different sounds.

    All done now. Back to Ry Cooder.
     
  11. Bilavideo
    Quote:

    Originally Posted by krmathis /img/forum/go_quote.gif
    I most probably can not tell a difference most of the time.
    Of course depending on the source material and the MP3 encoder used.

    But for sake of mind I use lossless, to not need to worry about if I loose out of something or not. [​IMG]




    Bingo! We have a winner!

    Whenever I have a choice, I go with the highest bitrate I can. If lossless is available, that's what I do. But when it comes to the hearing part, I have never heard any difference between the two. I can hear it when the movie reel is out of balanced and the title music sounds like an LP on a badly-balanced turntable. I can hear it when the headphone cups are reversed. I can hear it when sounds drop out in a recording that was layered one track too many. I just don't hear a difference between 128k and 1 meg.
     
  12. DaveBSC
    It depends on the production quality of the source. FLAC on a terrible sounding record like "Death Magnetic" is pretty much pointless, it's already been ruined in the mixing and mastering stage and there's no going back.

    If the production is at least half decent though (as opposed to a solid brick wall that is the current standard for mass market pop and rock) FLAC is definitely an improvement over 320 or 256+ VBR mp3.
     
  13. plonter
    I never tried ABing flac and lossy before,but i always use losless files (cd at home or flac in portable) for the sake of mind and knowing that I don't compromising anything in the chain.
    anyway..the difference exists,whatever we can hear it or not. and i believe that one shouldn't compromise if he doesn't have to. losless formats were invented for using with a low space machines, so if you have a low space DAP lossless is a must...but today all the majority of players have more th an enough space for handling lossless files.

    even if the recording is bad..you ought to hear it as it is...IN LOSELESS!
     
  14. dasmb
    I've always found mp3 at any bitrate to have trouble with percussion -- cymbals and high hats especially. So much depends on the accurate timing of those transients, and the MP3 format just doesn't have the tools to model them.

    From that perspective, Lossless is a winner. It's the best you're going to get from a given source forever. It's even a winner on bad recordings -- remember, your lossy encoder can't always tell what's desired noise and what isn't. I recently dubbed an old 10 inch of Champion Jack Dupree, and you could easily tell the lossless from the lossy when the guitar started battling record noise. On the lossless, the two were equals; on the lossy, one or the other would lose out as if at random.

    Of course these are all personal trade offs, which is why I couldn't vote in your poll. I'd never use FLAC because the iPod doesn't support it (which was due to playback constraints...FLAC took more resources to decode than ALAC) and I'd choose AAC over mp3 because it handles transients and highs without sacrificing portability. Others choose Monkey or Ogg Vorbis or WMA10, choosing more advanced technology.
     
  15. PiSkyHiFi
    If you have a DAC with 120+ S/N ratio and a really good quality amp to match with high current and a great set of cans with high grade cables, maybe around 2-5% of recordings may reveal something different in 320 Kbps MP3 and FLAC versions. I think if you are missing one of these in the chain, you are probably ascribing incorrectly the effects of using MP3 compared to FLAC..

    I found one recording that did it for me in the very low level background noise of the original analog component of the recording equipment used to make it - the noise was dithered in the 192Kbps MP3 version, not random as it is was on the FLAC version.

    For me, its still noise - it shouldn't be there, but it is - choose your flavour of noise but this one cannot be served sans noise.

    Maybe we need a special "added analog noise" module to make some high end DACs sound more like vinyl.

    I am listening John McLaughlin's "Floating point", hearing quite precise artificial spatial positioning, beautiful crisp snappy percussion and solid strong bass with well designed analog to midi to analog instruments - sounds perfect in a LAME encoded extreme setting variable encoding, does not need anything more to get all the details here.

    Listening to my new Little Dot Mark V with John's "Live at Festival Hall" practically put me on stage with the musicians, hearing nuances in their respective playing styles and getting a feel for the shape of the instruments - from MP3 format. Trilok Gurtu has one of the most diverse range of percussion instruments here - all comes through very clearly in MP3.

    I still really like Mp3s and only use FLAC for a few reference recordings to help establish realistic limits to combinations of equipment and encodings.

    Long live Mp3! but please, no lower than 192 Kbps!
    FLAC ? If the recording is not up to it - you will always have to deal with files 4 times less efficient to handle in every step.
    PCM encoding is useful in the studio where more mixing is required just like when using more bits to reduce mix errors creeping in.
     
First
 
Back
1 2 3 4 5 6 7 8
10 11 12 13 14 15 16 17
Next
 
Last

Share This Page