Extensive Format Comparison WAV FLAC APE MP3 MP3PRO MPC OGG WMA(french)

[Febs]

Originally Posted by Audax

i agree that your first reference should always be your ears. But a Bandwidth test is a useful tool when setting up a sound system. There is a reason to that, a microphone will always hear more than you do ...

Sure, but you're not talking about setting up a sound system. You're talking about evaluating lossy codecs. No-one disputes that lossy codecs discard a substantial portion of the information that is contained in the original. They're designed to do exactly that. What is really relevant is the audible impact that the process of discarding that information has on how the files sound. You can glean some useful information from graphs such as the ones linked to in this thread, but given the fact that the goal of lossy codecs is to discard information that is not audible and retain information that is audible, the graphs will not necessarily tell you how the file sounds.

Also, the fact that the graphs presented do not have a consistent X-axis makes it somewhat difficult to evaluate them.

[Audax]

Originally Posted by Febs

Sure, but you're not talking about setting up a sound system. You're talking about evaluating lossy codecs. No-one disputes that lossy codecs discard a substantial portion of the information that is contained in the original. They're designed to do exactly that. What is really relevant is the audible impact that the process of discarding that information has on how the files sound. You can glean some useful information from graphs such as the ones linked to in this thread, but given the fact that the goal of lossy codecs is to discard information that is not audible and retain information that is audible, the graphs will not necessarily tell you how the file sounds.

100% true, only newborns and youg children can hear above 20 Khz.

Also, the fact that the graphs presented do not have a consistent X-axis makes it somewhat difficult to evaluate them.

Yes, but they are consistent enough to show where the drops occur.

i appreciate your input and do not hesitate to correct me in case i talk bullsh"t

[krmathis]

Interesting!
Graphic comparison is not accurate though, as it tell you little/nothing about the audible difference.

[jilgiljongiljing]

Careful folks I just saw some NSFW ads on the page.

[jiiteepee]

There are some graphical results from a test I made ~year ago (graphics are generated recently) behind the link given below (and another link on the next page).

There'll you find spectrum displays for:

Original data file - WAV, 16-bit/44.1kHz, Linear Sweep 20Hz-20kHz -6dB)

IIS Fraunhofer Pro v. 3.3.2.144
Lame 3.96
WMA v. 4.02
OGG Xiph. Org Vorbis I
AAC

http://jiiteepee.fortunecity.com//te...omparison.html


jiitee

[ADD]

Originally Posted by EnOYiN

Try and explain to me why that would not be correct. Try it unless you want to start a lossless vs. lossless debate. In that case don't try to explain it please.

Maybe I have super ears, but the difference between an uncompressed WAV file and the highest possible 2 channel stereo bitrate MP3 (320), WMA (high 400s) and OGG (505) are so obvious to me it's not funny. That is not to say that the high bitrate compressed files sound bad - but they do sound different and are lacking accuracy, detail, tonal timbre and dynamics compared to the original. Apart from that, WMA adds a "sheen" to the extreme high end that is not there in the WAV file, the Fraunhofer softens the high end and LAME hardens it. So if for nothing else, they all have their own sound signatures at maximum birtates that make then easy to distinguish from the originals.

PS: not at all interested in the advice about not attacking Bigshot. Would much rather be banned please. I can't stand know-it-alls who...umm...don't and who furthemore give idoitic and misleading advice to people who just suck it all up without using their own judgement.

[lordmozilla]

Originally Posted by ADD

Maybe I have super ears, but the difference between an uncompressed WAV file and the highest possible 2 channel stereo bitrate MP3 (320), WMA (high 400s) and OGG (505) are so obvious to me it's not funny. That is not to say that the high bitrate compressed files sound bad - but they do sound different and are lacking accuracy, detail, tonal timbre and dynamics compared to the original. Apart from that, WMA adds a "sheen" to the extreme high end that is not there in the WAV file, the Fraunhofer softens the high end and LAME hardens it. So if for nothing else, they all have their own sound signatures at maximum birtates that make then easy to distinguish from the originals.

PS: not at all interested in the advice about not attacking Bigshot. Would much rather be banned please. I can't stand know-it-alls who...umm...don't and who furthemore give idoitic and misleading advice to people who just suck it all up without using their own judgement.

I'll just chime in and say that the codec your using to play the file could be the thing adding/taking away this detail you are finding.

However I disagree with bigshot, a 505kb/s ogg will not sound exactly like the original.

[ADD]

Well I will clarify once again that my point is the compression does not sound the same as the original - not that it necessarily sounds bad. And that is where I take extreme exception to Bigshot's comments. If he had said "to my ears, there is no difference between the highest bitrate compression and the original", I would have let things go, but he gets away with these things time and time and time and time again! He never ever offers his opinions - he just offers everything as a hard fact

For my own part, I happily admit to using Ogg 505 ABR for all my audio compression and I am very happy with th results - it still sounds extremely good to me. What I do clearly know however is that to my ears, even the best compression sounds different and in all cases that equates to worse as well. But for OGG 505, the difference is small enough to only annoy me on direct switching A/B comparisons - not if I just get up the next morning and listen to the OGG instead of the WAV file.

And just in case I am hanged and banned let me say just this to all impressionable people. I cannot sleep tonight without saying this, because I feel that comments made on this website by certain members are delivered in such a manner as to potentially hamper the long term musical enjoyment of other members. So...


1. Do not take anyone's view of audio sound quality, including mine, as factual. It is only their opinion, even if they appear to state it ad infinitum as a hard fact. The problem I have is when opinions are continually stated as patent facts to individuals who may not have enough experience to make a judgement call for themselves.

2. Trust your own ears - not those of others. But do not trust your own ears until you have given them the best opportunity and sufficient time to develop and maximise their forensic, musical hearing capabilities. Sometimes this process takes years and for many it continues to develop throughout life (upgrade-itis being the most notable symptom).

3. If you are new to audio, let those hearing capabilities develop over a period of at least 12 months before deciding for yourself what sounds good / better / best.

4. Don't transcribe that vinyl collection to CD unless you really don't think that for the rest of your days you will ever be able to hear the difference between doing that and transcribing it to 24-96, 24-192 or DSD. It would be terrible if you did that because someone kept telling you in a factual manner till the cows come home that CD copies sound no different. Especially if you don't have matured listening capabilities to make a judgement call for yourself (as opposed to doing that just because someone told you so).

After all, years down the track if you did discover you could hear the improvement, what a huge waste of your life that would have been transcribing 500 LPs to CD.

And if you can never tell the difference, that's fine and good for you too - you will save a lot of money on equipment and a lot of money on disc and archiving space.

5. Same principle with the topic under consideration as regards compression. Don't transcribe your music collection to even the maximum bitrate compression unless you are certain you will never hanker for more quality. If you spend 5 years developing your musical ear and can't tell the differences then fine - you will make some great savings. But if you go to all that time and effort then down the track you hear a lossless source (having gotten so used to compression you treat it as the norm), you might sincerely regret all that time you spent. Especially nowadays with storage space getting cheaper and cheaper.

It's comments about compreesion being inaudible that I feel are extremely damaging to the potential for good sound quality in the industry. If we accept that compression and originals sound the same, then there is no motivation any more for ultra high fidelity and music vendors will have no motivation to offer anything more than those MP3 downloads that have had the gritty guts of the music ripped out of their binary souls...

That's fine for those who can't hear the difference, but it's a grave consequence to those who can hear the difference.

[Febs]

Originally Posted by ADD

For my own part, I happily admit to using Ogg 505 ABR for all my audio compression and I am very happy with th results - it still sounds extremely good to me.

If you're going to use a bitrate that high, why not just use FLAC?

[lordmozilla]

Originally Posted by Febs

If you're going to use a bitrate that high, why not just use FLAC?

files would still be about half smaller than flac...

[jiiteepee]

Added some spectral results prepared from comparison between MP3 /lame/Fraunhofer) : WMA : OGG : AAC.

http://www.geocities.com/jiiteepee/comparison.html (transfer limit)
or
http://jiiteepee.fortunecity.com/pic...omparison.html


jiitee

[stevenkelby]

Does anyone here think that if a bitrate comparison/difference is inaudible to their ears with their system, it will therefore be inaudible to my ears with my system?

[LukeV]

EDIT <Nevermind >

[stevenkelby]

Originally Posted by LukeV

Edited

SHHH! Keep it down!

[LukeV]

Originally Posted by stevenkelby

SHHH! Keep it down!

Alright then XD