Slashdot Mirror
What Sounds Better, MP3 or Ogg?
I've never been able to make a clear decision on the subject. These days I rip all my CDs to MP3 at 160kbs which means about 80 megs for a longer album. With a 100g drive on order ($220. I remember paying more then that for .1% of that space) disk space isn't really the defining issue, but that doesn't mean I'm gonna rip everything at 300kbs just because I can. I'm curious what people think sounds better, and what bit rates they find to be acceptable for both casual listening, and more picky listening. Don't forget to mention what sort of equipment your listening on so we know where you are
coming from.
This sounds similar to a previous /. story. Although the tests were apparently run with a variety of people in the musical arena, the tests weren't run blindly (apaprently the panel knew if they were listening to an mp3 or an oog file.)
But, it's still worth a read, imho.
--You will rephrase your request for me to go to hell. Goto statements are not acceptable programming constructs
I like mp3 a lot more than ogg. I have an album or 2 ripped with ogg as well as some randoms songs from compilations. I did them around 200kbps VBR and my mp3s are 192 kpbs CBR. I'm listening on cambridge soundworks 4.1 surround speakers on an MX300.
I found the ogg files really tinny and light, so I'd stick with mp3.
Personally I think using r3mix on LAME mp3 encorder makes the mp3 sound exactly like you are listening to the cd. And if you rip the cd with EAC, you have a perfect copy. I never really liked VBR before but it is actually starting to prove itself to be worthy. Check out http://www.r3mix.net for more info.
I actually just did a pretty vigorous test of this the other day. I tested 128, 160, 192, and 256 bitrate mp3s and oggs against the source wav file. At 128 they both sounded similar, but the ogg file did seem a little brighter and clearer than the mp3, and the wav file of course blew them both away. At 160 ogg vorbis really shines... the mp3 remains kind of dull, muddy, and the high end is very "sizzly" compared to the ogg file which sounds brilliant and clear. I barely noticed a difference between the wav file and the ogg at this bitrate. Going up to 192 I found the difference between the ogg and the wav indistinguishable while the mp3 STILL retained some of that annoying high-end sizzle and midrange mud. If you've got the space... 192 oggs amazing... I'm doing mine at 160 because while disc space is cheap, the difference between 160 and 192 is negligible. As for 256... don't bother doing oggs at this level... it's just a waste of disk space. As far as mp3s go... IMO you'd have to encode them at 256 to get the same fidelity as a 160 bitrate ogg vorbis file.(your milage may very... i have been an audio engineer for a while and have picky picky ears.)
Now, if only I could flash my Rio into decoding these files i'd be in digital audio heaven! Also... I'm cannot wait for the 1.0 Ogg encoder to come out... encoding times should be much faster and fidelity even better. Amazing work!
Hope this helps.
-auttie
--->auttie
MPEG is lossy compression. Period. Even if you encode at 320Kbps, you are still losing data. You lose a lot less data at 320Kbps than you do at 128Kbps, but you still lose things.
128Kbps is NOT CD quality. A CD is a 44KHz, 16 bit PCM data stream, uncompressed. It's usually decoded using a 1 bit DAC, IE, via pulse width modulation.
Nothing is "CD Quality" except uncompressed audio or audio compressed with lossless compression like ZIP, RAR, ACE, gzip, bzip, et cetera. "Multimedia" compression is without exception lossy compression - Even our beloved DivX ;-) MPEG-4 High Speed compression is lossy, it's just less lossy than most.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
Do me a favour, everyone:
1) Rip your CD to 128kb mp3s.
2) Re-burn it to CD. (use a rewrite if you're a cheapskate)
3) Listen to the two side-by-side.
Big Fscking Difference!
192 is the best bang for your searching efforts, because any higher takes up too much bloody space. But that encode I can burn to a custom CD and it'll sound fine.
Now, back to Ogg... Ogg sounds about the same at 160kb as an mp3 at 190, (debate and argue all you want...) which is why I like its compression system. Still, I wouldn't touch a 128kb ogg either
Yes and no, but mostly no. I have a pair of Bose 601 Series II speakers attached to my stereo. They reproduce sounds *very* well. The stereo in question is a Sony STR-DE635 reciever with dolby digital, et cetera. It has 80 watts x 5 channels of discrete amplification. While there are stereo systems with a higher signal to noise ratio, this is a pretty damn good setup. I play mp3s from either my dreamcast (analog output) or my new Apex AD-3201 DVD player, which has a truly crappy interface for playing mp3s, but sounds okay, and plays VBRE without any trouble. It's got a coaxial digital connection to my stereo, and it spits mp3s out at it at 44.1KHz, 16 bit PCM (after decoding).
With all this said, I can definitely hear the inconsistencies in lower-bitrate (like 128Kbps) mp3s. The only encoding rates I'll use any more are mid-high VBRE (which will go up to 290Kbps or so) and 320Kbps for archival purposes of very touchy music, like classical pieces. If I have something which is purely spoken word, and it doesn't involve screaming, I'll sometimes drop down to 64Kbps mono just to make the files smaller, but generally I encode them as VBRE along with everything else.
The thing you really tend to lose a lot of in 128Kbps mp3s is bass. Deep bass tends to get crunchy VERY fast, even at slightly higher bitrates like 192Kbps. You can actually hear that even on computer speakers (I use a microsoft digital sound system in analog mode only) or on your car stereo (I burn mp3s back to CD fairly frequently) but especially on a high-end stereo, which will more faithfully reproduce the sounds its given. So actually, on a higher-end stereo, you will hear every bad frequency caused as a compression artifact.
I have no idea what low-bitrate audio sounds like on true 'prosumer' level home theater sounds like, but I bet it's really atrocious. It's bad enough on my only somewhat upscale layout.
"You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
I just read this article about Croteam using it for their next game:
"We did a major change in the sound engine between FE and SE. And its name is Ogg Vorbis. Yeah, that's right, we're using ogg for music playing. In case someone hasn't heard of it yet, Ogg Vorbis (http://www.vorbis.com) is a patent-free, open source audio codec project. Or in english: a music compressor that plainly rocks. Make sure you check it out. We've tried encoding all the music for SE with Oggdrop at 64kbps and the quality was perfect even at such low bitrate. In the final version, since we won't need the extra space, we'll be shipping with 128kbps music tracks, for even higher fidelity. The guys there are really helpful and supportive and the whole project is surprisingly functional already. There are plugins for all major music players and other music programs."
Ant(Dude) @ Quality Foraged Links (AQFL.net) & The Ant Farm (antfarm.ma.cx / antfarm.home.dhs.org).
The cake is a pie
Of course, with a drive that size, you could go all-out and use Monkey's Audio, lossless audio compression (you can decode to get *exactly* the same WAV file that was encoded. Compression ratio of only 2:1 or so, but again...what's the 100 GB drive for?!! Get on Google and search around for some comparisons, and make an educated choice.
I sent them an email asking for ogg support and they said if there was enough interest they would implement it.
course. someone'll have to mod this up first. ;).
If you're running Windows, you can get ABX from http://www.pcabx.com/. On UNIX systems, the LAME source code comes with an ABX program (in the misc/ directory, I think).
Here is an example of a test that took place using a slightly different testing methodology, more akin to MUSHRA (which is used to evaluate lots of encoders at the same time): http://www.ff123.net/128tests.html.
-- Help Digitise the Public Domain at DP.
You get almost the same quality at half the size with WMA
Sadly, you can't. Listening tests have shown that WMA 8 has sacrified sound quality at medium/high bitrates over WMA 7 to improve quality at the low end. So it's great for music over a modem, but at 96k and above it is no better than Ogg Vorbis.
-- Help Digitise the Public Domain at DP.
Iomega has promised a firmware update for their HipZip supporting Vorbis as soon as 1.0 is released.
And yes, the quality:bitrate ratio in ogg kicks mp3's ass.
--
grep "xercist"
No, it's not, not if you're playing music originally from CD. CD's are stereo. Not 4-channel, not 5.1. Do you expect your surround system to magically figure out what speaker to send a signal to?
With that said, four-speaker stereo can significantly increase the size of your room's "sweet spot" and reduce the stereo distortion effect you hear when turning your head. Add a subwoofer for deep bass response, and that's about the most you'll need for accurate playback of any two-channel source.
Actually, the primary objection to Mp3 is not the compression. Rather, it is the licensing issues surrounding Mp3.
Read all about it at http://www.xiph.org/about.html.
Daniel
I haven't directly compared OGG and mp3, mostly because I'm very happy with the quality of the mp3 encoding.
In my own testing, the r3mix.net settings were pretty much indistinguishable from the original in terms of frequency response. I did notice some changes in spatial effects. One of my CDs in particular was affected, Deepforest 2. With the original CD playing, the sound tended to bounce all around your head when wearing headphones. After being encoded by LAME, the sound still moved some, but it was much more granular. Most of the effect was lost. However, the actual FREQUENCY RESPONSE was awesome, and the only way I could really tell the difference was by listening very intensely. It is more than adequate for normal listening.
I did these tests about a year and a half ago, on LAME 3.81, and apparently it has improved quite a bit since. That team respects the r3mix site enough that they actually added in an '--r3mix' command line switch to implement all of their suggested settings at once. Apparently LAME now keeps more of the original signal; it's not quite so enthusiastic about assuming you can't hear certain kinds of noise. I'm hopeful this may have fixed the encoding issues I had with the earlier version.
Basically, given the fact that he has tons of space available, and given that there's all sorts of portable MP3 players in the world, I think he may still be happiest with MP3. I certainly am.
Equipment used: Non-golden ears, but decent ones. Soundblaster Live Platinum 5.1 (which has some frequency response issues with REAL audiophiles), Sennheiser HD 580 headphones for 'real' listening, Midiland S2 4100s (the older 2 speaker model) for casual music and gaming.
Aside: The 580s are AWESOME headphones, and you can often get them very cheap at auction. I got mine about two years ago for about $125. They have a reputation of having flaky connections. Mine did indeed have a problem when I first got them, which I solved simply by removing and replugging the wire in the bottom of the headphone. They are fully modular, easy to disassemble and clean, and sound INCREDIBLE. Two downsides: they really need an amplified headphone jack to reach their true potential, and they are big headphones. They're very comfortable but large.
Aside on the early model Midilands: great quality speakers, dismal amp. Hissy at any volume. Someday I'll move the way-cool little satellites onto a real amplifier, and will toss the subwoofer/amp in the trash.
BTW, for more in depth discussion that has been ongoing, have a look at the forums at r3mix.net and the Ogg-specific forums at Hydrogen Audio. I keep up with both forums, and the folks there tend to make prerelease build binaries available for people to play with. For up-to-date detailed information without the overhead of the Vorbis-dev list, those are the places to go.
One more link for folks who want to know more: The beginning of the document describing Vorbis stereo discusses good terminology and qualification of subjective fidelity. It's nothing new to most posters I expect, but it might help keep the discussion consistent.
Happy hacking,
Monty
xiph.org
Actually the concern is the ~$20 per unit that the LICENSE to use an encoder costs. That's any encoder, not just theirs. And this isn't so much the problem, as is the fact that Fronhofer (I believe that's who) can change this price at any time, on a per customer basis. Say maybe at the urging of the RIAA in order to eradicate all mp3 encoders in favor of SDMI (or whatever).
Open formats are critical to open information exchange, this is exactly why there is such a fight to keep patents out of the w3c standars.
The quality of your connectors is more important than that of your sound card. Bring the audio to your receiver over SPDIF or TOSLINK, not over analog RCA cables! Sound cards --- ALL of them --- have really awful RCA connectors.
While digital interfaces bring a theoretical possibility for a quality change over analog links, this is _not_ due to the properties of the cables or jacks.
Short of a connector totally covered in corrosion, no jack or reasonable cable will ever influence signals in the audio band.
Even SPDIF and TOSLINK aren't lossless
Yes they are, these are straight digital interfaces. Short of malfunction no data will be lost through them.
I can't tell the difference between 256 and anything above. VBR improves sound quality when you set a floor of 256 and a ceiling of infinity; otherwise, it's just a silly hack to save disk space at the expense of your MP3 files. It may not noticeably damage audio quality, but it sure as hell makes your MP3 files more complex, harder to analyze and play with/sort/etc. MP3 is just a poor file format for what VBR asks it to do.
VBR is part of the mp3 stadard, so it's not a hack by any stretch of the imagination.
VBR is IMO the Right Way(TM)to do audio coding as it essentially let you select a target quality instead of a target bitrate.
Current implementations of VBR are good enough to not degrade the sound noticeably so there is no real reason not to encode with VBR.
can't tell the difference between 256 and anything above. VBR improves sound quality when you set a floor of 256 and a ceiling of infinity; otherwise, it's just a silly hack to save disk space at the expense of your MP3 files. It may not noticeably damage audio quality, but it sure as hell makes your MP3 files more complex, harder to analyze and play with/sort/etc. MP3 is just a poor file format for what VBR asks it to do.
If joint stereo is a hack, then what do you call all the other techniques that make up mp3/ogg/whatever encoding.
JS simply utilizes the fact that significant signal is common for both channels and encodes this only once. Storing this information twice makes little sense.
JS is a efficient way to reduce space, which can be used to increase overall sound quality by using less aggressive compression on areas which actually matter.
SHN, perhaps?
TO BUY A NEW CAR WOULD MAKE YOU SEXUALLY ATTRACTIVE.
Sadly, this often isn't valid for quite a few reasons.
Quite often, encoders will use very different procedures to encode a low bitrates than they would use to encode at high bitrates. They will probably use a hearing model (which models the ATH - absolute threshold of hearing) which is less demanding, for example. They may even automatically low pass the music, or resample it to a lower bitrate.
For example, Ogg Vorbis has different methods of channel coupling. At very high bitrates, no stereo information is lost. At medium bitrates, no stereo information is lost for the sounds we are most sensitive to, but for others the phasing is quantisised. The degree of compression determines the range of lossless coupling, and also the amount of quantisisation -- and each may have its own distinctive artifacts.
MP3 can't encode stereo 44kHz (CD quality) sound at 48kp/s without sounding truely terrible. If you try with LAME, you will find that it automatically resamples, and uses one of the 'extensions' to the official MP3 specification which encode better at low bitrates by resampling the sound.
-- Help Digitise the Public Domain at DP.
Ok, we got many things (using lame style names):
CBR = Constant bit rate = Variable quality
VBR = Variable bit rate = Constant quality
ABR = Target bit rate = Variable but not as much quality
OGG normally uses a form of ABR, but is capable to do true CBR and true VBR as well (not sure which versions enabled for).
Also, even if you are using true CBR, there is little room for flexibility in the form of the "bit reservoir"; you can save some bits in the "easy parts" so they can be better spent in the hard parts.
Second, mp3, being open in some way or another, has the side effect of many encoders available. Different encoders produce different quality. Take 4 192kbps mp3s encoded with 4 different encoders, and you will discover quality differences as day to night.
And to use Lame properly, first, let me suggest that you *at least* use Lame 3.89b. Lame 3.70 is *too old*. If you get Lame 3.90a, even better.
Want to be on the safe side? use this single option:
lame --dm-preset standard
This will produce near 256kbps files, and its the hightest quality you can get out of mp3s.
If you think you can live with 192kbps like files, then use
lame --r3mix
Otherwise stick to the normal, don't apply options you don't know much of. Typically you *always* want -h, and -b for the desired bitrate in case of CBR, or minimun frame bitrate for audio in the case of VBR (usually 112 or 128). ABR is VBR attempting an average bitrate. And no, it is not wise to use option -B at all (let the encoder use up to 320kbps frames when using VBR).
If this topic of lossy compression is of interest for you, then you should visit:
Proyect Mayhem, channel #Project_Mayhem at irc.openprojects.org
and
r3mix.net, channel #r3mix at irc.openprojects.org
Um... on side note, have you seen The Wavelet Tutorial yet? Wavelets are planned for Ogg Vorbis 2.x, stay tuned... :)
Artix
Your Linux, your init.