BDawg
I disagree, I think *ALL* mp3's sound like crap compared to the original, the difference is huge really. Then again, I listen to stuff out of Event 20/20bas speakers which show every detail with a magnifying glass...
I disagree, I think *ALL* mp3's sound like crap compared to the original, the difference is huge really. Then again, I listen to stuff out of Event 20/20bas speakers which show every detail with a magnifying glass...
Workin'
Would using Joint Stereo make a difference (good bad?) with static bitrates up around 192 and 256 kbps? I've seen a lot of recommendations to use Joint Stereo (e.g. on the r3mix.net site), but I am always afraid to use it because of horrible effects when using it with the Xing encoder a long time ago.
Would using Joint Stereo make a difference (good bad?) with static bitrates up around 192 and 256 kbps? I've seen a lot of recommendations to use Joint Stereo (e.g. on the r3mix.net site), but I am always afraid to use it because of horrible effects when using it with the Xing encoder a long time ago.
That link that subhuman posted... it would be helpful if the fellow who did the comparison posted exactly which version of LAME, and which COMPILE even, he used to do the comparison. The latest LAME stable beta - 387bMMX is MUCH better than the current stable non-beta - 370, I think. Lame takes first place for being the most fervently worked-on encoder. They've got guys incorporating all sorts of psycho-accoustic theory into the product daily, while the Fraunhofer guys are just sitting on their butts. 
"BDawg
I disagree, I think *ALL* mp3's sound like crap compared to the original, the difference is huge really. Then again, I listen to stuff out of Event 20/20bas speakers which show every detail with a magnifying glass..."
-------------------------------
In february 2000 c't magazin organised a blind listening test. 300 Audiophiles were involved, finalists tested 17 1-min clips from different artists (classic and pop):
original CD recording
128 Kbit/s Joint Stereo [MusicMatch (FhG) v4.4] encoded PC decoded Mac
256 Kbit/s Joint Stereo [MusicMatch (FhG) v4.4] encoded PC decoded Mac
all on cdrs and played in a Recording Studio on:
B&W Nautilus 803, Marantz CD14 with amp PM14 (Straightwire Pro cabling and extra's) [DM30000- so bit more than $15000]
Sennheiser Orpheus Electrostatic Reference-headphones with tweaked accompanying amp (digital and analog out) [>$10000]
Conclusions:
90% of the 128 Kbit material was picked out
MP3@256 was rated to have the same music quality as cd!
If you find MP3@256 to be of inferior quality compared to the original cd, you're probably doing something wrong with the test or you have failing music equipment or ears. Maybe this is something for you. You can always read the article in the german c't 6/2000 on p92.
The treshold of mp3 transparency lies somewhere between 128kbit/s and 256kbit/s, depending on the kind of music and your hearing and equipment.
-----------------------------------
from http://www.r3mix.net/ click on quality, i remember some site had the real data, no one picked out the cd over the mp3, it was like 3 would pick the 256mp3 out form the cd, but 3 others would pick hte cd saying it was the mp3 quality.... i'll look for it
I disagree, I think *ALL* mp3's sound like crap compared to the original, the difference is huge really. Then again, I listen to stuff out of Event 20/20bas speakers which show every detail with a magnifying glass..."
-------------------------------
In february 2000 c't magazin organised a blind listening test. 300 Audiophiles were involved, finalists tested 17 1-min clips from different artists (classic and pop):
original CD recording
128 Kbit/s Joint Stereo [MusicMatch (FhG) v4.4] encoded PC decoded Mac
256 Kbit/s Joint Stereo [MusicMatch (FhG) v4.4] encoded PC decoded Mac
all on cdrs and played in a Recording Studio on:
B&W Nautilus 803, Marantz CD14 with amp PM14 (Straightwire Pro cabling and extra's) [DM30000- so bit more than $15000]
Sennheiser Orpheus Electrostatic Reference-headphones with tweaked accompanying amp (digital and analog out) [>$10000]
Conclusions:
90% of the 128 Kbit material was picked out
MP3@256 was rated to have the same music quality as cd!
If you find MP3@256 to be of inferior quality compared to the original cd, you're probably doing something wrong with the test or you have failing music equipment or ears. Maybe this is something for you. You can always read the article in the german c't 6/2000 on p92.
The treshold of mp3 transparency lies somewhere between 128kbit/s and 256kbit/s, depending on the kind of music and your hearing and equipment.
-----------------------------------
from http://www.r3mix.net/ click on quality, i remember some site had the real data, no one picked out the cd over the mp3, it was like 3 would pick the 256mp3 out form the cd, but 3 others would pick hte cd saying it was the mp3 quality.... i'll look for it
Possum I would think Joint Stereo would help at those bitrates. Of course, there will probably be times when it might not help, but from my experience, 97 times out of 100 JS sounds better. The best you can do is try it on a wide variety of music and see what happens.
ALL mp3 encoders have a lot more trouble properly encoding "noise-like" sounds like cymbals and even saxophones compared to "tone-like" sounds like a bass or cleanly-struck piano. So you have to try a lot of different situations before you can universally say "X is better than Y", because there is always going to be some situation where you will be wrong.
ALL mp3 encoders have a lot more trouble properly encoding "noise-like" sounds like cymbals and even saxophones compared to "tone-like" sounds like a bass or cleanly-struck piano. So you have to try a lot of different situations before you can universally say "X is better than Y", because there is always going to be some situation where you will be wrong.
PliotronX
Diamond Member
- Oct 17, 1999
- 8,883
- 107
- 106
LAME is simply the best at this point. The upcoming versions are going to have better support for ABR which is like VBR, but you can specify the target bitrate, it encodes in the amount of time it takes to encode a CBR, but comes out sounding and functioning like a VBR file.
I've read that using stereo adds about 30Kbps of filesize, and more often than not sounds like crizzap. I've done a lot of trials on all the major encoders, and LAME sometimes doesn't reproduce basslines as well with the VBR settings as recommended at R3mix.net as it does with CBR 256. The Xing codec sounds like BS even at CBR 320Kbps. With all encoders, at lower bitrates, the most obvious artifacts are the watery flanging effect. At higher bitrates, the flanging effect is reduced dramatically, but the artifacts are loss in volume at high frequencies.
So, it boils down to, if you are an audiophile and can't stand the MP3 compression, but you don't have HD space to spare, the LAME codec using the command line as suggested on R3mix.net will suffice. But if you have a gigantoid HD just sitting there, 256Kbps CBR using the LAME codec will surprise your ears
There is simply no point in using 160Kbps, as it's not better than 128Kbps, and it creates larger files. 192Kbps is basically the same. 192Kbps may sound the best to most people, but it's just not as satisfying as LAME's VBR
I've read that using stereo adds about 30Kbps of filesize, and more often than not sounds like crizzap. I've done a lot of trials on all the major encoders, and LAME sometimes doesn't reproduce basslines as well with the VBR settings as recommended at R3mix.net as it does with CBR 256. The Xing codec sounds like BS even at CBR 320Kbps. With all encoders, at lower bitrates, the most obvious artifacts are the watery flanging effect. At higher bitrates, the flanging effect is reduced dramatically, but the artifacts are loss in volume at high frequencies.
So, it boils down to, if you are an audiophile and can't stand the MP3 compression, but you don't have HD space to spare, the LAME codec using the command line as suggested on R3mix.net will suffice. But if you have a gigantoid HD just sitting there, 256Kbps CBR using the LAME codec will surprise your ears
There is simply no point in using 160Kbps, as it's not better than 128Kbps, and it creates larger files. 192Kbps is basically the same. 192Kbps may sound the best to most people, but it's just not as satisfying as LAME's VBR
Using stereo mode with Lame VBR *on average* adds perhaps 15 bits to the bitrate. It's unfortunate that the CT magazine tests were only done with headphones, which make it psychoacoustically impossible to really discern soundstage properly, no matter how good the headphone system (effects of the sound being generated right at your ear). Listening with dynamic speakers reveals that sometimes joint stereo does indeed give a looser picture of the soundstage. Less distinction across the L-R span, less front-back distinction, a lesser sense of 'presence'. No joint stereo for archival mp3s, for me.
I get pissed off when i download a 128 kbit/sec 'stereo' mp3.
The general rule is this:
128kbit should be in 'joint stereo'
192kbit can be in either 'stereo' or 'joint stereo', if the right channel produces completely different sounds then the left channel (e.g. when the right is instrumental and the left are voices), it's better to use 'stereo'.
The general rule is this:
128kbit should be in 'joint stereo'
192kbit can be in either 'stereo' or 'joint stereo', if the right channel produces completely different sounds then the left channel (e.g. when the right is instrumental and the left are voices), it's better to use 'stereo'.
Yes, joint for 128kbps; either, depending, for 192kbps.
That's why I said stereo for 'archival' quality mp3s. (192kbps doesn't quite qualify)
That's why I said stereo for 'archival' quality mp3s. (192kbps doesn't quite qualify)
for full stereo LAME, you might as well go for the full monty, 320kbits/second, because that dedicates a full 160 kbps to each channel, thus keeping the defects to a minimum (if you encode at anything less then 256 full stereo, you begin to see effects you wouldn't see with joint stereo until you go lower then 192 kbit/second).
I ripped Hampsterdance with the latest (I think) version of LAME, at 320kbps Full stereo, and viewed the resulting file in Cooledit.. it doesn't use frequencies of ~20000hz very often, which makes me wonder what version of LAME I was using..
I ripped Hampsterdance with the latest (I think) version of LAME, at 320kbps Full stereo, and viewed the resulting file in Cooledit.. it doesn't use frequencies of ~20000hz very often, which makes me wonder what version of LAME I was using..
I've sent people sets of 2 wavs before (uploaded to them, or sent CD-Rs) of the same bits of music, one being the wav rip, one being encoded with Lame 3.87 using -V1 -b 128 -h -ms -q1 and then decoded with MAD. I have not informed them which is which. Thus far, as often as not, they've picked the mp3 as the better sounding of the two - but only if they *had* to choose - in no case has any difference greater than the barest been reported. (Which is also my impression.) Granted, the sample size isn't very large...
Becks2k --
I don't care what other people think they hear -- if you make the music, you know exactly how it is supposed to sound like, and even 256 mp3s do strange things to the sound. Yes, it's "pretty good," but it is still different and lacking compared to the original source. If it sounds good to you, then use it. I know that it doesn't to me, especially on Event 20/20bas monitors or larger soundsystems (10 foot JBL stacks) outdoors at a party for example. To me, the difference makes it worth buying the CD for.
Snarko --
If you read the analysis I linked above, you'll see that 192 lacks compared to 256, 256 is the sweet spot for LAME (at least, as mentioned, that paricular version), as LocutusX mentions.
Again, different encoders for different types of music! Not all music has the same frequency content, and not all encoders are created equally -- each one has different trade offs and emphasis on different areas of focus for accurate reproduction.
Use your ears and get some high quality listening gear!
I don't care what other people think they hear -- if you make the music, you know exactly how it is supposed to sound like, and even 256 mp3s do strange things to the sound. Yes, it's "pretty good," but it is still different and lacking compared to the original source. If it sounds good to you, then use it. I know that it doesn't to me, especially on Event 20/20bas monitors or larger soundsystems (10 foot JBL stacks) outdoors at a party for example. To me, the difference makes it worth buying the CD for.
Snarko --
If you read the analysis I linked above, you'll see that 192 lacks compared to 256, 256 is the sweet spot for LAME (at least, as mentioned, that paricular version), as LocutusX mentions.
Again, different encoders for different types of music! Not all music has the same frequency content, and not all encoders are created equally -- each one has different trade offs and emphasis on different areas of focus for accurate reproduction.
Use your ears and get some high quality listening gear!
<< -- if you make the music, you know exactly how it is supposed to sound >>
Not true. Ask Pete Townsend.
Point is this: In scientific tests, people can't tell the difference between CD and 256.
Hm. I said 192 is not quite archival quality. Certainly didn't say I used it. I didn't even say I use the -V1 -b 128 -h -ms -q1 parameters I related that others have as often as not chosen as the better sounding (MAD-decoded) wav over the source, myself. I don't - I use a stereo 256kbps when I want top quality.
But thanks for the affirmation.
But thanks for the affirmation.
glen - dude, so you're saying that the differences i'm hearing are entirely in my head, that i'm not supposed to hear any difference between a 256mp3 and the original wav, even if i authored the original wav? uhhh, no.
how about this. i will send you a 10 second pure WAV file, you encode it using whatever codec, then decode it back to wav, and send me both wav's back, renamed, whatever. i promise not to look at the wave inside of an audio program, just listen on my 20/20bas. i will easily be able to pick out the 256 mp3. scientific tests my butt! =)
Snarko - sorry 'bout that -- but, it is interesting that 192 is so much worse than 256 for filesize vs. quality, don't you think? even stranger is that 320 is worse than 256 (at least according to the tests from
airwindows, i don't ever use 320 myself, just a zipped wav file at that point).
how about this. i will send you a 10 second pure WAV file, you encode it using whatever codec, then decode it back to wav, and send me both wav's back, renamed, whatever. i promise not to look at the wave inside of an audio program, just listen on my 20/20bas. i will easily be able to pick out the 256 mp3. scientific tests my butt! =)
Snarko - sorry 'bout that -- but, it is interesting that 192 is so much worse than 256 for filesize vs. quality, don't you think? even stranger is that 320 is worse than 256 (at least according to the tests from
airwindows, i don't ever use 320 myself, just a zipped wav file at that point).
<< glen - dude, so you're saying that the differences i'm hearing are entirely in my head >>
yes.
<< scientific tests my butt! =) >>
Not to pick on you, but I hope you are not serious!
I think you did hit it on the head in an earlier post - different encoders have different properties, and it is a good thing to be informed/knowledgable about all of the tools available, so you can pick the one that fits your (critical) situation. What's the saying? When all you have is a hammer then everything starts to look like a nail?
Blanket statements about "X is better than Y" have no meaning unless backed up by scientifically valid testing. Otherwise it is just an opinion. Unless the opinion is that LAME is the best
Just kidding, there. But that IS my opinion.The whole mp3 encoder debate has elements of a holy war. It is always amusing because no matter how many times the question is asked, it generates lively debate (by debate I guess I mean flaming).
But at least folks are passionate about it, that can't hurt in the long run.
So "scientifically" MP3s encoded 256 with LAME sound *IDENTICAL* to the original source WAV? I mean, "scientifically" they are massively different, this isn't a loss-less conversion! This is like saying "most graphic artists can't tell the difference between a JPG and the original uncompressed image."
I can't understand how people can say this (glen, Workin'), it's rather scary actually.
I use ZAP from emagic for archival myself.
I can't understand how people can say this (glen, Workin'), it's rather scary actually.
I use ZAP from emagic for archival myself.
Ummm, MP3 regardless of bitrate is still a lossy format. Information is lost, period. I sincerely doubt everyone hears the same and I know people don't see the same. How do you know what you see is blue is the same blue someone else sees? How do you know it's not green? Same thing with sound, most people can't hear dog whistles, but some people swear they can. Get over it man, you may not be able to tell the difference but he very well could.
<< Information is lost, period. >>
True.
But, the point of MP3's is that they leave out info humans can not hear.
To see if they are succesful, we can do testing. If no one can pick the original from the mp3 copy more than 50% of the time.... what do you think?
rahvin, I've always wondered about that :Q
What if people see a completely different set of colors from what we see? - What if their blacks were our whites, etc.
However it works, I like my color set
What if people see a completely different set of colors from what we see? - What if their blacks were our whites, etc.
However it works, I like my color set
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 23K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
News NVIDIA and Intel to Develop AI Infrastructure and Personal Computing Products- Started by poke01
- Replies: 411
Facebook
Twitter