Is LAME the best *encoder* for MP3?
- Thread starter xtreme2k
- Start date
Fraunhofer IIs seems to be the best Encoder. I personally don't much like their Decoder though. I use LAME and I think it work great, I encode all my Mp3s at 192kb/s and I haven't had any problems with LAME.
I'm kind of Partial to Winamp's Nitrane for a decoder (which they took out AGAIN...bastards, fortunately I have a copy of the Nitrane decoder from 2.65 and it's easy to install it into newer winamp versions).
I'm kind of Partial to Winamp's Nitrane for a decoder (which they took out AGAIN...bastards, fortunately I have a copy of the Nitrane decoder from 2.65 and it's easy to install it into newer winamp versions).
which option do u use for lame?
-b 256 -m j -h -k
this is what I use
CBR 256Kbit
Highest Quality
Joint Stereo
another question, what is the difference between 'stereo' and 'joint stereo'. I have heard J-Stereo is better for 128kbit, but what about 256kbit? From what i know, even audio experts cannot tell difference between 256kbit and true CD Audio.
-b 256 -m j -h -k
this is what I use
CBR 256Kbit
Highest Quality
Joint Stereo
another question, what is the difference between 'stereo' and 'joint stereo'. I have heard J-Stereo is better for 128kbit, but what about 256kbit? From what i know, even audio experts cannot tell difference between 256kbit and true CD Audio.
256kbit's sound is a lot more 'natural' if you hear them from some decent speakers
128kbit is acceptable but noticeable that it arent as good.
128kbit is acceptable but noticeable that it arent as good.
The concensus wisdom seems to be the best encoder (ranked by both
quality of encoded signal AND amount of compression) is LAME
using Variable Bit Rate (VBR) @ the highest level with a floor
of 128 kbps and a quality index of "1".
Now here's my question: Why is Fraunhoffer @ the highest quality
setting so incredibly slow? LAME according to the above prescription
encodes on my machine at around 6x the length of the track.
Fraunhoffer @ the highest quality setting encodes at around 0.30x.
Whattadifference...
Anyone know why?
Kwad
quality of encoded signal AND amount of compression) is LAME
using Variable Bit Rate (VBR) @ the highest level with a floor
of 128 kbps and a quality index of "1".
Now here's my question: Why is Fraunhoffer @ the highest quality
setting so incredibly slow? LAME according to the above prescription
encodes on my machine at around 6x the length of the track.
Fraunhoffer @ the highest quality setting encodes at around 0.30x.
Whattadifference...
Anyone know why?
Kwad
Hmm, well it's not really that cut-n-dry.
You might want to check out this link which shows that SOME PARTICULAR codec is much better but only at HIGHER bitrates. It's a good read and will help dispell some of the myths and misconceptions floating around here...
Short version: Frah. Codec kicks LAME in the butt at lower bitrates, but surprisingly, at higher bitrates, LAME is better.
You might want to check out this link which shows that SOME PARTICULAR codec is much better but only at HIGHER bitrates. It's a good read and will help dispell some of the myths and misconceptions floating around here...
Short version: Frah. Codec kicks LAME in the butt at lower bitrates, but surprisingly, at higher bitrates, LAME is better.
<< you like nitrane?!? you know that that piece of junk doesn't decode properly, right >>
Yes and yes
I said I'm partial to it, I didn't say it was good I'm just used to the way it sounds. Fraunhofer is probably better, but Fraunhofer won't play some of my MP3s (poorly encoded probably, but the fact remains).SSP: Depends on how sensitive your ears are. I don't much like the sound quality at 128kb/s, at 160kb/s it's better, but I like 192kb/s even more. To my ears 256kb/s doesn't sound noticeably better than 192kb/s but 192 is certainly better than 128 or 160. But I find my ears are bit more sensitive than most of my friends (hense my extreme hatred for the Promedia's
lol), most of my friends think that 128kb/s sound great.I use AudioGrabber with the LAME plugin, I think the settings I use are equivilant to -V1 -b192 -mj -h -q1 I'm not 100% sure though becuase I've never run it command line.
Well, Subhuman, it really is that cut-and-dry, if you follow
the collective wisdom, and it's as I said before: Lame VBR
at the highest rate, with q=1 and a base rate of 128 kbps.
In my opinion, NO MP3 encoder is particularly good at 128Kbps or
less. Sure, they sound fine on your Walkman (well, sometimes)
or through your mediocre computer speakers (and just about
ALL speakers that are marketed as "computer speakers" are
mediocre or worse...) But play them through some good quality
headphones (I recommend Sennheiser HD580 or HD600 or Stax), or
through some real high quality speakers, and 128Kbps just rots.
Not close to anyone with discerning ears.
This is fact, not opinion. OK, perhaps the word "rots" is
opinion. Perhaps I could say '128kbps does not create an
encoded file that cannot be distinguished from the original
source in A/B comparisons'. There, now that I've said that,
I won't need to see the "oh, I like 128Kbps just fine" response
that is just waiting around the next posting bend...!
So to me, which encoder is best at 128Kbps or less isn't
particularly important. To my ears, they all sound pretty bad
when listened to in a critical environment, and they all sound
passable when listened to in a non-critical environment. CERTAINLY
128Kbps is not archival. I have used the Fraunhoffer encoder at
128Kbps CBR and it sounds fine. So does LAME. Spectrum analysis
implies that Fraunhoffer is a bit better. But only a bit. And
psychoacoustics being what they are, it's hard to say on the
basis of the modest differences whether either is really "better".
Kwad
the collective wisdom, and it's as I said before: Lame VBR
at the highest rate, with q=1 and a base rate of 128 kbps.
In my opinion, NO MP3 encoder is particularly good at 128Kbps or
less. Sure, they sound fine on your Walkman (well, sometimes)
or through your mediocre computer speakers (and just about
ALL speakers that are marketed as "computer speakers" are
mediocre or worse...) But play them through some good quality
headphones (I recommend Sennheiser HD580 or HD600 or Stax), or
through some real high quality speakers, and 128Kbps just rots.
Not close to anyone with discerning ears.
This is fact, not opinion. OK, perhaps the word "rots" is
opinion. Perhaps I could say '128kbps does not create an
encoded file that cannot be distinguished from the original
source in A/B comparisons'. There, now that I've said that,
I won't need to see the "oh, I like 128Kbps just fine" response
that is just waiting around the next posting bend...!
So to me, which encoder is best at 128Kbps or less isn't
particularly important. To my ears, they all sound pretty bad
when listened to in a critical environment, and they all sound
passable when listened to in a non-critical environment. CERTAINLY
128Kbps is not archival. I have used the Fraunhoffer encoder at
128Kbps CBR and it sounds fine. So does LAME. Spectrum analysis
implies that Fraunhoffer is a bit better. But only a bit. And
psychoacoustics being what they are, it's hard to say on the
basis of the modest differences whether either is really "better".
Kwad
Try CDex. It uses Lame. It was recommended to me by someone on this forum. Sure it has its bugs, but it does one heck of a job and it it free...http://www.cdex.n3.net/
You'll love http://www.r3mix.net. It has probably the best comparison on the web.
I redid the tests they did, and Lame is indeed by far the best encoder there is.
I redid the tests they did, and Lame is indeed by far the best encoder there is.
"The concensus wisdom seems to be the best encoder (ranked by both
quality of encoded signal AND amount of compression) is LAME
using Variable Bit Rate (VBR) @ the highest level with a floor
of 128 kbps and a quality index of "1"."
Amen, my brother. Cut and dried, slam dunk, no questions asked. And don't forget to use Joint Stereo! I use a floor of 96 kbps, but I don't think it really makes a difference in practice.
Joint Stereo allocates the bitrate bandwidth more efficiently. Instead of forcing the encoder to encode ALL the data in BOTH channels, JS encodes the stuff that's the same between both channels and then the the stuff that's different between the channels using a matrixing scheme that gives you a higher "effective" bitrate than if you used "real" stereo. A common artifact from using Forced Stereo is "flanging", which is that swishy sound you hear a lot. Joint Stereo can help reduce flanging, especially at lower bitrates.
quality of encoded signal AND amount of compression) is LAME
using Variable Bit Rate (VBR) @ the highest level with a floor
of 128 kbps and a quality index of "1"."
Amen, my brother. Cut and dried, slam dunk, no questions asked. And don't forget to use Joint Stereo! I use a floor of 96 kbps, but I don't think it really makes a difference in practice.
Joint Stereo allocates the bitrate bandwidth more efficiently. Instead of forcing the encoder to encode ALL the data in BOTH channels, JS encodes the stuff that's the same between both channels and then the the stuff that's different between the channels using a matrixing scheme that gives you a higher "effective" bitrate than if you used "real" stereo. A common artifact from using Forced Stereo is "flanging", which is that swishy sound you hear a lot. Joint Stereo can help reduce flanging, especially at lower bitrates.
I don't know how this affects the argument, but I'm sitting here, listening to mp3s recorded with Fraunhofer codec at 160kbps on my nice Sony headphones, and they sound great. Pretty near perfect with a ton of resolution.
No, this isn't with crappy computer speakers, but quality headphones (I've never hooked up my computer to my Yamaha/Klipsch home theater system )I think I would be hard-pressed to distinguish from the CD.
No, this isn't with crappy computer speakers, but quality headphones (I've never hooked up my computer to my Yamaha/Klipsch home theater system
)I think I would be hard-pressed to distinguish from the CD.TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 24K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 20K
-
-
-
Facebook
Twitter