MP3 sound quality: Good enough(newyorker.com)
newyorker.com
MP3 sound quality: Good enough
http://www.newyorker.com/online/blogs/sashafrerejones/2009/09/dithering-jonny-greenwood.html
8 comments
Hmmm... Well I agree with you about him taking it to far. I know monster cables are bull, but true audiophiles ($25000 and up per system) will use non-monster oxygen-free cables, and they swear by them. I will give it to them, some of those systems really sound great. I don't know if it's due to the lack of oxygen in the lines though.
I promise you, it's not. There's a lot of articles floating around recently about the placebo effect in medicine. I'll be kind to these people, and let them off with the excuse that placebo effect, which is still not understood, is the reason their $500 cables make them think their audio sounds better.
Audio is something we understand pretty well. We can measure it's delay and phase and amplitude and distortion. We can measure these things, and we can see when they change, and we can see when they don't change. We can see these changes to levels that the human ear can't detect, but even if you believe that an audiophile can detect things that others can't, we can see these changes on our instruments.
I'm not certain how far to beat this horse. It's a little like arguing that gold doesn't fall faster than lead, even though gold is heavier. At the end of the day, you either believe in measurable science, or you don't. As somebody who has spent countless hours in front of an audio analyzer, tracking down actual audio effects all the way down into the noise floor of the system, I am confident on this issue.
Still, the guys selling those cables are richer than I am, so who's the idiot? Certainly not the salesmen.
Audio is something we understand pretty well. We can measure it's delay and phase and amplitude and distortion. We can measure these things, and we can see when they change, and we can see when they don't change. We can see these changes to levels that the human ear can't detect, but even if you believe that an audiophile can detect things that others can't, we can see these changes on our instruments.
I'm not certain how far to beat this horse. It's a little like arguing that gold doesn't fall faster than lead, even though gold is heavier. At the end of the day, you either believe in measurable science, or you don't. As somebody who has spent countless hours in front of an audio analyzer, tracking down actual audio effects all the way down into the noise floor of the system, I am confident on this issue.
Still, the guys selling those cables are richer than I am, so who's the idiot? Certainly not the salesmen.
Further robotrout's point: a friend of mine did some research on actual sound quality for generic versus high-price cables.
He measured the noise floor with Monster cables at about -160dB, whereas cheap crappy cables were about -125dB. However, he also found that basically all of the difference is explained by higher quality connectors on the Monster cables compared to his ultra cheap bargain basement ones; when he replaced the cheap connectors with some modestly priced RCA connectors that weren't utter shit (but were not Monster cable made), the two cables were indistinguishable.
For the record, even -125dB is below the noise floor of arguably every person on earth.
So really, the most important thing for cables is to always use good connectors. You cannot hear the difference between "super duper copper" and "layman's copper."
Another popular bit of audiophile snake oil is "ribbon cables." These use wide, flat conductors instead of round ones because---they claim---the round conductors have higher impedance because of the skin effect (http://en.wikipedia.org/wiki/Skin_effect ). Of course, this is absolute bullshit, because skin depth at audio frequencies is more than the diameter of the round conductor. What's more, since ribbon cables aren't coaxial, they're actually more susceptible to EMI.
...and people still spend hundreds or even thousands _on_the_cable_!
He measured the noise floor with Monster cables at about -160dB, whereas cheap crappy cables were about -125dB. However, he also found that basically all of the difference is explained by higher quality connectors on the Monster cables compared to his ultra cheap bargain basement ones; when he replaced the cheap connectors with some modestly priced RCA connectors that weren't utter shit (but were not Monster cable made), the two cables were indistinguishable.
For the record, even -125dB is below the noise floor of arguably every person on earth.
So really, the most important thing for cables is to always use good connectors. You cannot hear the difference between "super duper copper" and "layman's copper."
Another popular bit of audiophile snake oil is "ribbon cables." These use wide, flat conductors instead of round ones because---they claim---the round conductors have higher impedance because of the skin effect (http://en.wikipedia.org/wiki/Skin_effect ). Of course, this is absolute bullshit, because skin depth at audio frequencies is more than the diameter of the round conductor. What's more, since ribbon cables aren't coaxial, they're actually more susceptible to EMI.
...and people still spend hundreds or even thousands _on_the_cable_!
I agree with you about the cables, but this:
> It's a little like arguing that gold doesn't fall faster than lead, even though gold is heavier.
is a bad example, because heavier items do fall faster. If the difference in masses between them is very large (like a box vs the earth), the difference is not measurable, but it's there. It's off topic, but if you want me to show you the math I can.
> It's a little like arguing that gold doesn't fall faster than lead, even though gold is heavier.
is a bad example, because heavier items do fall faster. If the difference in masses between them is very large (like a box vs the earth), the difference is not measurable, but it's there. It's off topic, but if you want me to show you the math I can.
The problem isn't that $format doesn't sound good, the problem is that when you convert $format to $otherformat for playback on $system, you lose far more quality than if you converted from the original audio file.
Two words: repeat sales.
Those who care about audio quality won't abandon their hardware playback platform for exactly this reason. Even those who don't understand the technicalities are vaguely aware of the problem.
The solution is FLAC, but since it's free, lossless, and (typically) 2-3x larger, it has had fairly slow growth. Still: http://flac.sourceforge.net/links.html
Those who care about audio quality won't abandon their hardware playback platform for exactly this reason. Even those who don't understand the technicalities are vaguely aware of the problem.
The solution is FLAC, but since it's free, lossless, and (typically) 2-3x larger, it has had fairly slow growth. Still: http://flac.sourceforge.net/links.html
ACK. Additionally, FLAC is perfect for future-proof archiving. MP3 is already declining (thanks to Apple) and while you will probably still be able to play MP3 files in twenty years, it will hopefully not be the format of choice. Other formats are superior to MP3 even today -- not only in sound quality, but in other areas like gapless playback and metadata capabilities as well.
But hey, Jonny is still one of my heroes. :)
But hey, Jonny is still one of my heroes. :)
"MP3 is already declining (thanks to Apple) "
How do you know that? And what format is supposed to be replacing mp3? It's the only audio format I run into on the Web (but then I don't use Apple products to play music).
How do you know that? And what format is supposed to be replacing mp3? It's the only audio format I run into on the Web (but then I don't use Apple products to play music).
I don't use Apple products either but I know that they sell tracks AAC encoded. And that's exactly why I see MP3 declining: before iTunes, there was no serious competitor for it.[1] But now the biggest online retailer of music uses an alternative. MP3's "market share" may only have gone down by a few percent. But thanks to iTunes, AAC songs can now be played almost anywhere. Since most people couldn't care less about the format of theit digital music collection, as long as they can play it, MP3 has lost a lot of its relevance. It will probably survive as a household term for digital music, irrespective of the actual encoder, though.
[1] Yes, I do know Ogg Vorbis. It is even my lossy format of choice. But it's never been a threat to MP3.
[1] Yes, I do know Ogg Vorbis. It is even my lossy format of choice. But it's never been a threat to MP3.
"MP3" can mean a lot of things (to some people it just means "digital music"). Quality depends on the bitrate, on CBR vs VBR, on the encoder used, on the playback device, etc.
Personally, I ripped most of my CDs in 256kbs VBR AAC (no DRM), and that's fine for me. But I wouldn't go back to 128kbps MP3s (especially if encoded with something else than LAME).
I'd love to use more OGG Vorbis, but iTunes just plays better with AAC (even if you install the quicktime codec, you can't have the artwork inside the files).
Personally, I ripped most of my CDs in 256kbs VBR AAC (no DRM), and that's fine for me. But I wouldn't go back to 128kbps MP3s (especially if encoded with something else than LAME).
I'd love to use more OGG Vorbis, but iTunes just plays better with AAC (even if you install the quicktime codec, you can't have the artwork inside the files).
> "MP3" can mean a lot of things (to some people it just means "digital music"). Quality depends on the bitrate, on CBR vs VBR, on the encoder used, on the playback device, etc.
Yeah but every time you need to transcode your music to the next format of the day, you're going to lose a bit of quality. And the losses multiply too, they don't just add from one lossy format to one another.
If you store your music in a lossless format, you just have to come back to your lossless copy to create each lossily-generated copy, and you'll only ever have the losses of this transcoding.
Yeah but every time you need to transcode your music to the next format of the day, you're going to lose a bit of quality. And the losses multiply too, they don't just add from one lossy format to one another.
If you store your music in a lossless format, you just have to come back to your lossless copy to create each lossily-generated copy, and you'll only ever have the losses of this transcoding.
He started this series by saying that he thinks vinyl is "the best available method for reproducing recorded music."
I believe that anyone who thinks this has never bothered to learn how audio reproduction works in theory or in practice.
(I say this as someone who loves vinyl for many reasons... just not for its reproduction.)
I believe that anyone who thinks this has never bothered to learn how audio reproduction works in theory or in practice.
(I say this as someone who loves vinyl for many reasons... just not for its reproduction.)
Well, if CD sound quality is good enough..
I was just talking to friends about this the other day. MP3s are convenient, but the fidelity isn't the same as a CD (or the nuisances of a record). Are we lowering our standards as a society, because we don't have time for Xbox, TV, Music, and Family. For the most part music is timeless. Will we get to the point where MP3s will be the only distribution medium and we lose the timelessness of Music? Perhaps we already have.
>Are we lowering our standards as a society
My MP3 player sounds a hell of a lot better than the cheap ghetto blaster on which I played cassettes (often filled with songs recorded off the radio) as a child in the early 1980s.
My MP3 player sounds a hell of a lot better than the cheap ghetto blaster on which I played cassettes (often filled with songs recorded off the radio) as a child in the early 1980s.
"the fidelity isn't the same as a CD"
At what bitrate? I don't think I can tell the difference by 256...
At what bitrate? I don't think I can tell the difference by 256...
What are you listening on? I expect that most people could tell the difference if they listened on a reasonable quality separate CD player, amp and speakers and compared that with an MP3 player plugged into the same amp and speakers. Unfortunately I don't have an experiment I can cite but having tried it myself, I can say the difference is substantial.
You're probably experiencing the placebo effect.
If you care there's software that will let you double-blind test yourself to see what bitrate is transparent for you.
LAME above 165 should be transparent (i.e. sound the same as the source) for most people even with good equipment.
If you care there's software that will let you double-blind test yourself to see what bitrate is transparent for you.
LAME above 165 should be transparent (i.e. sound the same as the source) for most people even with good equipment.
notice what he said:
"quality separate CD player, amp and speakers and compared that with an MP3 player plugged into the same amp and speakers"
"MP3 player"
I'd assume that it means iPod-style device, which probably doesn't have as high quality output as "quality separate CD player". Yes, its comparing apples to oranges, but its different from placebo
"quality separate CD player, amp and speakers and compared that with an MP3 player plugged into the same amp and speakers"
"MP3 player"
I'd assume that it means iPod-style device, which probably doesn't have as high quality output as "quality separate CD player". Yes, its comparing apples to oranges, but its different from placebo
Good point. I haven't listened to high bitrate MP3s through something like an iPod but with a high-end outboard DAC. With a setup like that I believe it could become hard/impossible to tell the difference.
well I think rphillips has proven his point with these replys
I've done ABX testing with cd and mp3s and I always choose cds correctly
those that can't tell the difference i suggest you stop using your ipod earbuds and get something better
"For the most part Music is timeless."
I think music is, speaking generally, a product of its time. Most remaining music from past centuries is the small portion that was timeless. That doesn't mean that the majority of music through history was timeless.
Edited - no need to be rude
I think music is, speaking generally, a product of its time. Most remaining music from past centuries is the small portion that was timeless. That doesn't mean that the majority of music through history was timeless.
Edited - no need to be rude
MP3 causes a hearing disorder called Tinnitus. There's a research paper about it here: http://weltenschule.de/Logologie/info/MP3-Gefahr/MP3-risk.ht...
I don't mean MP3 is totally bad but, I believe it might affect a sensitive person when the bitrate is below 160kbps and badly encoded.
I don't mean MP3 is totally bad but, I believe it might affect a sensitive person when the bitrate is below 160kbps and badly encoded.
If only the problem would be in the transfer medium and not in the MAX GAIN LOUDNESS IS HEAVEN -style mastering.
FLAC is a good means of conveying audio losslessly. Monster Cables are snake oil. These things are not the same.
(And I'm on the older cusp of the "MP3 generation," so this isn't somehow an age thing.)