Neil Young speaks out against excessive compression

[projectMalamute]

You should really take a closer look: Invalid Link Removed

And please, relax a bit. This has nothing to do with making shoddy recordings.

alright, I did.

I see statistically significant preferences for one thing over another all over those listening tests.

If this is a solved problem they shouldn't be hearing any differences.

and, one more time, that's not even the point. almost, maybe, could be, under some circumstances, for some people, as good as 30 year old technology is crap.

Why choose to use tools that you know are marginal when better tools exist? The answer is of course convenience, which is fine. Cassettes were convenient too, and I have a bunch of them.

 

[Febs]

I see statistically significant preferences for one thing over another all over those listening tests.

If this is a solved problem they shouldn't be hearing any differences.

Thanks for taking the time to take another look. A few points in response to this.

First, these tests were conducted in many instances in conjunction with the people who were actually developing lossy codecs, not to prove that lossy compression was a "solved problem," but to identify issues so that they could be addressed in the development of the codecs, which was (and undoubtedly still is, though I haven't followed it closely in a few years) ongoing.

Second, if you look at the results of most recent public test, which is the October 2008 128 kbps listening tests, the ultimate conclusion was that there was a statistical tie among all of the codecs being tested:

(Note that in the test, l3enc, which was the first mp3 encoder from 1994, was not among the codecs being tested but was included as a "low anchor," i.e. a known inferior encoder that was included as a check on the results).

Third, although it's not immediately obvious from the results page, my recollection from following the tests at that time was that the reason the October 2008 test used 128 kbps encoders was because by that point, participants generally could not differentiate among the codecs at higher bitrates and therefore, comparison among them would not have yielded statistically significant results.

and, one more time, that's not even the point. almost, maybe, could be, under some circumstances, for some people, as good as 30 year old technology is crap.

Why choose to use tools that you know are marginal when better tools exist? The answer is of course convenience, which is fine. Cassettes were convenient too, and I have a bunch of them.

Sorry, I'm not sure what you're getting at with the sentence of this quote. With respect to the rest of it, I think that you are unnecessarily (for the purposes of this discussion, anyway) conflating the recording process with the choice of medium used to deliver the result of that process to the consumer. As to whether those are "marginal," are they really marginal if there is no audible difference between them and the "better" options, and the "better" options come at some cost (be it storage space, bandwidth, etc.)?

 

[swspiers]

Y'all do realize that if the results of the majority of tests are statistically inconclusive, that none of our claims are validated. Which is why the debate rages.

If someone were trying to prove that 'all codecs sound the same', the results for the most part do not support that hypothesis.

 

[PJRL]

I don't understand his complaint. He sounds like an old crank to me. Does anyone actually think that Neil could tell the difference between a lossless digital recording and a 256k AAC file? Please. Even if he had the most sophisticated playback equipment on the planet he wouldn't be able to tell the difference. And neither could I. Only the most golden-eared audiophiles could even hope to do so. Old Neil probably has enough hearing loss to preclude him ever successfully distinguishing the two.

Basically all I heard when I read his remarks was "get off my lawn!"

And the bit about downloading an uncompressed song in 30 minutes? He's on drugs if he thinks that's the kind of "innovation" anyone wants.

Yes I think he can.

I remember when the first CD's came out. I loved the music but hated the cold dry sound. I remember Neil being one of the fist ones to talk about it back when digital and CD's were becoming the norm. It has improved since then but I still think vinyl and analog sounds better. If an uncompressed song sounds better I would would spend 30 minutes to do it once in a while.

 

[longfinger]

I think there is a problem with your source that is merely being exposed by the AAC compression, rather than a problem with AAC compression itself.

That does not make sense to me. What makes more sense is that the original file is raw, simple and untouched enough for people to be able to hear differences in sound brought out by AAC compression, that cannot be heard in files of sounds that are already altered with various effects, plugins, overdubs and mixing.

 

[eqvolvorama]

That does not make sense to me. What makes more sense is that the original file is raw, simple and untouched enough for people to be able to hear differences in sound brought out by AAC compression, that cannot be heard in files of sounds that are already altered with various effects, plugins, overdubs and mixing.

If that's the case, the test should be easily reproduced in blind fashion, using source program material that's as raw as you say. Let's see the results! The time for sighted tests, anecdotes, and theories is over!

 

[MonkeyBass]

I don't understand his complaint. He sounds like an old crank to me. Does anyone actually think that Neil could tell the difference between a lossless digital recording and a 256k AAC file? Please. Even if he had the most sophisticated playback equipment on the planet he wouldn't be able to tell the difference. And neither could I. Only the most golden-eared audiophiles could even hope to do so. Old Neil probably has enough hearing loss to preclude him ever successfully distinguishing the two.

Basically all I heard when I read his remarks was "get off my lawn!"

And the bit about downloading an uncompressed song in 30 minutes? He's on drugs if he thinks that's the kind of "innovation" anyone wants.

Um... I can tell. But, I was trained by Dolby Labs. But despite that, every good audio engineer I know can hear the difference.

 

[projectMalamute]

Yes I think he can.

I remember when the first CD's came out. I loved the music but hated the cold dry sound. I remember Neil being one of the fist ones to talk about it back when digital and CD's were becoming the norm. It has improved since then but I still think vinyl and analog sounds better. If an uncompressed song sounds better I would would spend 30 minutes to do it once in a while.

There is no reason for it to even take 30 minutes. 192/24 is the highest resolution commonly being used today, that is 70MB per stereo minute. It's a big file, don't get me wrong, but you are getting the exact same data the artist was hearing when he decided that the music was finished.

Think about that for a minute. Technology has progressed to the point that you can listen to the same media that was in use in the studio. That's the equivalent of being able to hear the 1/2" 2 track master of Thriller. and it is orders of magnitude cheaper and more convenient than even listening to Thriller on cassette.

Instead we are attempting to find the absolute minimum level of fidelity possible before people start to complain, a race to the bottom. I genuinely cannot wrap my head around how anyone who cares about music would think this is a good idea.

 

[eqvolvorama]

Um... I can tell. But, I was trained by Dolby Labs. But despite that, every good audio engineer I know can hear the difference.

Prove it. I think you are exaggerating your credentials, past experience, and hearing perception. Prove me wrong.

 

[scottfeldstein]

Um... I can tell. But, I was trained by Dolby Labs. But despite that, every good audio engineer I know can hear the difference.

Ok. But is there any actual proof that what you're asserting is true?

 

[scottfeldstein]

we are attempting to find the absolute minimum level of fidelity possible before people start to complain, a race to the bottom. I genuinely cannot wrap my head around how anyone who cares about music would think this is a good idea.

No. I'm not trying to find the absolute minimum level of fidelity before people start to complain. I'm trying to find the absolute maximum level of fidelity that human beings can actually detect so that we don't waste our time, money and effort chasing mirages. I don't see why everyone doesn't wish to join me in this.

 

[DWBass]

Instead we are attempting to find the absolute minimum level of fidelity possible before people start to complain, a race to the bottom. I genuinely cannot wrap my head around how anyone who cares about music would think this is a good idea.

This is old thinking. This was because 10 years ago, there were no iPods or any mp3 device that had more than 32mb or 64mb of memory so make the file as small as possible to fit as many songs as possible. This is no longer the case! In addition, 10 year old software did not achieve the same sonic results as today's software!

Lastly, the % of audiophiles to average music listeners is vastly different! Could be 100 out of 1,000,000,000 or less! Ever hear a car go by and all you hear is boom boom and not much else? That's the average music listener with no discerning ear! We are different. I think most of us here do care about audio quality but some more than others. Portable music devices are what they are.

 

[seamonkey]

Um... I can tell. But, I was trained by Dolby Labs. But despite that, every good audio engineer I know can hear the difference.

And so it's good audio engineers.
It can be learned to hear the difference?

It's like training to learn to see the blind spot in your eyes. What for? The blind spot can is ignored by most. Don't bother spending time to learn to sense something the majority can't unless there's a real benefit to it. Measuring tools can accurately measure the difference better than humans could.

Yes is A/B test maybe golden ears can hear a difference in raw versus Any CODEX. In the bell curve - No, people don't hear a difference.

And it's proven again and again that people do not have memory for this. If you listen to A and then a minute later listen to B, then a minute later listen to X - you are not going to remember what X sounds more like - A or B. Change blindness. People in the bell curve are blind to huge changes when you add in time or distractions between the samples.

 

[projectMalamute]

No. I'm not trying to find the absolute minimum level of fidelity before people start to complain. I'm trying to find the absolute maximum level of fidelity that human beings can actually detect so that we don't waste our time, money and effort chasing mirages. I don't see why everyone doesn't wish to join me in this.

We aren't there yet. Even 192/24 on the highest quality monitoring imaginable doesn't sound as good as standing 10 feet in front of the Julliard String Quartet.

We don't need to be downgrading listening standards we need to be upgrading them.

This all assumes we care about how music sounds. I don't even mean that pejoratively. News reporting, telecommunications, incidental music for video games, anything temporary or disposable; in all of those situations and many more excellence in fidelity is probably not that big of a deal.

If the goal is a permanent record of music someone, somewhere might care about we should be providing it in the highest fidelity we can achieve.

 

[Rockman]

Prove it. I think you are exaggerating your credentials, past experience, and hearing perception. Prove me wrong.

I believe he can hear a difference, I sure can.

Heres the real problem, Neil Young has zero technical knowledge and has no idea about what he's proposing. He proves this by saying "some rich guy", and that a file would take hours to download. The broadband data rates available now are pretty amazing and there is plenty of signal analysis that proves the minimun data rate (and thus size/second) of any sinusoidal source to make it completely lossless. I haven't been on a network that got less than 10MB/sec, which transfers to 35GB of data, or only about 30 song PER TERABYTE. Processors and hard drives generally have trouble dealing with huge files like that, that is why a huge part of computer engineering/programming/research is ways to break things up or shrink file size. This is stuff that any sophomore electrical engineering student would understand. Also these types of thing aren't developed by "some rich guy" they are developed by some very hard working super intelligent electrical engineers and computer scientists, the fact he thinks they are developed by "some rich guy" shows his ignorance.

If you want to know the maximum audio-bitrate(fidelity) then you simply need to talk to any audiologist, or even better a microbiologist that specializes in audio processing. All these things have been and are studied.

The only thing that he said I agree with is prodcuts like "Beats" try to make a crappy source sound good. Also his comment about piracy, which is a surprising comment to hear from such an old fashioned man.

In closing: Neil has a valid point about the standard compressions used in todays technologies but his "solutions" are insane and generally impossible.

 

[sevdog]

Man, this thread had a short stint of reasonable and technical discussions but has gone right back to "I can't hear a difference so no one can. Prove it or I'm right."

I can hear a difference. There are others who can too.

When you stare at the horizon YOU can't see the curvature of the earth so you claim the world is flat.

I don't know why this thread has gotten under my skin. Maybe it's the know it alls that won't even give a "maybe there's a possibility that I'm wrong."

I hope Neil Young gets his way and you all have to start buying music in a format that is higher def than you want it to be. C ya.

 

[scottfeldstein]

We aren't there yet. Even 192/24 on the highest quality monitoring imaginable doesn't sound as good as standing 10 feet in front of the Julliard String Quartet.

I see what you did there.

Instead of comparing one listening format to another, you're comparing a listening format to something that isn't recorded in the first place.

So, when you invent a playback mechanism that is capable of sounding exactly like being there, then I guess we'll need a recorded format which can equal it.

We only need recording methods, playback technologies and audio storage formats that meet our ability to appreciate them. And no more.

I don't need a television that has extra definition in the infrared spectrum. I don't care how many TV aficionados claim they can tell the difference. It's a fact that people can't see this material and therefore, we don't need to make or buy televisions which do this.

If someone can show me empirical data indicating that human beings can distinguish between what Apple sells in their stores from the master recordings, then I'll join you in asking for better quality. And I'll keep on asking for it--right up to the point where nobody can distinguish it anymore.

I say this last bit with an acute awareness of people's seemingly inexhaustible capacity for self-delusion. People believe all kinds of things. 9/11 was an inside job. Vaccines cause autism. Echinacia cures colds. You can hear the difference when you use expensive audio cables. They're sure these things are true, even though they aren't. It gets under my skin.

And I do detect a whiff of this phenomenon in this discussion. I am the kind of person who cares a lot about whether the things he believes are actually true. I want to know what the reality of it is. Can anyone hear the difference or not? it's a pretty easy thing to prove. The fact that no one has is a bit of a red flag for me.

 

[eqvolvorama]

Man, this thread had a short stint of reasonable and technical discussions but has gone right back to "I can't hear a difference so no one can. Prove it or I'm right."

I can hear a difference. There are others who can too.

When you stare at the horizon YOU can't see the curvature of the earth so you claim the world is flat.

I don't know why this thread has gotten under my skin. Maybe it's the know it alls that won't even give a "maybe there's a possibility that I'm wrong."

I hope Neil Young gets his way and you all have to start buying music in a format that is higher def than you want it to be. C ya.

You are misrepresenting my argument. I am not questioning that you can hear a difference because I cannot. I question it because:

1. There are no published tests indicating people can tell the difference we have proposed (256kbps AAC vs. Lossless or high res)
2. Proving one could is so trivially easy that your reluctance to even try is suspicious.

I owe your claims no more automatic respect than if you claimed to have seen a UFO land in your backyard.

 

[3506string]

I can hear a difference. There are others who can too.

Hear the difference in what exactly? What Neil is proposing is a new format that is superior to what is already available. What people are questioning is whether the human auditory system can even perceive a difference between lossless sources and the higher quality lossy formats that are currently readily available and have been for some time. I have not seen significant evidence that there are many people that can repeatably hear a difference in these sources, let alone enough that would warrant an entire new probably proprietary(if apple has anything to do with it) format.

One problem in areas like these is that people tent to trust their ears. It has been proven time and time again that our auditory system isn't great at giving us unbiased information on what are ears are perceiving. If this wasn't so there would be no need for double blind testing.

 

[projectMalamute]

I see what you did there.

Instead of comparing one listening format to another, you're comparing a listening format to something that isn't recorded in the first place.

So, when you invent a playback mechanism that is capable of sounding exactly like being there, then I guess we'll need a recorded format which can equal it.

We only need recording methods, playback technologies and audio storage formats that meet our ability to appreciate them. And no more.

I don't need a television that has extra definition in the infrared spectrum. I don't care how many TV aficionados claim they can tell the difference. It's a fact that people can't see this material and therefore, we don't need to make or buy televisions which do this.

If someone can show me empirical data indicating that human beings can distinguish between what Apple sells in their stores from the master recordings, then I'll join you in asking for better quality. And I'll keep on asking for it--right up to the point where nobody can distinguish it anymore.

I say this last bit with an acute awareness of people's seemingly inexhaustible capacity for self-delusion. People believe all kinds of things. 9/11 was an inside job. Vaccines cause autism. Echinacia cures colds. You can hear the difference when you use expensive audio cables. They're sure these things are true, even though they aren't. It gets under my skin.

And I do detect a whiff of this phenomenon in this discussion. I am the kind of person who cares a lot about whether the things he believes are actually true. I want to know what the reality of it is. Can anyone hear the difference or not? it's a pretty easy thing to prove. The fact that no one has is a bit of a red flag for me.

So the current state of the art is good enough? Rather than pushing for higher fidelity with each advance in technology we should arbitrarily decide that a CD is good enough, and just focus on convenience?

Tons of music was recorded at a higher fidelity than the playback technology of it's day could reproduce. That's why I get to hear Fritz Reiner conduct Bartok with pretty awesome sound.