Mastered for iTunes: how audio engineers tweak tunes for the iPod age | Hacker News Reader

Mastered for iTunes: how audio engineers tweak tunes for the iPod age

arstechnica.com

57 pointsshawndumas44 comments

When I saw this headline on HN I thought to myself "Wow, Apple is finally implementing ReplayGain in iTunes and on iDevices?". Sadly this is not the case. It could do more for the quality of music on iTunes than any increases in bit depth or sample rate.

A quick word on bit depth: 16 bit audio gives us a potential signal to noise ratio of 96dB, which is plenty. The reason why we record in 24 bits is for increased headroom. If I make a 24 bit recording (potential S/N ratio of 144dB, but 124dB with current converters), but my loudest peak is at -18dB, I still have a S/N ratio of 106dB. I can then bring it into a DAW, give it 18dB of digital gain, and my S/N ratio of the 16 bit output file will still be 96dB.

Having a higher fidelity acquisition format than delivery format is not unique to audio. This is why photographers may shoot in RAW, but output to jpeg or tiff and why HD video is edited in 145Mbit/s, but delivered on blu-ray in 40 Mbit/s. It allows for some tweaking in post-production that doesn't come at the expense of not maximizing the potential of the delivery format.

As for bit rates, I think most of the negative perceptions about digital audio come from back when iTunes's default encoding was 128 kbps and ADC technology was still maturing as well as a knee-jerk reaction to lossy compression in general. When I make classical recordings available for web release I use LAME at the V2 setting. Obviously, what bit rate is "good enough" is dependent on program material, but for me that's a reasonably small file size where I don't hear compression artifacts. I know they take some crap for it, but I think Apple choosing 256kbps VBR AAC to be their iTunes plus setting was a good choice.

I don't attempt to fully understand the business implications of these decisions, but "mastered for iTunes" appears to be more gimmick than substance. It may be that Apple is holding on to the high resolution master files for future ALAC release. The engineers quoted in the article talking about having to compensate for AAC's losses are almost certainly talking about 128kbps. A great-sounding recording mastered at 16bit/44.1khz, will still sound great when properly encoded.

Also, Apple's "Mastered for iTunes" technology brief seems to be written with hobbyist engineers in mind. I can't imagine any competent mastering engineer finding it useful. Just further evidence that audio mastering as a craft is on its way out.

When I saw this headline on HN I thought to myself "Wow, Apple is finally implementing ReplayGain in iTunes and on iDevices?".

Isn't that what Apple calls 'Sound Check'?

Apparently, it does the same thing, but is not the same field. Explains why I was getting volume differences on my iPhone even though I had Sound Check turned on (and ReplayGain'd files - made via conversion in Foobar2k from FLAC to AAC using the Nero encoder - need to investigate more)

http://www.vdberg.org/~richard/rg2sc.html http://en.wikipedia.org/wiki/ReplayGain

having done a decent number of mp3/vorbis listening tests, the curve of diminishing returns in audio quality is very steep...as well as the price of the equipment and hearing ability needed to detect it. 85% of users will be unable to discern a well-encoded 192kbps VBR from 320kbps CBR from the original (no matter how high the quality).

a lot of difference which CAN be noticed results from bad encoders and the generic settings used - not strictly related to compression but to other aspects of the psychoacoustic model. classical music encodes better with one set of params, heavy metal with another. there were times when 160kbps has been transparent for me from the original using hi-fi, DACs and headphones. on recording, 24 bit makes a difference vs 16 bit, on output not so much in today's nil-dynamic range records.

telling the difference often requires constantly comparing to the original master where certain aspects sound just slightly different (not necessarily worse), and do not justify a 3x increase in compressed size to get perfect.

the biggest disappointment with a lot of lossy music is bad encoders, encoder settings, poorly (re)mastered originals and non-existent dynamic range at the source rather than a limited quality distribution format.

This. Sample rate, bit depth, and even compression bitrate/settings matter little these days. Between the ultra-loud limiting/compression, poor quality source material and samples, and cheap playback equipment 95% of the target audience will never know the difference between high and low quality delivery formats.

Well put.

I also find that the nature of the musical content matters as well. For me, the most easily detected artifacts occur in complex material (e.g., symphonic or multi-tracked rock), especially when a lot of upper-bass and lower-midrange energy is present. Even so, the defects are slight.

As a disclaimer, I listen with extremely high-end gear and have access to high-resolution source material. It's possible that these effects are undetectable with normal kit.

Do you have any links to 'mp3/vorbis listening tests', or were you rolling your own from your local media?

The reason I ask is because I bought the WAV version of the most recent Radiohead album and tried to blindly discern a difference between them and the MP3s and failed. I've been to plenty of loud concerts, so probably have degraded hearing, but I'd love to try some more tests.

unfortunately it's been quite some time. most of the tests done were to compare different encoders and settings for different types of music. some i did encode myself from CDs. search around online, it might take a bit of digging, but there's some stuff around still.

here's a quicky: http://www.noiseaddicts.com/2009/03/mp3-sound-quality-test-1...

i was surprised that more than 50% got this one wrong. i got it correct even on my single shitty LCD-attached speaker at work at safe-for-work volume levels. the difference between 128 and 320 can be discerned pretty much 97% of the time in all but the quietest, gentlest of music.

128 sounds flat, like you're listening through a wall, but unless you have the higher quality version for reference, even this can be hard to tell. kind of like "perfect pitch" http://en.wikipedia.org/wiki/Absolute_pitch

Thanks! I guessed wrong for your link.

"It was my quest to make the AAC files sound as close to the CD as possible..."

"I can see that it has the potential for making the AAC encoded masters sound truer to the CD and LP versions..."

This is nonsense. If encoding the same data that's on the CD doesn't produce the AAC file that sounds the closest possible to how the CD sounds, it would obviously be the encoder that needs to be fixed, not the source data that needs to change.

I refuse to buy music off iTunes and generally end up having to pirate artists' music if it's not available to buy in a lossless format.

I don't understand how bandcamp can offer me a choice of whatever format I want, yet apple still expect me to pay the cost of a CD for an inferior substitute. The galling bit will be when they do start selling in ALAC, they'll probably charge more for it - at which point I'll still refuse to pay for anything from iTunes..

EDIT: Downvoted why? Because I buy music off places where it's offered in good quality and not from places where it's not? This is not a hard problem to solve, hell bandcamp solved it ages ago - as did what.cd....

The galling bit will be when they do start selling in ALAC, they'll probably charge more for it - at which point I'll still refuse to pay for anything from iTunes..

On one hand, Apple did charge 30 cents to upgrade to the higher quality AAC from the DRM'd versions. On the other, they're providing even higher quality versions of any song they can now, for free.

(Not lossless, but, 256kbps AAC is far beyond the point where most people can tell the difference.)

Exactly, they're not lossless. I don't know if they're deliberately doing it, but they've decided that audiophiles don't matter, much as they always do (see people who used to use expresscard on MBP's, film makers etc etc), and just plain ignore people who want to purchase high-quality content. Thankfully bandcamp exists and I buy off there all the time.

Do you really think that most people care about their music being lossless? Let alone know what "lossless" even means? Most people want more music on their phones and iPod's, not less. Now, and not later in the afternoon after they've downloaded something 5-10x the size for a quality difference that they cannot notice.

Most people don't fit into your niche usage.

Apple is doing the best they can for the majority of the people, while still making the labels happy. Yes, it sucks if you happen to be in the minority[1]. But, you can't claim Apple isn't making a good business decision.

1. I speak as an iPhone and Mac developer who is used to Apple doing the best thing for the user, and not the developer.

I wasn't questioning apple's business decisions, I was questioning their emerging monopoly on digital content that is making the digital market even more homogenous than the physical market was. At least in the physical market anyone could open up a music store because there were independent distributors whose sole incentive was to move product, now the incentive is to extract margins on product as Apple and Amazon negotiate directly with the labels and shut out competitors from being able to properly compete because they can't get the content. Some platforms are doing ok in this (bandcamp) but most aren't as the reality of music retail is not the long tail but the centre of the bell curve.

I'm also an Apple developer.

If you really cared (about the artist), you'd buy music in CD form and rip it, wouldn't you? Does not buying in iTunes mean that pirating is the only other option?

The music lobby has decided that (in some cases) iTunes is the only digital retailer they'll supply, so actually yes, it sometimes does.

Sums it up nicely: http://theoatmeal.com/comics/game_of_thrones

Uh. Can you give an example of music that you want to download, but can't buy on physical CD?

There's plenty of older music that is not available on CD anymore (or has never been), but can be purchased via iTunes. To give an example: most of Sam Rivers' albums. Or Old and New Dreams (ECM).

A random recent viral example: http://amzn.com/B003MV33JY

Video game soundtracks tend to be hit hard by this too -- roughly half the soundtracks I want are download only, and about 2/3 of the ones available on CD are Japan only (and frequently out of print).

Minus the pirating bit, this; a thousand times this!

This is one of the reasons that I'm actually sad to see the CD go. When producers mastered for CD, you got the best quality that they could produce or that the CD could hold. (Whichever came first.) Now that the CD is fading into irrelevance, theres the worry that digital distribution will make the average quality of a track worse instead of what it should be, better than CD. (Though at this point I'd take parity with CD quality.) Yeah, I get it, lossless tracks take up a lot of memory. They should still be an option to purchase though. Lossless is the default state for recordings anyway.

I don't even have the equipment required to hear the difference. It just doesn't sit right with me to support a business that blatantly panders to the LCD at the expense of everyone else.

"Mastered for Itunes" Just reaffirms this for me.

I can’t tell the difference. I really can’t. Why do you think you would be able to?

How much have you tried? At lower bitrates (160 and under, mp3) I wouldn't notice a sound quality issue so much as I noticed that I consistently became fatigued faster than when I listened to CDs. I honestly don't know if the same holds true of higher bitrates, since by the time I started encoding those, I was listening less and the convenience outweighed potential irritation.

For reference, here's the Wikipedia article for the Loudness War (http://en.wikipedia.org/wiki/Loudness_war). In my opinion, hip hop producers like Timbaland and Dr. Dre really ratcheted up the effective loudness of songs by introducing heavy compression.

It's the (sad) reason why the Beatles' albums have been "re-mastered" and re-released so many times over the years -- our 21st century ears are so used to compressed music that old Beatles albums sound too quiet to us.

> our 21st century ears are so used to compressed music that old Beatles albums sound too quiet to us.

I wouldn't mind so much that an album is quiet, if any MP3 player actually allowed me to turn it up to a reasonable level (letting me sacrifice the listening experience on my own terms, when and if applicable.) Why is VLC the only* piece of software to combine volume level and compression-gain into a single output slider?

* Amusingly, the iPod's software knows how to apply compression just-in-time--but it only does it if you've applied a "volume adjustment" pragma to the song in iTunes. And then whenever the song comes up, you have to turn down your volume slider from 80% to 30% to avoid having your ears fall off, because "loud" on desktop speakers is a very different thing than "loud" on in-ear headphones.

That's because music listening has traditionally been about source reproduction and I hope this remains the case. Sure you can apply compression or whatever if you want, but you're going to destroy the sonic quality of the recording even more, and, as an audio engineer, I hope software manufacturers keep "limiting" your freedom in this regard (sorry for the pun...).

Reproducing source audio level doesn't help when you're e.g. listening to something on a plane, with a large amount of background noise, and you want to actually hear the damn thing. Users' opinions count for something too.

Confusingly, the article is about two different types of compression. The type of compression you're talking about is compression of dynamic range, which has been an issue since before mp3 players.

Then there is lossy data compression, which also distorts sound, but in different ways. It tends to make high end percussion sound like it's being filtered through bubbly water, for example.

Then they were also taking about increasing sample rates and bit depths, which is another question altogether, and (I think) much less significant than the other two issues.

While the compression is obviously a problem, I am always just as worried about the fact that the mixes are for the lowest common denominator. For years most engineers I knew would mix so that the song was listenable on NS-10's (which are complete crap). Now I would assume the mix is probably tested more on a set of white earbuds.

A factor in the prevalence of compression in popular music that is not often mentioned is the effects of the music production technology, specifically the cheap akai samplers that were used a lot in the 1980s and 90s. They only had a 12 bit sample rate and so had a very small dynamic range. This led to a fetishisation of the "chunky" sound it imparted on samples, especially drums. Then producers seeking this effect further compressed their samples to make them sound even chunkier. Here's an example from 1995: http://www.youtube.com/watch?v=3fkWMrw4nFY#t=1m39s and another from 1993: http://www.youtube.com/watch?v=zu9Ml51vDrM

In these scenarios, over-compression is a legitimate artistic choice: it's meant to sound like that. The problem is obviously when this expectation bleeds out into the rest of the music industry and is forced upon other types of artists by their labels.

In my understanding, 44.1 kHz was chosen because it's twice the maximum of human hearing (22 kHz), and thus you can reproduce all audible sounds without worrying about aliasing (as per the Nyquist-Shannon sampling theory). What is the point of going higher?

When downsampling or recording, and when playing back 44.1 kHz audio, the data must be low-pass filtered to eliminate aliasing effects.

But filters aren't perfect. Even a decent low-pass filter (say 3rd order Butterworth) requires an order of magnitude of bandwidth to drop the output 60 dB (practically inaudible). This means that with a Nyquist limit of 22 kHz, you're either attenuating everything above 2.2 kHz (the "knee"), or you're letting some aliasing noise through.

With a 192 kHz sampling rate, the filter's knee can rise to 9.6 kHz, and the stuff between 9.6 kHz and 20 kHz won't be appreciably attenuated.

It's important to note also that this attenuation can't be fixed by simply boosting the high end -- filters are linear, so such an adjustment (with an equivalent order filter) would merely cancel out the low pass filtering and reintroduce aliasing noise.

(Edit: I am not an audio engineer but I have a strong signal theory background. So actual audio engineers please feel free to correct me.)

You have the right idea, but the wrong numbers. Terribly, terribly wrong numbers, it's quite clear you're making those numbers up. No filter for an audio ADC would ever have a cutoff as low as 2.2 kHz, not even if it were for a telephone. (You have a strong signal theory background? No offense, but really?)

When you use a 44.1 kHz sampling frequency, any frequency above 22.05 kHz will be "aliased" and recorded as a lower frequency. This sounds incredibly nasty, like the sounds you'd get out of a broken Commodore 64. So you have to remove frequencies above 22.05 kHz in order to get a clean recording. But human ears can hear up to 20 kHz or so depending on age (e.g., NTSC TVs with cathode ray tubes have a 15 kHz horizontal refresh which drives me nuts, but my parents can't hear it at all).

The trick then, is to design a filter that will let the 20 Hz - 20 kHz band through while stopping everything above 22.05 kHz. We call 20 Hz-20 kHz the "pass band" and 22.05 kHz and above the "stop band". We don't really care what happens to the frequencies between the pass band and the stop band: the range from 20 kHz to 22.05 kHz which can't be heard well enough to be worth preserving and doesn't cause aliasing so it doesn't need cutting out. This is difficult because the stop band is only 1.1x the frequency of the pass band -- for you musicians out there, that's less than the difference between C and D on the western scale. (Just think for a minute: design a filter that lets middle C through, but completely filters out the D above.)

Heavens no you wouldn't use a Butterworth filter for such a task. We want an elliptic filter, probably. 3rd order is no good either, it won't give a sharp enough cutoff. 8th order is better. This will get you a cutoff around 20 kHz with something like 60 dB attenuation at 22.05 kHz. People need a lot of these filters, so you can actually go out and buy a 20 kHz low-pass 8th order elliptic filter as a monolithic chip.

Let's suppose you chose a 48 kHz sampling rate instead. Now the stopband starts at 24 kHz instead of 22.05 kHz. It sounds like a small difference (22.05->24 kHz cutoff) but it's actually a factor of 2 (2.05->4 kHz transition band). This means that with the same components, you can get 80 dB or more attenuation in the stop band.

Now go to 96 kHz. You have to design a filter that rolls off between 20 kHz and 48 kHz. That's easy peasy, and you can reduce the ripple, increase the attenuation, maybe reduce the order (affecting noise) and make all sorts of design tradeoffs that are much easier.

Now think about 192 kHz. What's the point? What does 192 kHz get you that 96 kHz doesn't have? It's already easy enough to design a very nice system at 96 kHz. I think 192 kHz is a bunch of bunk as far as audio is concerned.

That's recording. Now let's talk about playback.

Playback is very similar, everything goes in the opposite direction. You start with a digital signal, convert it to analogue, and put it through a low-pass filter. The aliasing noise is still there, except instead of reflecting high frequencies to low ones, it reflects low frequencies to high ones. So you get the same trade-offs.

The difference is that playback requirements are not as difficult as recording requirements. In particular, the required SNR of a playback system is lower than that of a recording system. I think 48 kHz is fine for playback.

The problem is these stupid 192 kHz systems have backers with big names who never bothered to do proper double-blind tests to figure out if the difference is actually perceptible. You can even get a 384 kHz system these days, which would be overengineered for dogs and is more than good enough for bats.

If I understand you correctly, going from 44.1kHz to 48kHz would be worth it on the playback side of things?

That wouldn’t seem like it would be all that hard to do. CDs are a legacy format now and AAC files don’t care about the sampling rate (if you don’t want to go beyond 96kHz).

What’s stopping that? Have all the audio engineers pipelines that are only capably of outputting 44.1kHz? (I imagine someone at Sony Music sitting in a dark room and ripping CDs all day – probably not true but a funny enough picture.)

Then again, after doing a blind test (256kbps AAC, CD) and being unable to tell the difference (yeah, I know, that’s not the same as a difference in sampling rate) I’m skeptical of all supposed small improvements in audio quality on the playback side.

Last year I bought a 14 input, 24 bit, 44.1/48/88.2/96kHz audio interface for $100. Granted, it's not supposed to cost so little, no one cares about MSRP but it runs twice that on Amazon, more at Guitar Center, and typically no less than $170 on eBay. I just happened to get a good deal on Craigslist.

Anyway, the issue isn't one of hardware limitations; Call it old guys fearing technology, call it the Red Book cartel, call it whatever you want, it's inertia.

192 kHz is a "bigger number = better" marketing ploy imo. Dont get me started on interpolating "240hz" TV's.

The move from 16 bit to 24, however, is significant. And not only to headroom.

As far as SR systems, I really like how 96 can sound. And I'm more comfortable intuitively with considering the limit of human hearing somewhere closer to 24 or 25k than 22.

It's also worth pointing out that for the last 20 years, most audio converters have been of the "oversampling" type, one of the main advantages being that a much gentler anti-aliasing can be used, with fewer audible artefacts down into the audible frequency range.

Hence the disclaimer :) thanks for the better numbers; I didn't realize such high-order filters were common in consumer equipment.

Partly yes, but also because of the data requirements of the format (CD). Bit depth is important too because it defines the number of locus points available to you when measuring a wave, so the more bit depth you get the closer the digital representation of the wave becomes to the original analogue source.

This is a point that often gets lost in these discussions: the whole point of digital audio is mean't to be a more efficient capturing and distribution of analogue source, as such the more information you can capture about each instance in time the closer your digital reproduction can get to the original analogue wave.

To anyone who thinks they can here the difference between 48 and 96 sample rates, I've got some Monster cables that will make it sound even better. Neil Young probably can't hear anything over 10 kHz, yet he's getting all worked up over 256 Kb ACC? None of this makes any sense.

I suppose as I don't have gold coated cables, I can safely ignore these issues for now...

Regarding the headline image for that article -- where'd they find a MacBook with bright red keys? Is it even a MacBook?

Keyboard cover.

http://www.amazon.com/IVEA-Keyboard-Silicone-Aluminum-Unibod...

I've got the pink one, and I get groups of them to have on hand for friends - they're generally less than $2 from amazon (the one above is 26 cents, but you've got shipping on that in some cases too).

Thanks!

Reply on news.ycombinator.com