Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Mastering engineer Bob Katz: The Loudness War Has Been Won (Read 22951 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Mastering engineer Bob Katz: The Loudness War Has Been Won

“The debilitating loudness war has finally been won,” said mastering engineer Bob Katz on the eve of the Audio Engineering Society (AES) Convention in New York City. The last battle will be over by mid-2014.”
http://www.digido.com/forum/announcement/id-6.html

Perhaps he's being too optimistic, but if iTunes Radio manages to become a reference for how modern music should be mastered then there's hope. Loud, over-compressed music will be reduced in level and sound worse than more dynamic music.

The news is that iTunes seems to have the power to enforce new standards, something the mastering engineers can't manage on their own. In 2009 I attended a convention at Galaxy Studios (Belgium) about loudness. It was demonstrated that Orban Optimod equipment, which is used in many/most radio stations (to meet legal transmission standards), makes over-compressed music sound worse.
Apparently this hasn't provoked mass complaints from radio listeners. In that sense it's good that iTunes takes the lead.
Mastering isn't cheap and it will cost extra time to make separate versions for iTunes and CD. Let's hope clients will choose the less-compressed iTunes version.

Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #1
Although I would prefer that Apple not be portrayed as our only hope (more than normal), if iTunes Radio can set a good example for everyone else, that can only be a good thing on balance.

Thanks for posting!

Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #2
I think Spotify has about as much pull now as iTunes. (I initially thought iTunes Radio would compete directly with Spotify but the features are way too different.)

...but in any case I think we should all reconvene here in mid-2014 and hopefully celebrate the victory (or heap scorn upon Bob Katz and declare him a false prophet.)
The Loudness War is over. Now it's a hopeless occupation.

Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #3
Now if only iTunes featured lossless downloads...
Project Leader of DDResampled

Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #4
Whether the method of delivery is lossless or lossy is totally irrelevant to the question of whether or not a particular master is overcompressed. A file downloaded from the iTunes Store in AAC at 256 kbps is highly unlikely to sound any different from its lossless counterpart, and perceived compression is no different in this regard. So, did you actually think that detail mattered, or did you just want to vent your opinion about something unrelated?

Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #5
Is it correct to assume that Sound Check is the main ingredient which might end the loudness war? Then ReplayGain and David Robinson should be heralded as our saviours just as much as Apple.
It's only audiophile if it's inconvenient.

Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #6
I though Spotify already did that?

Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #7
Apple also has these Mastered for iTunes albums which however sound terrible to my ears and are also heavily compressed, so not sure how important iTunes Radio is.

Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #8
Despite what people may want to believe, Mastered for iTunes does not discourage the use of dynamic range compression.  I'm betting some of the issues attributed to sound quality may have to do with the moronic idea that one can equalize prior to lossy encoding in order to compensate for lossy artifacts.

Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #9
I gained the impression it's a control thing. Generalising, mastering engineers are rarely happy with the effect that lossy encoding has on their finely tweaked masters. When MfiT was first announced there was a lot of discussion on how to tweak the sound pre-iTunes to make the result sound the way they wanted. There was surprisingly little discussion on the good habits encouraged by MfiT, such as keeping the peak levels low enough to avoid clipping during the decode. I hope this means that such habits are so well ingrained that they didn't need discussion, but I didn't really get that impression.
Regards,
   Don Hills
"People hear what they see." - Doris Day

Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #10
If this means that a loudness level (similar to ReplayGain or a RMS value or a EBU R128 related value) is being transmitted along with the meta data then I'd call this a big win.

Add to this fact that traditional radio stations have begun adjusting to EBU R 128 this is even more awesome. http://tech.ebu.ch/docs/techreview/trev_20...ess_aarseth.pdf (on page 7 it says that -15 LUFS is the standard target for digital radio in Norway)

And with iTunes Radio and iTunes (and Spotify?) being in the same area loudness-wise, I'll have to agree with Bob's statement. The Loudness war is over, now we just need to wait for the rest of the industry to get the memo. I look forward to more artists to play with a extended headroom (compared to previously).

Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #11
When MfiT was first announced there was a lot of discussion on how to tweak the sound pre-iTunes to make the result sound the way they wanted.

Exactly.  A sure-fire way to bolster your credibility!

Generalising, mastering engineers are rarely happy with the effect that lossy encoding has on their finely tweaked masters.

Based on properly controlled double-blind tests, of course.

Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #12
Kees has the intent of my evangelistic effort down pat! Many people have claimed that I said "the loudness wars have been won". But that is not what I said, and Kees has it correct. Of course we have a long way to go. But the last big battle has just been fought and there is a winner: All of us won.

Think of it this way:

---- The audio Berlin wall has now fallen, with the introduction of iTunes Radio. This was the turning point of the war, we now can see the light at the end of the tunnel. Communism is dead. Freedom is now here. Freedom is headroom.

----Mastering engineers, producers and mixing engineers need to tell their clients, starting right now, to turn on Sound Check, and to audition their references through Sound Check so as to hear what they will sound like "on the Radio" against the "competition". This collaborative act alone among producers will hasten the end of the loudness war. We also should encourage them to play their i-devices in the car with Sound Check turned on. And leave their CDs at home.

----Even if no one follows the above practice, the war has still been won. It will just take longer to end as more useless battles will be fought amongst ignorant "peak maxterizers" who think they can fool their clients for a few months more. Resistance is futile. 

----So the war is won, eh?  Who won? The answer is:  Everybody wins. If one engineer or artist wants to apply more compression than the taste of another engineer or another artist, they can still do it. But their product will no longer sound louder than the other artists'. As Ralph Kessler pointed out at the Convention, "Pink" and "Pink Floyd" are now at the same loudness.

---During the interim before it all clears up, the wounded soldiers will have to deal with a lot of drek. One suggestion is to make two masters, one for CD and one for the loudness-normalized media. That will be a pain in the butt, and will cost clients money or if we don't charge them, it will cost us time.

----And there's more but you get the idea. It is true that I have been prophesizing the end of the war, after 33 years of digital war there is finally light at the end of the tunnel. That's what we should rejoice over! Perhaps my announcement was premature because iTunes Radio has only made it in the U.S. so far. Perhaps I should revise my "mid-2014" estimate to later. But my goal was not to be held up as a prophet, but rather to be an evangelist, to get the troops jumping for joy and to clear our heads from this terrible psychological depression that has been lingering for so long. iTunes Radio is a genuine turning point. Let's rejoice over that and not try to say, "Oh, Bob Katz was wrong, see, it didn't happen by mid-2014". Instead, keep on educating your clients, artists, and producers why the war is now over and everyone work together and work forward. Then the war will still be over much sooner than we all thought it would ever be, maybe by 2015... if you get my drift.

----Someone praised Spotify here?  According to Kessler's research, Spotify's target level is much too high! And its algorithm is extremely inaccurate. We need to get that fixed. It would be ironic and tragic if loudness normalization happens throughout Internet radio but there is still a loudness war between stations! So we have to get everyone down to at least Apple's target level.

Hope this helps!

Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #13
Now I just hope that record companies and music vendors won't force "hi-res" music down our throats. My big fear is that the rise of "hi-res" formats will coincide with the reinstating of dynamic range, and everyone will attribute the improvement in sound to the new technology instead of the "new" way of recording, mixing and mastering. Then we all have to waste humongous amounts of bandwidth and disk space on ultrasonics and the possibility to store the sound of a H-bomb exploding at 1 feet distance.
It's only audiophile if it's inconvenient.

Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #14
----Mastering engineers, producers and mixing engineers need to tell their clients, starting right now, to turn on Sound Check, and to audition their references through Sound Check so as to hear what they will sound like "on the Radio" against the "competition". This collaborative act alone among producers will hasten the end of the loudness war. We also should encourage them to play their i-devices in the car with Sound Check turned on. And leave their CDs at home.

While turning on Sound Check will demonstrate to clients that peak limiting and extreme compression won't make their product louder, I fear there is another factor that means the war is far from won...

Listening habits have changed radically. A lot of people now listen on the move in noisy environments. And that requires quite a lot of DRC in order for the quiet bits to be audible. So while the hunt for extreme loudness might go away, the perceived need for extreme compression won't - especially if the clients check their masters in the car! Until mobile playback devices routinely incorporate compressors, I fear we're going to be stuck with over-compressed masters. Which remains bad news for those of us who like to listen at home on decent systems.

Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #15
What Apple giveth Apple can also taketh away.

If Spotify felt the need to pander to their suppliers by making the quiet louder rather than the loud quieter and leave it as an option anyway why would Apple end up any different?

What's in it for them? Next years we are going to see a blizzard of attempts to sell us, or rather resell us, DSD. All the big labels and manufacturers will be in on it.

Would it make any commercial sense try and flog 'better'  lossy and 'better' lossless at the same time?

Mmmh. Though. Always follow the money, right? Where is it going to flow?

Could be interesting but by no means a done deal.

Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #16
----Mastering engineers, producers and mixing engineers need to tell their clients, starting right now, to turn on Sound Check, and to audition their references through Sound Check so as to hear what they will sound like "on the Radio" against the "competition". This collaborative act alone among producers will hasten the end of the loudness war. We also should encourage them to play their i-devices in the car with Sound Check turned on. And leave their CDs at home.

While turning on Sound Check will demonstrate to clients that peak limiting and extreme compression won't make their product louder, I fear there is another factor that means the war is far from won...

Listening habits have changed radically. A lot of people now listen on the move in noisy environments. And that requires quite a lot of DRC in order for the quiet bits to be audible. So while the hunt for extreme loudness might go away, the perceived need for extreme compression won't - especially if the clients check their masters in the car! Until mobile playback devices routinely incorporate compressors, I fear we're going to be stuck with over-compressed masters. Which remains bad news for those of us who like to listen at home on decent systems.


Your point is well taken, Clive, but I think that we are in such a state with the digital loudness war that compression practices are far beyond what is necessary for the car. I'll make a bet that the practices we performed to make a good-sounding LP will be more than adequate for the i-device world. That the LRA (loudness range) of a good-sounding LP will work fine for idevices, even in the car. I could be wrong, but that's my premise. Have made many masters that have been auditioned in the car, and on the few occasions when the clients complained that some part was too soft I quizzed them: "Were you driving at 50 miles per hour with the windows open?"  (or something like that). It usually turns out they are stretching what you can expect from within any car and no sensible client or engineer in the new normalized loudness world would take the master to that extreme and ruin it for playback everywhere else beside the car.

By that time, DRC algorithms will have become so sophisticated that the portable player will use its microphone, asssess the background noise, and raise the gain only during that time. So the war will still be won. That's what I'm hoping and predicting.

Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #17
...It is true that I have been prophesizing the end of the war, after 33 years of digital war there is finally light at the end of the tunnel. That's what we should rejoice over! Perhaps my announcement was premature because iTunes Radio has only made it in the U.S. so far. Perhaps I should revise my "mid-2014" estimate to later. But my goal was not to be held up as a prophet, but rather to be an evangelist…

...Someone praised Spotify...


I was merely teasing, Mr. Katz.  I think you're doing a good job of evangelization and value your efforts to see the end of the Loudness War.

(I don't personally use Spotify other than to preview a few potential album purchases as a free-user. I prefer iTunes but know Apple can't really take a lot of Spotify's premium user-base. iTunes Radio really is like radio whereas Spotify Premium users that I know are caching entire albums in offline-mode instead of synching their own digital files to their phones and iDevices. Apple will never offer this sort of feature as they are the biggest retailer of music and it would obviously cannibalize their album sales.)
The Loudness War is over. Now it's a hopeless occupation.

Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #18
Thanks for only teasing, Engelsstaub! I got sliced and diced to some extent over at Gearslutz and dozens of other forums by people who have no understanding of what loudness normalization is about, so it's nice to know I have friends at HydrogenAudio!

Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #19
Kees has the intent of my evangelistic effort down pat! Many people have claimed that I said "the loudness wars have been won". But that is not what I said, and Kees has it correct. Of course we have a long way to go. But the last big battle has just been fought and there is a winner: All of us won.


I have little meaningful to add to the discussion, so here's a .GIF expressing my exultation.


Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #20
SoundCheck has been with us for many years.  The one significant change we have had recently is, iTunes Radio has SoundCheck on by default, and iTunes and iPods default to SoundCheck off.  People won't typically change defaults, so SoundCheck should now be considered "standard" for iTunes Radio.  This would be significant if/when iTunes Radio achieves significant market penetration.  But iTunes Radio is just over a month old, and USA only.  By comparison, Pandora has been up since 2005, has 200M global subscribers, 70M of which are active regularly, and offers commercial-free subscriptions.  Pandora isn't using SoundCheck, but is clearly handling loudness pre-processing in some way, as their levels are quite consistent track to track, but without the obvious additional dynamics processing that radio stations apply.

To my knowledge, SoundCheck is meta data that sets a static gain adjustment per file, similar to ReplaGain, but using a different loudness determining algorithm.  There is no dynamic processing in SoundCheck...unless there's a new development.

I find this sentence in the cited article, therefore, puzzling: "iTunes radio will not just ‘turn down the volume,’ but may peak-limit the important transient peaks of the material and make the song sound ‘smaller’ and less clear than its competition.”  That seems to imply dynamic processing, which would have to happen pre-streaming, and outside of a single SoundCheck meta-tag. 

It would be interesting to know where that statement came from, and on what basis. 

You could cite Pandora to clients as a determining battle to the loudness war as well, a battle won since 2005, and it's had far wider impact for far longer than iTunes Radio (so far), but with similar results.

Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #21
SoundCheck has been with us for many years.  [...]  Pandora isn't using SoundCheck, but is clearly handling loudness pre-processing in some way, as their levels are quite consistent track to track, but without the obvious additional dynamics processing that radio stations apply.

You could cite Pandora to clients as a determining battle to the loudness war as well, a battle won since 2005, and it's had far wider impact for far longer than iTunes Radio (so far), but with similar results.
Since the loudness war is still ongoing, and if your information is correct, Bob Katz' assessment is wrong, since an entity with a large impact did not win the loudness war for us before, why now? I'm not saying it cannot happen, but at least it has not happened before when it could have.

On the other hand, nobody wanted to buy Microsoft tablets some decades ago, but everyone is flocking to Apple now, so it might still happen. They have a knack to thrust into markets and initiate developments.
It's only audiophile if it's inconvenient.

 

Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #22
Hello, DC to light! It's important that other people also measure the output of iTunes Radio to see if anything that I discovered is incorrect.

I reached the conclusion that sound check might be peak limiting because for SOME songs I measured a different PLR (peak to loudness ratio) between the output of Sound Check in iTunes and the output of iTunes Radio. Then, after the press release went out I had a discussion with Thomas Lund who said possibly the explanation is that the bitrate of songs downloaded from the iTunes store may be higher than that of the broadcast, so the peak to loudness ratio could be reduced. It was a natural mistake on my part. Pity that Apple doesn't publish their Sound Check algorithm and so we cannot check the "black box" except empirically.

As to whether this is an important development or not my opinion is that it is! Because I believe this marks the turning point in the end of the loudness war. Apple has the power to reach major audiences and we can hasten the end of the loudness war by telling everyone to turn Sound Check on. Consider this the beginning of the mass consumer education campaign! If I have stepped on any parallel efforts that you good guys at Hydrogen have already started which I was not aware of, I apologize in advance.

Best wishes,


Bob



SoundCheck has been with us for many years.  The one significant change we have had recently is, iTunes Radio has SoundCheck on by default, and iTunes and iPods default to SoundCheck off.  People won't typically change defaults, so SoundCheck should now be considered "standard" for iTunes Radio.  This would be significant if/when iTunes Radio achieves significant market penetration.  But iTunes Radio is just over a month old, and USA only.  By comparison, Pandora has been up since 2005, has 200M global subscribers, 70M of which are active regularly, and offers commercial-free subscriptions.  Pandora isn't using SoundCheck, but is clearly handling loudness pre-processing in some way, as their levels are quite consistent track to track, but without the obvious additional dynamics processing that radio stations apply.

To my knowledge, SoundCheck is meta data that sets a static gain adjustment per file, similar to ReplaGain, but using a different loudness determining algorithm.  There is no dynamic processing in SoundCheck...unless there's a new development.

I find this sentence in the cited article, therefore, puzzling: "iTunes radio will not just ‘turn down the volume,’ but may peak-limit the important transient peaks of the material and make the song sound ‘smaller’ and less clear than its competition.”  That seems to imply dynamic processing, which would have to happen pre-streaming, and outside of a single SoundCheck meta-tag. 

It would be interesting to know where that statement came from, and on what basis. 

You could cite Pandora to clients as a determining battle to the loudness war as well, a battle won since 2005, and it's had far wider impact for far longer than iTunes Radio (so far), but with similar results.


Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #23
Hello, DC to light! It's important that other people also measure the output of iTunes Radio to see if anything that I discovered is incorrect.

I reached the conclusion that sound check might be peak limiting because for SOME songs I measured a different PLR (peak to loudness ratio) between the output of Sound Check in iTunes and the output of iTunes Radio. Then, after the press release went out I had a discussion with Thomas Lund who said possibly the explanation is that the bitrate of songs downloaded from the iTunes store may be higher than that of the broadcast, so the peak to loudness ratio could be reduced. It was a natural mistake on my part. Pity that Apple doesn't publish their Sound Check algorithm and so we cannot check the "black box" except empirically.

As to whether this is an important development or not my opinion is that it is! Because I believe this marks the turning point in the end of the loudness war. Apple has the power to reach major audiences and we can hasten the end of the loudness war by telling everyone to turn Sound Check on. Consider this the beginning of the mass consumer education campaign! If I have stepped on any parallel efforts that you good guys at Hydrogen have already started which I was not aware of, I apologize in advance.

Best wishes,


Bob




I'll sample some iTunes Radio stuff and see if I see what you did. I doubt if the kind of bit rate reduction they have to do in iTunes Radio would result in any significant change in DR though, that would be pretty nasty if it did.  But, as a "radio" service, it would make some sense for them to pre-process at least a bit, since even SoundCheck goofs fairly often.  I'm capturing some iTunes Radio now, will have to wait until I hear something I have an original of...

The reason I still think SoundCheck is a fixed gain tweak with a single block of meta-data is that some iPods have little or no DSP for dynamics processing (like the Classic), and couldn't handle the task.  A computer or IOS device probably could do it just fine, though.  In fact, if a track has a significant downward gain adjustment applied by SoundCheck, a Classic may play a few milliseconds of the file without the adjustment before it pops down to what it should be.  Kind of slow and stupid, but that's life.

The other indicator is, in an iTunes library you can reset the meta-data outside of SoundCheck by turning SoundCheck off and running a ReplayGain scanner on the files, setting a ReplayGain data block, then after synch with a player, turn SoundCheck on in the player and it responds to the new ReplayGain data instead.  This doesn't work in iTunes software because just turning on SoundCheck initiates a scan and resets any ReplayGain meta-tags.

One iTunes Radio comment so far... watching data transfer, as you might expect it throws songs at you in bursts of data, not continuously.  So there's no live pre-processing, it would have to be done on a per-file basis.  And, there's a LOT of "dead air" between tracks.  Odd, when you think of how it works.  They could easily provide a tighter seque if they wanted to...  And those commercials...yuck. That will drive me back to Pandora, even if the audio isn't as good. 

Mastering engineer Bob Katz: The Loudness War Has Been Won

Reply #24
There's a function on the mac terminal called afinfo which in OSX up to Mountain Lion gives the Sound Check normalization gain in decibels. OSX Mavericks has some new normalization functions (I've learned) but I don't know what they are. One thing is that afinfo in Mavericks no longer reveals the Sound Check normalization gain.

Clearly this is in the metadata for the AAC header. For Wavs, however, Apple keeps its own database. All of what you/we have learned and figured out about the behavior of Sound Check is based on behavior and inference and in most cases you are right, but it doesn't seem that Apple wants to confirm or deny any of this.

Do you have an FAQ on what you know about Apple's database and where it is stored and how it can be manipulated? It's one thing to trick iTunes into responding to ReplayGain data, it's another to feel comfortable that iTunes is continuing to do it reliably. Forgive me if I'm talking about things which you guys have long known or figured out, I don't spend much time at Hydrogen Audio, gotta make a living  :-).