Mastering in DP (making things LOUD!)

[magicd]

I've seen a few threads on this general subject. Here is one more!

As some of you know, I work with Al Kooper. I've had the pleasure of helping him make his last two solo CDs. I maintain his home studio.

Last year, Al celebrated 50 years in the music business. Part of that celebration was a re-release from Sony of 50 of Als solo songs from his Sony catalog. The masters were released earlier this year. Here's a link:
http://mp3.rhapsody.com/al-kooper/50-50

I was the mastering engineer for the project. Al looked over my shoulder and let me know what he was looking for.

The project was a real challenge because we were dealing with recordings that spanned seven years, two continents, multiple studios and multiple musicians.
Sony owns the original masters and provided us with 44.1khz/24 bit transfers from the original two track tapes.

I have a great respect for Als past work and I knew his fans would be upset if we radically changed the sound of the recordings. I've heard remasters that I thought butchered the original recordings and I did not want to be guilty of that.

We loaded all the tracks into DP and went to work.
The songs were all loaded into the same sequence chunk, onto separate stereo tracks. There was a single master fader for the session. The last plug-in on the Master Fader was the MOTU MW Limiter. The MW Limiter was responsible for final output level, as well as rounding from 24 bit to 16 bit, and dithering.

On the individual song tracks, we used a variety of other plug-ins. I used the MOTU MW EQ and MW Compressor here and there. I also used a few select third party plug-ins. In particular, I used the RNDigital phase linear EQ, which is quite amazing. In addition to the phase linear EQ function, it's also has a spectrograph and EQ "learn" function. This was a very handy tool for comparing frequency curves between tracks.
I also used the Waves Ren De-Esser on a few tracks that had sibilance problems. Because the bass varied so much from track to track, I used the Waves Multi-Maximizer. The Multi-Maximizer can be used as a final mastering limiter, but I did not use it for bit quantization or dither. I just used it to tame specific bass and sometimes 2-6k trouble frequencies.

On the first pass of masters, I was very hesitant to do any serious limiting. My goal was to even out frequency response between the tracks, and make sure level was up to the top of the sample, without any obvious compression.

On the studio speakers, Al was very happy with the initial sound. However, when he put the tracks onto his iPod and compared them to other current commercial releases, he was dismayed to hear that the tracks were significantly quieter than just about everything else he was listening to.

This is a common problem. You make a mix. You like the way it sounds. It isn't as loud as anything you can buy on CD.

So we did another three complete passes, each time bringing the levels up to get a stronger average signal. In the end, we had tracks going through as many as three different dynamics processors, with various EQs before and after the compressors, depending on what was needed. Keep in mind that we were working with stereo mixes, so it's likely that there were several additional layers of dynamics control in the initial mixes. My personal rule of thumb is no more than 4db change on any one EQ or compressor. By the time we were done, we had raised average signal strength on the tracks by as much as 12db. So far the reviews of the remaster have been positive. Whew!

The irony here was that in the past, Al has told me not to squash the mixes too much. He likes his dynamics. In the case of the remasters, he was telling me to "make the mixes louder, but don't lose the dynamics". Oh yeah.

Now is a good time to point out that mastering is a singular art and craft. There are people who post on this forum who are far more skilled than myself. I think the only way that I've been able to get into any kind of quality mastering work has been to put aside tracking, mixing, songwriting, etc, and just approach mastering as a unique skill. I've been mastering recordings since the 1980s and I still feel like I'm on the first page of the "Mastering For Dummies" book.

So, if you can boot up a plug-in or two and get a good master from the git-go, you are a Master!

Eventually we got the tracks to a point that Al was happy. When is a project complete? When the artist (or record company) says so. If it was up to me, I'd probably still be tweaking EQs and listening to CDs on every different speaker I could find. But we got the job done. As a side note, it took me nine months from completion of the project to actually get paid by Sony/BMG. I had to haggle with accountants in India! I realize this is the way the music business works.

I know that there are lots of great tools available, and as I mentioned, I did use some non-MOTU plug-ins on the job. But I'd also like to point out that the MW Limiter was the final mastering point, and it was good enough for Sony and Al.

Look at it this way, learning to master is easier than learning to play the violin. In order to do either, you do need to approach the skill with respect, and learn how to use the available tools.

And when you get good at it, you can squash the crap out of the mix dynamics and then go make some Chilli Peppers records!

Dave

[Frodo]

Dave--

Thank you for this. Your posts are greatly appreciated.

[cbergm7210]

Great post, Magic Man.

Chris

[FMiguelez]

And when you get good at it, you can squash the crap out of the mix dynamics and then go make some Chilli Peppers records!

Dave

Yes. You should probably add that care must be taken in order to NOT miss ANY spots... things must look like a sausage. If they don't, just lower the threshold until you are sure no spots were missed

All kidding aside, nice post, Señor Mágico!

[David Polich]

Dave, I met Al at last winter NAMM, over at the Marriott Cafeteria during lunch. I recognized
him and introduced myself to him and we talked for awhile and then walked over to the Yamaha
booth. On the way he told me that you were a 'great friend of his" and you had been over to
his place countless times to help him out.

This story brings back the memory of that afternoon. Yep, Al is one of the legends. An incredible
list of musical achievements. Great producer and a great keyboard player.

Thanks for the post. It also reminded me of Brian Eno's statement, "No one ever finishes a record - they
just give up on it". Or something to that effect.

[Eleventh Hour Sound]

Thanks for the insight Dave. I appreciate the candid view of your approach! I wonder how the final "loud" masters would sound on a stereo compared to the first masters just turned up louder.

I just thought of a solution to the "loudness wars." Wouldn't it be cool if the master volume on Stereos, iPod's, etc had a CC# assigned to them that could be encoded in the master recording? Then everyone could master for the best sound and dynamic range and let the listening device "turn it up" <Grin>

*Patent Pending (This would be an exclusive DP feature of course.)

[waxman]

Thank you for your humble post md. I am now choking on a big reality sandwich. I spent last week mastering with many failures.

What little hair I have left stood on end when you said,

I think the only way that I've been able to get into any kind of quality mastering work has been to put aside tracking, mixing, songwriting, etc, and just approach mastering as a unique skill

I finally resorted to putting the mastering tools in the DP inserts to get an idea what is going on with the multitrack. When I pull them out I realize I would never mix the multitrack like that. However the master it is coming out better.

I am the one on Page 1 of mastering for dummies. I think most likely you are on page 1 of the last chapter...

I believed I could do it but now after last week and reading your post I am having doubts but I am going to keep trying for a while... I will probably feel better about writing a check to Capitol or Future Disc if I can't get it. So bravo...

[Dave Connor]

Well that is all interesting in too many ways to count. Very much appreciated.

[magicd]

I wonder how the final "loud" masters would sound on a stereo compared to the first masters just turned up louder.

Yes, I think you would hear a difference between the original LP and the new digital master. The digital master would sound "louder", and have more bottom end.

I've done digital mixes that sounded great on CD, and were then transferred to vinyl. I could hear the digital ceiling on the vinyl (bummer!). If I had been mastering for vinyl in the first place I would have gone for much more dynamics.

I think the difference is that vinyl and tape don't have a fixed ceiling. Digital (at least PCM digital) does. With digital, your master is competing with every other recording that is hitting the same ceiling.

Digital also allows you to have low frequencies that you can't reproduce on vinyl, unless you do master at a much lower volume. The old Kooper recordings had been made with vinyl in mind. They were bright, but with fairly tame bottom end. Most of the dynamic difference in the tracks were mid and upper-mid transients. If we had just normalized the original tracks, those transients would have kept the average gain of the track around -25db. I haven't done an actual test, but I'd be willing to bet that if you digitized a random pop LP from the 60s or 70s, you'd find that transient peaks were about that much above average gain.

Pop CDs now have an average gain of around -12db.

I just thought of a solution to the "loudness wars." Wouldn't it be cool if the master volume on Stereos, iPod's, etc had a CC# assigned to them that could be encoded in the master recording? Then everyone could master for the best sound and dynamic range and let the listening device "turn it up"

There's a lot more to it than one slider!

And man, I hate iPods. I hated cassettes, and when CDs came out, I thought we had left lo-fi behind.
Who cares about fidelity in the mix when you're listening to mp3 encoded audio through earbuds? Bleah.

Dave

[waxman]

And man, I hate iPods. I hated cassettes, and when CDs came out, I thought we had left lo-fi behind.
Who cares about fidelity in the mix when you're listening to mp3 encoded audio through earbuds? Bleah.

That is the eight hundred pound gorilla in the room. I can't stand to even think of the stuff going to ipod, compressed crap downloaded from itunes and finally put through those wonderful hifi earbuds.

Why bother with DP and all my stuff listed below when I can just record and master with this awesome piece of gear and sound like this group... it rivals Abbey Road and the Beatles.

http://www.prosoundnews.com/Blog.aspx?id=22892

[Eleventh Hour Sound]

Hi Dave, what I meant was, I wonder how the new "soft" digital masters (the first pass you did, not the vinyl) would sound compared to the "loud" digital master (the last pass) if both were played on the same system using the system's volume control to get the same overall level.

[richardein]

I think, in any discussion of mastering, there should be a mention of Bob Katz's masterpiece, Mastering Audio, http://www.digido.com/mastering-audio-book.html

The book is as clearly written as any book on music or audio can be. Yet it will take you quite a while to understand completely the material he covers.

Bob's book is not the end-all and be-all on the subject of course. You need to master audio recordings for a very long time before you know when and how to use the amazing tools that are available.

Be sure to read Bob on the loudness wars. Very enlightening.

[BobK]

So we did another three complete passes, each time bringing the levels up to get a stronger average signal. In the end, we had tracks going through as many as three different dynamics processors, with various EQs before and after the compressors, depending on what was needed. Keep in mind that we were working with stereo mixes, so it's likely that there were several additional layers of dynamics control in the initial mixes. My personal rule of thumb is no more than 4db change on any one EQ or compressor. By the time we were done, we had raised average signal strength on the tracks by as much as 12db. So far the reviews of the remaster have been positive.

So, was most of the level increase was done on the individual tracks rather than via the limiter on the Master fader?

As for the mp3 thing - 'why fuss with making things high quality when it's just gonna be run through compression algorithms and listened to over cheap earbuds?' I've heard the argument that the better quality material you start out with, the better the end product will be, even if it's just a 128 kbps MP3. Makes sense to me.

Thanks for the great post, Dave. That one's going in my archives.

[Shooshie]

Thanks, Magic Dave, for the great story. I appreciate that.


Shoosh

[magicd]

So, was most of the level increase was done on the individual tracks rather than via the limiter on the Master fader?

My rule of thumb is no more than 4db of change with any one processor. The MW Limiter on the Master Fader was primarily for rounding and dither, and final gain check. By the time the signal got to that point in the chain, the MW Limiter was there primarily to catch any stray peaks and to insure a consistent ceiling. Lookahead was set to max and release was set very fast.

Compressors and limiters earlier in the chain were more aggressive. The big challenges including controlling overall bottom end, as well as sibilance. First step was to control any sibilance problems. Next I looked to control transients, which were mostly mid range. Last step was to bring the overall mix up to the volume I wanted.

A lot of mastering (for me anyway) involves watching spectrograms and level meters. I know what RMS (average gain) I need for the mix. One approach is to work backwards. If you set the final limiter so that you have the average gain you want, you will probably hear all kinds or artifacts that you don't want. That's what I used the RNDiginal EQ/spectrum analyzer for. It could show me what was peaking in the mix. I could then apply EQ and compression upstream of the MW Limiter to grab the problems.

I think it's a relatively common thing to use multiple dynamics processors in the overall signal chain. Each step gives the signal a minor "haircut" until you have the level you want.

As for the mp3 thing - 'why fuss with making things high quality when it's just gonna be run through compression algorithms and listened to over cheap earbuds?' I've heard the argument that the better quality material you start out with, the better the end product will be, even if it's just a 128 kbps MP3. Makes sense to me.

And to me. I just don't like what the mp3 compression does to the mix. In the case of Als 50/50 compilation, it is a download-only release. We knew at the start that the final product was going to get mangled in the mp3 process. That had a lot to do with why we needed the tracks to compete with other tracks on an iPod.

Dave