Hi guys!

I've found this amazing forum looking for mastering related information and I've read some of your threats.

Cool!

In one I've read a Bob Katz answer where he told about the need of the right mix (and all the previous production chain...) to achieve the maximum loudness.

Which properties must a mix have to be raised up at very high level without affecting the sound quality?

I'm referring to many pop productions that sound very clean (and great in my opinion) at an incredible loudness (for example Britney Spears, Michael Jackson...). I had different experiences in great mastering facilities with top level engineer who made a really great work with my mixes, but I didn't achieve the loudness I wanted.
In these cases I avoided to ask to the engineer to rise up the level as he told me that, in his opinion, an excessive level will get worse the sound.

What's your opinion about it?

Thanks you all!

Gianni

Very wise ME.

>and I've read some of your threats.

Hilarious! heh

I bet Bob Katz knows this really well but I will chime in with a few points. In order to get a REALLY LOUD and CLEAN mix you need to first of all capture the signal very efficiently. This typically means miking quite closely and use mics that don't generate a lot of distortion when doing so. In order to get a great focus you also need to combine this with a great pre amp. So this is the first step. The second step has to do with the transients. If there are a lot of peaky transients coming in to the DAW but the average volume is low these transients will become problematic when you start gaining. For this reason you should process the signal such that it comes in rather round. You can use a couple of good analog compressors, a tape recorder or whatever that smoothes out the peaks. Some op amps can saturate close to the clipping point and some op amps can even saturate after the clipping point. But there are also op amps that become extremely harsh sounding close to the clipping point, which in turn means you need to track quiter and you lose signal. Once the signal has been converted without clipping distortion you are ready to mix for loudness. What this means is to mix such that you lose as little of the signal as possible. Use the maximum amount of bit depth to introduce as little quantization distortion as possible. Cut on the EQ instead of lowering the volume faders. Use few effects. Make sure there is minimum amount of phase cancellation and frequency masking. Once you get this far you come to the last and most critical step of gaining. What you do is to push as much of the signal as possible through a really tight peak limiter ceiling. You can do this in many ways. Some prefer clipping converters and have a limiter applied on the mix bus, some prefer a good sounding peak limiter. I think the most efficient way is to push the whole mix through the same pre conditioning route that you were using during the tracking of the individual tracks and then use M/S mastering for the rest and have a good peak limiter applied on the mix bus. This peak limiter should be of the highest possible quality, since it can destroy the whole mix. Once you have full access to the mid and sides you can use a multiband compressor for additional cutting. You might want to cut some lows on the sides for instance and boost some frequencies on the center. At this point the mix is LOUD. But if you need to make it even louder there's one last step: Squeezing until AUDIBLE clipping distortion! For this you should really have a good monitoring setup to be able to notice the distortion artifacts and you should really test this on a lot of sound systems and volumes so that most systems can handle it without any kinds of distortion artifacts. (don't cut the audio in the wrong way when you do audio editing) What you do is to trim the signal to overloads. You then mark the points where the signal clips audibly and use accurate fast automation to cut out these points. You then trim even more and apply more automation. You can do this with the DAW editor in several cycles until you can't do it anymore. (you can preferrably do most of this in the mixing phase instead due to the higher isolation level) Then you un-trim the signal until there's no audible digital distortion left and until the mix sits like it should.

Warning! At this point the mix is LOUD, EXTREMELY LOUD! When you open the audio file it will be totally black! This is not audio that I prefer myself, I rather want to turn up the volume on a good sounding mix than down the volume on a bad sounding mix.

There is no substitute for a great ME. But, you have to have a level field of freq's. Try looking at a spek analyzer and you will see that a great ME will make everything sound "louder" but really it could be at a lower level. I think it is more about RMS than PEAK, they are good and thats why they make money!

Sounds like you got good advice from a professional ME.

If you're looking to provide mixes that have the potential to be even louder without immediately sounding a lot worse you'll need to make sure your mixes are under control
with respect to dynamics and that there is no out of proportion buildup in any frequency range. It's as difucult a skill to "master" as it is to describe.

The simple answer is no big dynamic swings.

Hi Rainbow,
And thank you for the quick reply.

Except that I've obviously done all what you wrote during recording and mix, you final accurate description is, in my opinion, something that is more mastering than mix related.

In fact you described an “own made” mastering process….that is surely something very good to do if you haven’t planned to go to a professional mastering studio, otherwise I think it should be avoided because if you bring to a mastering engineer a pre-mastered mix he (…or she, to avoid discrimination…) will haven’t “space” for his work.

The problem is: how a mix has to “sound” and which spectral and dynamic (…if it’s possible to encode…) behavior has to follow (roughly speaking, otherwise all mixes will sound the same) in order to have the possibility of making a “super loud and super punchy” master (as productions I’ve quoted)?

The question, in effect, is especially directed to mastering engineers…when you receive a track to master, and the producer asks you for the loudest sound, what do you verify to be sure that you’ll be able to do it without affecting sound quality?

Come forward mastering guys!

What do you mean in terms of db as "no big dynamic swings"?

There's only one way to tell: Experienced ears in a great room with great, uncompressed monitors. Then you listen as objectively as possible, comparing the source with the master at equal loudness. When the sound starts to go downhill, you try to stop. But this is often 1, 2, 3, or more dB lower than the client would like :-(.

BK

Depends on a lot of things.

How much louder you want to make it [past a certain point ANY dynamics will cause wild distortion]

The density of the arrangement [more sparse is genereally easier to make louder].

andd lots of other little variables.

In gereral, the louder you want your master the less dynamic your mix should be.

I should say for the record that this is not a recipie for the best possible sounding mix. Just the loudest.

Thanks Bob,
It’s an honor to have your answer!

Yes, I understand what you mean….but my specify problem (and the reason I’ve posted the threat…) is the following (I try to describe my experience to see if you and the other engineers can give me an opinion):

1) I’m a producer/sound engineer, and usually I mix my production by myself using a DAW with some other analog processors
2) I’ve finished a production in 2 languages (so the same album with the same mixes first in Italian and after some months in Spanish) and I’ve mastered it in two different high level studios
3) Both ME, when I’ve asked to have more loudness, told me that it wasn’t a good idea as they already pushed up the mix at the maximum possible
4) As I compared these masters with other similar as music style, arrangements and mix approach (…I think…) the mine was 4-5 db quite
5) So my concern is: as I’ve heard masters that sound louder than mine (and the sound quality is not affected) what could be the problem in my mixes that “blocked” ME work? What can I do in my future works in order to reach this loudness during mastering?

Hmm... If you do it like me but don't succeed it's either due to your ears or due to your monitors or or due to your room or due to your gear or due to your mixing skills or due to a combination or due to whatever, since I don't have the same problem when doing the same thing. Do you have the stereo pan law set to attenuate with -6dB or -4.5dB? Mine is set on -3dB... Are you not using an audio engine at 32-bit or more? Are you using some noisy plug-ins? The number of variables is endless...

Who knows what those mastering engineers thought... They probably noticed distortion in the signal itself (guitar pickup noise etc). They might also have noticed that the sound field was not very efficiently consumed, since they "couldn't" make it louder. That's typically for dimension reasons or for noisy sound sources. I think you should work more on signal distribution in the sound field. If you feel that is under control as well, change mastering engineer. heh

I hope that helps...

Umm. They might have noticed that it was starting to sound bad and advised against making it any louder. It's much more likely that the mix itself was starting to distort rather than any individual track noise [like guitar pickup noise].
Whomever it was, they did the right thing.

Yep, but typically track noise results in mix noise after the summing process.

Please post a clip and we could maybe here what the problem is... To me it seems like a problem with the way you have consumed the sound field. Either you have noise in the signal (due to whatever, everything from pickup noise to plug-in noise) or you have distributed the signal too inefficiently. As Bob Katz said, it's also a matter of what monitors and room the mastering engineers are using.

You asked what a mastering engineer does when he gets a mix in his hands and is told to make it loud and needs to verify if it is going to work. I don't think there's any standard answer to this, but he probably know a lot just by knowing the genre, the mixing engingeer etc. Maybe he starts asking questions about how the mix was done. Where you using this or that when you did this or that, just to find out what he is dealing with. How did you blablabla... He might also use a number of analyzing tools and techniques to find out things about the mix himself. I use reference speaker systems for this. It might include the application of different noise filters/gates in combination with other effects on different speaker systems to see what's left of the signal and pay attention to it. This can be very informative on an M/S setup for instance to find out how good the instrument separation is, which is a critical thing related to gaining. But professional engineers that have worked long in the field I think can use their ears. They don't need very deep minded approaches because they know the sound and their monitors so well and have done it thousands of times before and it has worked then. I think many just do what works and don't care too much about it. In this case they might have noticed it simply sounds bad to limit more (either they knew why or they didn't or they didn't have time to try make it louder), for whatever reason I think they thought it doesn't work, let's leave it at that... The most simple way of finding out why is to ask the mastering engineers that thought so.

The first thing I would do would be to analyze the high end isolated in different speaker systems. Just by limiting a few dB I would pretty much know if it's a good thing or not to limit more. I would then take another small step towards the decision, I would analyze the sound field consumption to see if there are any possible dimension or related problems related to limiting more. I might notice that by limiting more I get a chain of new problems, like for instance some important instrument in the center is becoming blurry or the interaction between the kick drum and the bass guitar doesn't work... Then I would try finding alternative approaches to that. If I wouldn't find good enough alternatives I would probably end up doing the same decision as your mastering engineer and realise it needs to be fixed in the mixing process because the material and tools I have can't target the problems isolated and transparently enough.

You know, professionals are quite familiar to different anomalies and know what to pay attention to. They learn by mistake and don't repeat their mistakes.

O.K. maybe we should talk about that instead.

Yep, it might really be the problem source in this case...

What plug-ins are you using when you mix? If you have placed any plug-ins on the mix output bus, what plugs? How many elements are you typically using on what kind of sample rate/bit depth? How have you set up the guitar pickups and what kind of pickups are you using? What converter are you using and how hot do you track with it? Do you record samples or hardware synths? What type of mics are you using? What kind of connection types are you using? What amps do you use? Do you use outboard DSP effects? What miking technique(s) are you using? Do you record in a damped or not very damped recording environment? What kind of monitors are you using when you mix? Are you typically applying a lot of effects or very few? Do you track and mix mostly ITB or OTB?

A lot of questions...

So to sum it all up:

- Clean and good sounding sound sources
- Great signal (successful tracking)
- Great signal distribution (track volume balance both on sub group and main group level, mono compatibility, imaging, realism, depth)
- Low amounts of digital processing, on highest possible bit depth, avoid mix bus effect processing
- Few negative compensations due to a big difference between the mixing engineer's and the mastering engineer's monitoring setup (also converter differences)
- Proper and clean EQing
- Clean plug-ins used properly
- Keeping the number of tracks and elements played simoultaneously low
- Learning to recognise sound anomalies, how to find them most efficiently and how to best avoid and target them
- Good transient control
- Effective use of automation
- Transparent mixing procedures
- Great gear
- Great ears

These can be summed up to any the following three problems that you have:
- Bad input signal
- Bad signal distribution
- Bad monitoring environment

So use the sound field more efficiently and I am sure you'll get the loudness/clearity ratio you are after. Of all these properties I think great signal distribution is the key. When you gain a lot you will soon lose the proportions of the original image. The worse it is before you start gaining the sooner you'll need to back off on the gaining...!

4-5 dB indicates a couple of possibilities, but from my experience, PROBABLY the ME. was right. ANYONE can make it louder. But can they make it better. And if the sound goes downhill then that's what the M.E. meant by saying that it was at the "maximum possible". He knows that he could go further but the sound would get worse.

The ONLY way you could tell if you could make it louder without losing quality is to actually be there in his room, and listen while the M.E. sweats bullets trying to beat the laws of physics. I'm betting it's not going to be that 4-5 dB that you heard. Objectively as you listen the 4-5 dB louder one is going to sound worse, I'm goign to bet!

Somewhere above a K-12 no matter what you do, the bass is going to start sounding Wimpy, the sound is going to get harsher and harsher and more edgy, and the depth is going to become as flat as a pancake. Laws of physics.

And we have to somehow find a better language. Instead of "louder" I prefer "louder in front of the volume control", or "higher RMS", because it's not really louder, remember, the listener has control of his volume control. And besides that, at some point as the RMS goes up, the loudness stops going up, because the transient response is a very important part of the loudness, and the transients are gradually lost as the RMS level goes up.

BK

Thanks a lot rainbow for your opinions.

In effects I've just finished a studio reequipment in order to fix some limitations of my past environment (the one used in productions I was talking about), especially in the pre-DAW chain (pre-limit-eq) and monitor setup.
It's surely right that the experience in this field is very important, as probably I have noted some errors in the mixes (due to uncorrect approach to mastering process) like:

- Keep BD and SD a bit more higher than the expected as in the mastering they will be lowered by the limiting process
- Too much dynamic in the whole song mix (for example in many cases I had a strong level difference from the beginning of the song to the end, the purpose was to have an interesting effect of "growth", but this obviously blocks the ME using limiter)
- Maybe a too rich arrangement....this is something I still have to investingate better

I'll check also your great list....

Ok, but I haven't understand this a part of your answer (sorry I'm not native speaker...)

I've read in many of your posts this "K scale", could you please explain me what does it mean?

This is surely right! In effects I've used the term "louder" meaning "louder in front of the volume control", in the sense that I've compared the "loudness sensation" listening to different masters without changing volume control....while my "4-5 dB quite" is measured as RMS, and this is not the only parameter that affects the loudness feeling.
My concern is, in effect: "how can I have a very loud master (so something that the listener can hear at a useful level even with volume control very low) without losing sound quality and punch?" as I have listened to masters that have these characteristics.....but I know that probably there's not a single answer to that!

Thanks!

I have to disagree with this one point. You should get the mix with the precise balance you want BEFORE mastering. Louder kicks and snares will only trigger the limiter earlier than you want causing you to push it harder to get the desired level but with much worse [probably] effect on the mix overall.

The musical arrangement plays a significant role as well. If you aren't blessed with a good one, in mixing, try to leave space. Just because they recorded 12 tracks of guitars and have them blaring full tilt from start to stop doesn't mean you need to mix them all up all of the time. Of course the artist or producer has to be in on this decision.

A well balanced spectrum helps too. Don't have evrything bunching up in one area, but don't overcompensate and think you need to look at an analyser and get a straight line. Use your ears, not your eyes. Basically what you're looking for is a good song and a good mix. No magic bullet available.

Dear Karibu, I hate to make you wade through pages of English, but the best way is to read the original article that was published in the AES Journal.

http://www.digido.com/portal/pmodule...er_page_id=36/

This is part one. The K-System is desribed in a link at the top of that page to part 2.

A K-12 is "roughly" material with a peak to average ratio of 12 dB.

You'll freak out at this answer: "Optimal Compression". I don't have another optimum answer in short. It's also tonality and presence and a great mix to begin with. If the mix engineer mixed listening at multiple levels then the master will be easier to make.

This new post re dance mixing may be useful!

https://www.gearslutz.com/board/so-much-gear-so-little-time/74223-dance-music-mixing-yes-kick.html

I think that if you want to sound loud, you have to prepare your mix to that.
When someone ask me to push their mix to highest level, I do it until it sounds ok.
You want more? You need to do a different mix. It's not only a mastering issue. It's a combination.

I agree with karibu regarding keep the transients high. (KD, SD).
I usually recommend (and I do it too) to put a hard L2 in the mix, just to preview what will happen to the mix when the RMS grows 9 or 10 dBs. It's just a preview ok?
You don't print the final mix with that.
Then, you end with a mix with much more air, and maybe exagerated dynamics, but in the end, after the mastering process, you will have similar dynamics like when you were mixing.
Of course, any good mastering setup, will outperform the L2 job.

And Bob, there's a third kind of fool, the ones that ask you to get their mixes louder and then ask to keep the dynamics untouched.

Mastering for loudness: what a waste of time. I spend more time than before, because of this.

That's right, maybe I haven't explained in the right way, but in my post I would mean this one as an error I made in the past, so I perfectly agree with you.

Sorry Hagen Daaz, but I've seen that maybe I wasn't clear about it in my post:

do you mean that you think it's a godd idea to leave BD and SD a bit higher in the mix or not?

Yes, you're surely right from technical/artistic point of wiew, but unfortunately (especially in pop and pop/rock productions, the one I work on) the loudness is always requested as a primary purpose. When A&R guys listen to a master the first comment is always "Great! It's very loud!" or "MMMMHHHH, it sounds low....", and this can determine a deal or not......

I know we can discuss about it for hours (I'm sure there are already many threats about this problem...) but this is the truth, in the "commercial side" of our work!

Thank you for your comments.

Leave BD and SD a bit higher in the mix.


Yes, I know that very well (I live that situation every day). That's why I'm saying it.

Higher then what?

I think it's clear. Puting an L2 at -8, obligues yo to expand the dynamics of your mix. When you listen to it without tyhe L2, it reads like the BD and the SD are really higher than your standard mix.
I'll try to upload a few samples of :
a mix A,
same mix with L2,
mastering of the original mix A.

if you want. The method works. It reminds me when I worked hearing the sync and repro head during a mix.
Remember, I'm dong all this **** just to sound al loud as Green Day, at client request.

I believe the original question was:
"Which properties must a mix have to be raised up at very high level without affecting the sound quality? "

The appraoch you are suggesting might be good if you are looking for interesting limiter or compressor effects [like pumping effects] but it will not help you achieve maximum loudness with the least loss of quality or most distortion. What you want for that is to CONTROL the dynamics as much as you can while retaining as much "punch" as you can. The louder you want it the less punch you will retain.
Purposefully making the kick and snare louder than they should be is the opposite of what you want to do here.