Mixing: Loudness missing

[jhonny.eriksson]

Hello,
One thing that I constantly struggle with is to get the right loudness in my productions. There are so many places you can adjust volume (Instrument volume, Master volume, MIR microphone volume, Expression: even Velocity) plus all the different effects that are there to help such as limiter, compressor et al. I realize that this is a huge topic but I really need some tips to make my life easier.

I've tried to follow the directions from the great Beat Kaufmann; following his step to Cubase mixing. My normal setup on the master is:
Master EQ
Multiband Compressor
Master EQ
Maximizer
Limiter
Stereo Enhancer
UV22HR

The Multiband Compressor I find it really difficult to work with as sometimes slight changes on any of the other volume impacting knobs make it sound dreadful. Usually I try to strengthen the low, lower the mid-low and higher the mid-high (and high when needed). The thresholds I usually tend to end up with less on the low and more on the highs.

I usually want to use Natural Volume in MIR to get the overhead I need to create my dynamics but mostly I just end up having to higher them when I master as no matter what I do with the plugins I just don't get as loud as I want.

I try to use Expression smartly to create the space I need for dynamics which I can detail with V.X-Fade. My normal working routine is that I place my piano on the MIR stage. Set it to Natural Volume; plus slight mixing EQ etc. and use that as a "starting point" volume-wise. After that I adjust Expression to create a "normal volume" for my instruments which I can later use to higher when playing forte or lower coming into a piano section. Does not seem to work very well tough...

Sample;

https://www.dropbox.com/s/w9dbofytbii3tw2/Mixing-Sample.mp3

Picture of master:

https://www.dropbox.com/s/93c1igqar4b20m8/Master-mix.jpg

[Beat Kaufmann]

Hi jhonny.eriksson

It's a pure chance that I visit this Forum. Sometimes I'm off over weeks...

Something about the row of mastering effects. With classical mixes it is not really neccessary to use a multiband compressor.

So I would use and try with the...

1. Compressor (Ratio: 1.5, 20ms Attack, 150 ms Release as a starting point) for "glueing" the sum a bit together

   If you used MIR you don't need this comprssor. MIR glues together all the instruments as well (sometimes it produces too much "glue" when the Wet ratio is > 40%)

2. Master EQ, OK, for searching bad frequencies...

(3. an EQ for Sound-Design?)

- No Maximizer

- No Stereo-Enhancer with MIR! 

4. ...without MIR: Why not a bit enhancing Stereo but controll the Correlation-Value. It should not go below 0 in average.

5. Limiter OK

6. Dither OK

A Mixing-Tip:

You should try that frequencies below ~ 100 Hz (or even a bit higher) are not treated by the used reverb system.

Unfortunately these frequencies are often blowed up by reverbs if you don't do anything.

Bad: you can't "repair" such a blowed up result with an EQ afterwards (with the mastering process)

I don't know how ever you want to reach this technically.

First of all you should give each instrument with the exception of the double bass a highpass filter. 

For the D-bass:

Either you can filter the low frequencies before they go through the reverb or

you can filter the reverb signal itself (the two convolutions of the SUITE EFFECTs are able to).

A further solution could be:

Copy the D-bass (so that you get two tracks of it) and insert a Highpass-Filter (100Hz) in track 1. This track goes now through your reverb

as usual.

Track 2 of the bass gets a low pass filter (it lets pass frequencies below 100 Hz). Route this signal directly to the output (not through the reverb).

Now you are able to adjust the low bass frequencies separately. Find a good balance then.

And also:

If you are using MIR: Try also to lower the WET-Ratio over all (40% wet/ 60%dry - 30% wet/ 70%dry)

The result is more "crispy" without losing the room impression

Best

Beat

[Dietz]

Some good suggestions by Beat! :-) Just to offer a sligthly different point of view:

There's nothing wrong with bus compression on a MIR Pro-mix (full band and/or multi-band), and there's nothing wrong with stereo-enhancers*) either - provided that you know why and how to use them.

That said, it's the best idea to start with those effects in a master chain you need, and - even more importantly - you know with uncanny sureness. :-) Leave away everything you've just been told on the 'net that "everybody usese this", and/or which doesn't show any desirable effect.

Personally I would start with nothing than a transparent, high-quality full-range limiter with its output ceiling set to about -0.3 dB. Then raise the input volume (most of the time by lowering the threshold level) until the limiter starts to kick in on the loudest hits, but not during longer sustained notes.

For sculpturing the sound, don't aim for any "coloring" plug-ins as long as you really know why you would want to hear them. Use a clean, high-quality parametric EQ instead (Viennna Suite's Mastering EQ is a great choice). A high-pass around 20 Hz is always a good idea, above that you will start to kill important fundamental frequencies. Its a better idea to use MIR's Character Presets for individual tracks, as they will always take care for LF-rumble (if any).

For "glue", a full-range compressor with settings similar to the ones suggested by Beat can make lots of sense. Personally I tend to "mix through" a compressor, which means that it's part of the master bus from the very beginning, so I hear its effect and consequently work with its impact on my mix. But this is no "fire and forget"-solution: I constantly re-visit the bus compressor for minor adaptions throughout the whole mixing process.

Re - Volume: I know that it can get confusing with so many possibilities of changing it. Assuming you're starting with MIR Pro's Natural Volume I would leave every other means of actual _volume_ control alone ("volume" in its technical meaning); the obvious main exception being the master bus volume. Controling your musical dynamics (opposed to the aforementioned "volume") should be done mostly by means of a MIDI-CC of your choice.

*) Regarding stereo-enhancers: Not enough people know about the "Shuffler"-mode of Waves' trusty old S-1Stereo Imager. Originally introduced by the late Michael Gerzon (one of the god-fathers of digital audio), this nifty little algorithm adds some interesting perceptional gimmicks to any signal that's derived from coincident microphones. MIR's impulse responses are captured by means of 4-capsule Ambisonics microphones, which are coincident by definition. The Gerzon Shuffler (used with care!) is able to make a MIR-based mix even more "enveloping" and appealing.

PS: Get rid of the UV-22 Dither. It's a very special "colour" of dither which is aggressivly noise-shaped. Only few mastering engineers would suggest its use for delicate orchestral music. Use something less abrasive as long as you don't see the absolute need for it. - In case you're using MIR Pro's RoomTone, you could actually do perfectly without any additional noise. ;-)

Hope that helped,

[Dietz]

Sidenote:

@jhonny.eriksson said:

My normal setup on the master is:

[...]

A brick-wall limiter should always be the final stage of processing a master (apart from dithering which should take place when changing bit-depth only). The Stereo-Enhancer _after_ the limiter makes no sense, actually, as it will change at least _something_ of the signal's peaks.

Kind regards,

[Beat Kaufmann]

Some further good suggestions by Dietz.

In addition from my side:

First of all: You always should do the enhancing process with monitors.

Why I personally advise against using stereo enhancers in connection with Convolution Reverbs?

There is a microphone method called ORTF (as an example). Two Mics, 110° angle, 17cm distance.

Recording with this empirically founded "stereo arrangemend" leads to a very roomy sound and you also get the depths of instruments nicely back while listening the result with speakers.

If you are going to use a stereo enhancer together with such a recorded track you completely destroy the nice stereo impression.

This happens of course with a real ORTF-recording. I don't know how VSL simulates the ORTF Mics within MIR maybe it is another process.

Nevertheless, using stereo enhancers in connection with good IRs (which normally are real room recordings) is not the same as with algorithmyc reverbs.

So even if you get a spectacular result in headphones you always should prove the result with monitors as well.

It would be a pity to lose the natural feeling of a room only for a getting a wider but unnatural and "phased" sound.

In other words:

Use the stereo enhancer gently and with caution in connection with convolution reverbs.

It is maybe better to use another IR which comes with the width you wish.

And of course as always: Less (effects) could be more.

Finally: As mentioned above, the correlation measurement is a great help for not doing too much.

Beat

[Dietz]

Hi Beat,

like mentioned before: MIR's impulse responses are all derived from true conicident microphones (opposed to "Near Coincident" like ORTF, wide "A/B" or Decca-Tree-like setups). The ORTF setup which can be found amongst MIR Pro's Output Format presets is of course a virtual one and still based on the same recording setup.

... here's a nice primer covering typical microphone placement techniques (with quite "conservative" aesthetical assumptions):

-> http://www.tufts.edu/programs/mma/mrap/StereoMicTechniques.pdf

Many Vienna Instruments have been recorded using ORTF setups (-> http://www.vsl.co.at/en/65/72/103/20.vsl), but as soon as we use any kind of panning device this is not an issue anyway. :-)

Kind regards,

[Beat Kaufmann]

Hi Dietz

I don't want to get in a war with you about stereo enhancers and Convolution Reverbs. And also, I don' have a problem to accept you as the Mixing-Guru here as well.

But believe it or not, there are others in the world who also are working with microphones almost daily...

So getting an ORTF-theory from you is a bit hard to take for me, sorry.

OK, for all others

My point of view (better no stereo enhancers after convolutein reverbs) is based on a lot of mixing experiences after real recordings with several

stereo microphone procedures. I will show it with a short recording example. I've recorded it in January this year in the KKL Luzern.

Listen to the Original-ORTF-Recording (with some additional microfones of course) >>> here

Here the same example, treated with a stereo enhancer >>> here (...a bit exaggerated, yes)

If you want to compare the two tracks with two parallel players you can do it >>>here (no 35.)

My observation:

At a first glance the second result seems to sound "better" (what ever better means) and louder (better?) as well.

But if you are listen a bit longer and compare the two examples (also with monitors of course) it is clear that the "enhancer effect" kills the precise information about the positions of all the instruments (L,R, and depth). If you are listening with speakers you also can make out a sort of phase-effect because it depends on how you turn your head...

And now my statement once more (which is hopefully proved now):

Good IRs are normally made with care and also with a certain microphone procedure as the upper receorded demo track was done.
So using stereo enhancers after a Convolutien Reverb (which works with natural IRs) can destroy the natural sound of those Impuls-Responses.

Nevertheless, everybody can enhance hell for leather just to his taste and as he likes it.

Beat

[Dietz]

Uh ... easy, Beat, easy. I don't understand why you react like this. [^o)] ... do you really think I call myself a "guru" of anything just by offering advices that (partly) differ from your opinion ...?

All I said is that MIR uses coincident microphone arrays for capturing the IRs. ORTF is a nice concept (and used VSL for many instrument recordings, like I mentioned before), but a different animal. 

The "theory" link I posted is for those readers of this forum who don't deal with real recordings that often (if ever). We have more than 50,000 subscribers here, not just you and me. ;-)

Best,

[Beat Kaufmann]

@Dietz said:

Hi Beat,

like mentioned before: MIR's impulse responses are all derived from true conicident microphones (opposed to "Near Coincident" like ORTF, wide "A/B" or Decca-Tree-like setups). The ORTF setup which can be found amongst MIR Pro's Output Format presets is of course a virtual one and still based on the same recording setup.

... here's a nice primer covering typical microphone placement techniques (with quite "conservative" aesthetical assumptions):

-> http://www.tufts.edu/programs/mma/mrap/StereoMicTechniques.pdf

...

Kind regards,

Hi Dietz

Maybe it's because English is not my mothertongue...

But with

"Hi Beat" followed by explainings about ORTF around MIR and finally with a treatment about ORTF... This wasn't meant for me? No instruction?

Not you - I called you Guru. Guru was the reaction to your topic which I obviously took the wrong way.

Normally I'm more a calmy boy - you know this as well.

Is it the full moon today... in any case: I'm sorry Dietz and thanks for clearing up the situation.  

Let's hope that this incident doesn't distract the forum readers from the real content of this thread.

All the best

Beat

[William]

I find this info from Dietz invaluable.  Somehow his approach is more understandable to me.  And he is definitely THE SOURCE on mixing ideas for me as his mixes are simply the best I have ever heard done with samples.  I feel people are extremely lucky to be able to hear directly from an expert like Dietz. 

I would add that there seems to be a lot of complication in that original setup.   To me complication is deadly in mixing.  I always need to start from total, raw simplicity and only add what is absolutely essential. 

Though I am trying to do something somewhat weird, not necessarily fast or efficient - I am trying to use VSL in its most raw state, because I feel that is what makes it great as opposed to the other libraries which are heavily processed.  And so I am extending that as much as possible to mixing which MIR - miraculously - allows one to do.  ESpecially with dynamic ranges and loudness, it is crucial NOT to use all the pop music limiters, compressors, processing that are normal, because they distort the purity of the orchestral sound which is actually captured in VSL.

[civilization 3]

All of this stuff on the master bus looks like you're ready to premaster the hell out of the thing like it's EDM music or something horribly smashed and entering the Loudness Wars full-on. There is no call for a maximizer and a limiter, and this following a MB Comp. There is little or no call for a maximizer here period.

Put things on inserts for an actual reason that you understand, not because you read it on the internet. You need to approach mixing before it gets to the master bus; you have a whole mastering setup and then some on the master bus. That's cheaping out in a big way.

Yes, the Mastering EQ is primarily meant for busses (or a two-file in an actual mastering process), but these big cuts in mids could really castrate some things in the mix. Are they there to fix big problems? It looks drastic for whole program EQ.

Learn to treat individual things first, make every instrument complement the other instruments by the FX plugins, work up to groups of sections maybe.

For instance I only ever put the limiter on the master bus in Cubase for overload protection and lately I'm liking that gone, because the transient excitement is at times mitigated to the point I lost some juice. If you don't know the techniques, resorting to limiting right off the bat is not wise. Using it to get loudness is not real suitable for straight orchestra music. I make other kinds of music primarily, some of it is {rock music-} loud but I never achieve it through limiting. Again, learn to deal with compression, EG case-by-case, approach mixing like a surgeon, you have some blunt instruments at the end of the path.

Yes, MB Comp is tricky or even treacherous, so study it some first. I think you didn't run before you could crawl as a musician so relax and cultivate some chops here.

Also, dithering today is pretty much about your end result is going to be heavily amplified before it's heard, and the REALLY quiet sections don't cover the noise floor; such as it will be heard in a movie theater. This is not a job for you right now, regardless.

[jhonny.eriksson]

Hello everyone,

... And thanks so much for all your great answers! I never thought that I would catch the attention of all great masters here on the forum, Beat, Dietz, William and Civilization that really hit the nail with his comment.

No. I do not really understand what I am doing to be frank. As I think it is for many my focus, all years writing songs, has been on composing. So when it comes to production, mixing and mastering I am a total novice; which is why I really needed your wisdom to push me in the right direction. I've tried to read the "Mixing Audio" reference but working full time as a networking engineer doesn't give you much time to read anything else (at least not pursuing the CCIEs certificates).

When it comes to actually trying your suggestions I instantly noticed a difference once I got the Compressor/EQ/Limiter setup. At least my meters are now moving (as if that proves anything). The problem with the Stereo Enhancer; so elegantly described by Beat; I instantly noticed. It did kill the stereo feeling. I was playing around with multiple master EQs because I felt that once I added the compressor (for example) it colored some not very pleasant frequencies which I tried to kill off. But working a little bit harder with the primary (and now only master EQ) seems to have solved it.

So now my master is bit more stripped down:
Compressor
Master EQ
Limiter

I do use MIR. The Teldex Studio Wide Venue with the MIC and EQ presets that comes with the download from the user area. I use MIRacle on the Master in VEP5; the Teldex Length Enhancer and I got my dry/wet ratio set to -30. Additionally I use an EQ on each instrument in VEP mostly playing around with presets. My instrument profiles in MIR mostly end up on the Pure option.  It’s beginning to sound a little better. Adding a picture of my new setup of anyone has any other suggestion to what I could improve.

https://www.dropbox.com/s/h0jo52dxhrlbw8p/Master-mix-2.jpg

Thanks all so much!

[Beat Kaufmann]

Hi Jhonny

Thanks for your kind words about all our feedbacks.

Happy Easter

Beat

[Dietz]

Hi Jhonny,

I can't see all the details on the screen-shot, but if that EQ on the right side is supposed to sit in the master-bus, I suspect that these huge cuts won't do any good. These are settings I would expect to see in a single channel, but hardly as part of a master-bus chain.

Personally I'm always trying to fix a problem at the source, and if you _really_ need these hard cuts, you should try to find our why you need them. It could very well be that there is an issue with the arrangement you're working on rather than with the actual sound.

Kind regards,

[Dominique]

Just a quick thank you for all posters here. This thread is immensely helpful.