Home Theater Forum and Systems banner
1 - 20 of 33 Posts

·
Registered
Joined
·
291 Posts
Discussion Starter · #1 ·
"How hot should my levels be when I'm recording or mixing?" is a question that is asked quite
frequently on all sorts of internet forums and in general conversations with aspiring recording engineers and those new to mixing.

Gain staging refers to one's signal path, going from the source through the microphone pre amps, converters and
finally to your DAW. This is a very critical moment for your signal; the goal is to have the cleanest signal possible.
(also refers to S/N ratio on your mixing console). How you calibrate your converters will play a heavy role in this equation.
Generally, you hear "record as hot as possible without clipping" this is true to a point, what I mean is, yes you want to record
as hot as possible, but that does not necessarily mean being in red/orange!

In the digital realm, one is able to record at 24-bit which offers an astounding amount of headroom (144dB dynamic range to be exact);
as such you do not have to push the signal as hot as you normally might with an analog console and tape machine. 0dBu is
equivalent to -18dBFS digital, so that is something one should keep in mind when calibrating their converters for recording.
A good recommendation would be to have your signal peaking around -6 to -5.

You may be questioning this, however every console is different, so finding the proper balance of signal to noise will also
depend on the console you are using, you may need to go a little hotter or a bit less, that's fine! Do your best to peak around -6 though.

When mixing try to keep your levels in the -18 range on your faders. Kicks, snares and vocals can peak around -6 which
is fine. All of this is leading you into the mastering phase so it is a good idea to leave plenty of headroom for your mastering
engineer to do his/her magic.

In general, you want to keep your master bus peaking around -3 which will give the mastering engineer plenty of headroom
to work with. Some engineers will make their mixes so hot and squashed there is almost no dynamic range left! Just a big giant brick
for a waveform after mastering (or sometimes before) which is totally unnecessary.

However, a client may claim the tracks aren't loud enough. It is part of your job as an engineer to explain why it is not
advisable to have the tracks so hot. Most musicians don't think about levels as we do because they forget that what they
are listening to is already mastered! They are unable to hear or see the final mix and just believe it needs to be super hot.
Inform them otherwise and stand firm about this subject. The tracks will sound much cleaner and clear when all is said and done.
Never forget that dynamics are what makes music interesting, so make sure those tracks have plenty of breathing room!

All of these tips and steps will help you to avoid distortion and noise in your tracks so I do hope that you follow this crucial
step in the audio recording/mixing process. If you are only mixing and get hot tracks, use a trim plug-in to get the gains under
control as well or use the fader, either way works, but a trim plug-in will never steer you wrong in keeping gains under control.

Look out for more of my articles and keep on improving.
 

·
Registered
Joined
·
63 Posts
Excellent advice, the simple notion being that it's only 1 bit for every 6dB. I always track with peaks at around -12dBFS for each track. It makes it so much easier to mix too, when you consider where the fader moves have to be to add a track to the mix.
 

·
Registered
Joined
·
291 Posts
Discussion Starter · #3 ·
Yes indeed, instead of large moves, all it takes are 1/10th of a dB moves :D. I say -6 peak just for S/N ratio sake btw. That just ends up
going into my comment about every console being different though.
 

·
Banned
Joined
·
66 Posts
Note that only a tiny select handful of 24bit A/D converters have real 144db dynamic range. Most, though 24 bits, behave more like 19 or 20 bits in real life.
 

·
Registered
Joined
·
103 Posts
It's more than just a headroom issue: plugins which emulate hardware (and some which don't) also have operating ranges and reference levels. Sure, you may not clip your 32FP plugs but in order to get the proper saturation/behavior/attenuation/etc. you need to be in the ballpark not just blasting -1dBFS into every input.

There is a huge thread over at Gearslutz (I think it's up to 117 pages now) about gain staging in digital ITB mixing. In that thread Paul Frindle said:

Paul Frindle said:
I helped to design much of the celebrated analogue consoles mentioned, much of the digital stuff - and more recently several of the most highly regarded plug-ins. And believe me, this is the single biggest issue affecting the quality of your digital work.

The biggest 'own goal' of the whole of the digital industry from the outset has been the total disregard for the notion of operating levels and headroom.

Do not accept it - if your production environment dictates that you must provide limited, maximised and level-blasted material, just do it at the end :)
Unfortunately there are some real misconceptions about gain staging and headroom in digital and I have seen people routinely attacked all over the web for suggesting that proper gain staging still has its place (or maybe even more of a place) in DAWs.
 

·
Registered
Joined
·
34 Posts
Great thread~! I especially agree that most music is mastered WAY too hot these days, with no dynamic range at all. There has been alot said already about this, but it can't be emphasized enough..

I can't even listen to some recordings these days, which contain otherwise good music, but are so squashed with no dynamic shading at all that they're un-listen-able. My reference for a modern recording with good dynamic range and the impact and drama that it adds to music is YES - Talk from 1993. Huge dynamic swings which, if weren't there, would remove much of the emotional impact of this great music.

Gain structure is something that is also important when setting up DSP's for installed sound. You need to make sure that you're not clipping the A/D converters at the front end, then have the signal hot enough but with adequate headroom for all the processing within the DSP so that you don't end up with a 'grainy' sound on the output. You can hear it, especially on spoken voice. If you're levels are too low inside the DSP you're not giving enough bit-depth for the processors to work with. When you make up the gain on the output side you're just amplifying the grunge.
 

·
Registered
Joined
·
91 Posts
I always track with peaks at around -12dBFS for each track. It makes it so much easier to mix too, when you consider where the fader moves have to be to add a track to the mix.
I agree with you at some point. I agree about -12dBFS. I work in film sound and usually my peaks are between -25-22 dBFS for quieter dialogues and between -18 - 12 dBFS on a louder staff with occasional peaks @ around -6dBFS. Most of the times I record @ 16bit (96 dB headroom) @ 48 kHz, the reason for that is a different mix levels in film post production. Usually calibrated at around 83-86 dB. A normal listening levels. So I don't really need to blast the levels during recording the dialogues. First of all by bringing up the gain levels you obviously bring up the SNR, which I personally don't think is good :nerd: bunch of unnecessary BG noises and etc. And the second reason is, from what I know according to a comparison of Analogue and Digital levels, analogue Peaks harshly @ -12dBFS in Digital domain well at least during recording. Thus when you run a analogue Pre of the console to a digital source you have to calibrate the gain structure to a -20dBFS, which is equal to 0VU. I strongly believe in to a rule that the quieter sound have to be a quieter in levels then a louder sound, i.e. different Headroom as well as a bit higher SNR, but of course you want to record it as clean as possible, that could be fixed by a close miking technique. It's important to keep the recording intelligible.

Kirill.
 

·
Premium Member
Joined
·
15,054 Posts
I cant agree more with what has been said so far. Withe the old analog boards clipping cold still happen without much noticeable distortion unless you really hit the roof hard. Clipping a digital signal means it becomes unusable.
-12db's is a very good target and like has already been said allot of the new recordings out there are so hot that there is no dynamic range and everything seems flat. I personally like a nice punchy bass and deep kick and with most recordings this gets lost.
With todays uncompressed formats available on BluRay discs dynamic range plays a big part in the movies feel. Jazz is a great example of recordings that need to be dynamic and if its recorded like some of the pop music today there is no feeling.
 

·
Registered
Joined
·
1 Posts
Hi,
I would highly recommend checking Bob Katz's site called digido. I've calibrated my studio to the K-14 system and had great results. FWIW
Scott
 

·
Registered
Joined
·
33 Posts
Great thread! I have been following that juggernaut over on GS for a long while now too.....


Two things I would like to add:

• One must understand the difference between peak levels (instant) and VU or RMS levels (averaged over time) and that simply setting a peak level might not be the right thing to do. Bob Katz's site is a great start to learn that.

• There is no adopted standard for 0 VU (+4dBu) in digital. The EBU broadcast standard is -18dBFS= +4 dBu whereas the US film community uses -20dBFS=+4 dBu.

Differing converters will take a true 0 dBu 1k tone and register wildly varying levels from different units. It depends on their internal operating level which is usually not stated in specs.

What you can do to figure this is look at the maximum operating input and output levels and subtract from there. for example:

RME ADI-8 has a max output level of +21dBu. Since pro levels are referenced against +4dBu, we can say that a nominal 1k tone coming into the unit at +4 dBu will show up as -17dBFS. In this case, you would have to pad down the inputs on the RME by 3dB to get a professional 20dB worth of headroom over nominal operating levels. Most pro console peak out at +24dBu for example, just a bit more for good measure.

Now the RME can still have a great deal of dynamic range even with this limitation. You just have to pad down the inputs and crank up the outputs by 3dB to adjust the nominal operating level to be in line with professional outboard gear. That way, all your killer preamps will function as they should. Going straight into the RME, a pro preamp will not be turned up as much as it should be, losing such things as transformer saturation and other potentially desirable things while having perhaps some more noise than usual.

-ashley
 

·
Registered
Joined
·
91 Posts
I agree with all the points here. I remember those little differences in meters. We had a whole chart in class when I was at sound school. I find it a bit stupid though. I understand the differences and staff but why can't people just make one standard for once, and in case if you change it, you change it every where at once...

Kirill
 

·
Registered
Joined
·
48 Posts
Ont thing not to forget: 24 bit digital with a headroom of 144db is a lot of level!!!

Analogue sources can provide about 110 to 132 db of clean signal. You don't need to push your analogue source to the limit, because your front end preamps will begin to sound harsh(even in the best converters on the planet). Please leave it to the digital domain to have more gain, i.e using maximizer (plug-ins or other tweaks) to do the job. At the end you will end up with clean soun that is well recorded, loud and clean... Dont forget that before going digital, your analog section of your converter is still analog...) so pushed to the limit it will distort...so leave it to the digital processing and you will have lots of "apparent gain" without the distortion artifacts and harsh sounding "loud" tracks.

Don't forget that some mics are able to handle 140+Db but your preamp does it???:gulp:
 

·
Registered
Joined
·
91 Posts
Very good point Maxserg. One thing you forgotten to mention. How to figure out the headroom in case of different bit rates.

As an example let's take 16 bit:

Every bit contains 6dB of headroom, thus by multiplying 16x6=96dB. Meaning that there are 96 dB of headroom in 16 bit recorder for example. So going to 24bit it becomes 144 (24x6=144dB) 32bit is even more, 192dB (32x6=192dB), which is devastating levels, BOOOOOOOOOOMMMMM!!! :sn:

Just a little addition to the Maxserg post.

Kirill :T
 

·
Registered
Joined
·
48 Posts
Let's not forget that this beautiful thing (24 BITS) is almost something of another world(CD=16bits). Who can enjoy the benefit of 24 bits dyn. range? We mostly listen from IPods or others brands. So the question remains: how loud can you record (anyway you can control the playback volume isn'it):sn:

I would like to se with all the technologies that we could listen to higher fidelity but it doesn't seem to be the case(sorry for my english):blink:
 

·
Registered
Joined
·
79 Posts
"How hot should my levels be when I'm recording or mixing?" is a question that is asked quite
frequently on all sorts of internet forums and in general conversations with aspiring recording engineers and those new to mixing.


[/SIZE][/FONT] Look out for more of my articles and keep on improving.
This entire thread should be required literature in every DAW system sold, along with the WARNING! DO NOT USE/INSTALL UNTIL YOU HAVE READ THIS!

I recently recorded a CD and the band leader decided to add some tracks (in Cubase) to the stereo reference mix I had given him.. he later brought it back for mastering and EVERYTHING was destroyed... I asked what had happened to the levels, and his reply was "Oh, all the levels were so low, I brought them all up so it'd be louder." Apparently, he has no "volume control" on his monitors.....
Zero headroom to work with, digital clipping, ruined an exquisite recording.
I won't even go into how bad the added tracks were.
 

·
Registered
Joined
·
192 Posts
Very good point Maxserg. One thing you forgotten to mention. How to figure out the headroom in case of different bit rates.

As an example let's take 16 bit:

Every bit contains 6dB of headroom, thus by multiplying 16x6=96dB. Meaning that there are 96 dB of headroom in 16 bit recorder for example. So going to 24bit it becomes 144 (24x6=144dB) 32bit is even more, 192dB (32x6=192dB), which is devastating levels, BOOOOOOOOOOMMMMM!!! :sn:
It's actually more than that, because 32-bit is generally a floating point rather than an integer value, which gives an effective dynamic range in the fairly insane range of approx 1500dB... :hsd:
 

·
Registered
Joined
·
103 Posts
As was pointed out earlier, gain staging in digital isn't really a matter of not clipping; as has been said, 140+dB is a lot of dynamic range, you'd have to really not know what you're doing to end up with a clipping problem. But when you start using saturation/distortion effects, compressors or any other effects (in or out of the box) which are very much signal dependent, your results will benefit not only from some uniformity but also in that they will be processed in the way the software/hardware was intended to work. I have plenty of software which assumes -18dBFS as 0VU and would produce pretty cruddy results if I try to drive it too hot.

The "gain staging in digital is all about even more headroom" paradigm is a red herring and one that's easily refuted (since there is plenty of headroom in digital). It obfuscates other real reasons to maintain "analog-style" gain structures in ITB/digital mixes.
 

·
Registered
Joined
·
28 Posts
Ont thing not to forget: 24 bit digital with a headroom of 144db is a lot of level!!!
Analogue sources can provide about 110 to 132 db of clean signal. You don't need to push your analogue source to the limit, because your front end preamps will begin to sound harsh(even in the best converters on the planet). Please leave it to the digital domain to have more gain, i.e using maximizer (plug-ins or other tweaks) to do the job. At the end you will end up with clean soun that is well recorded, loud and clean... Dont forget that before going digital, your analog section of your converter is still analog...) so pushed to the limit it will distort...so leave it to the digital processing and you will have lots of "apparent gain" without the distortion artifacts and harsh sounding "loud" tracks.
Don't forget that some mics are able to handle 140+Db but your preamp does it???:gulp:
Just where would I go to get an analog source that can handle 132 DB ? And where would I find a room that it mattered?
 
1 - 20 of 33 Posts
Top