Can someone explain -18 db mixing?
- Thread starter XxSicRokerxX
- Start date
that's usually about where i end up after leaving the loudest element of the mix at 0 and pulling everything else down as needed
Nimvi
Member
I disagree, because many plugins still benefit from working around this -18 mark (or -20 or -12, depending on calibration).
An example you may relate to is a VCC channel. Its saturation is fully dependant on the inputlevels it gets.
Let's say you have a track that is constantly peaking at -0,1dbfs before doing anything to it. It isn't clipping yet, so it isn't completely lost, but if you load VCC on this track, you will see that its VU meter is constantly hammering in the red. This causes a lot of saturation and possibly even distortion, which is probably not your intention.
By using a trim-plug before VCC (or the input knob on VCC itself), you can get the VU meter to hang around the 0dbVU mark within VCC again, and you will gain much more control over the amount of saturation by the plugin again.
You are completely right about the recording part though. Ideally, this kind of practice should be used from the start. Sadly, we often get assignments where the tracks have already been recorded too hot, so it's just damage control from there on.
And for the people being confused about where the peak/average should be, imagine a rubber ball bouncing up and down. The -18dbfs mark would be wherever the ball is when it is halfway up or down after a bounce. This doesn't mean it stops moving at this point (peaking). It's just an average. Sometimes the ball will bounce up a bit higher, sometimes a bit lower, but it still averages around that point halfway through.
So your tracks might be peaking at -12dbfs, and fall down to -30dbfs every now and then, but the average should reside around -18dbfs. Personally, I think this is easiest to see with a VU meter plugin, because the needle on those usually moves a bit slower than a daw-dbfs-meter, and it becomes easier to get a feel for the average level against the peaklevels. Just make the VU meter bounce a bit around the 0dbVU mark and you should be good.
And to avoid more confusion: to me, this counts only for the individual tracks. When they sum together on the masterbus, they will probably end up around -6dbs, give or take, like uros mentioned. Which leaves some nice room for the mastering engineer as well.
Uros
Sonic Incision
^^hahaha the rubber ball analogy was actually quite a clever way of describing peaks and average levels 
Btw I had a private message about this sent to me today, so I'll just copy here what I wrote about how I do it, systematically. It is not the only way of doing it, but my mixes have improved since I adopted it. I don't like that feel of being lost, so that's why try and follow it most of the time. This also answers The-Zeronaut's question.
If you do it like this, you will get output ~0dBVU on 2buss while using plugins on their operating levels at all times, which can't be a bad thing since most of them are modeled with that in mind (even if some of them are not, it still is a good practice), and you will get good final 2buss level without unnecessary 'wasting' of fader resolution, which you need for automation.
edit:
Ohhhh, and if you mix into a buss compressor, this way you can 'calibrate' it with a ~0dBVU signal. You can use 1kHz test tone for that right at the very beginning, since you already know signal level you'll gonna feed it with even before you start any mixing (Michael Brauer does it too, and he's a master of buss compression, so he can't be that wrong )!
Of course that it'll require some tweaking when you actually advance into later stage of the mix, but those tweaks shouldn't be too drastic (in fact, they should be quite minor), if you do it 'right'
Btw I had a private message about this sent to me today, so I'll just copy here what I wrote about how I do it, systematically. It is not the only way of doing it, but my mixes have improved since I adopted it. I don't like that feel of being lost, so that's why try and follow it most of the time. This also answers The-Zeronaut's question.
- trim input to ~0dBVU on every track (again, use Satson's, Klanghelm's or some other VU meter, that's the most intuitive way to do it IMO)
- apply plugins on channel inserts with unity gain in mind
- do desired balancing between tracks via channel faders
- send them to a buss (for ex. send close and OH mics to a drum buss). Repeat the same process on other tracks & busses (sums of vocals, guitars, keys, etc.)
- trim input on busses to ~0dBVU
- apply plugins on group channel inserts with unity gain in mind
- do desired balancing between busses via group channel faders
- send them to 2buss
- trim the 2buss' input to ~0dBVU. Apply plugins with unity gain in mind
If you do it like this, you will get output ~0dBVU on 2buss while using plugins on their operating levels at all times, which can't be a bad thing since most of them are modeled with that in mind (even if some of them are not, it still is a good practice), and you will get good final 2buss level without unnecessary 'wasting' of fader resolution, which you need for automation.
edit:
Ohhhh, and if you mix into a buss compressor, this way you can 'calibrate' it with a ~0dBVU signal. You can use 1kHz test tone for that right at the very beginning, since you already know signal level you'll gonna feed it with even before you start any mixing (Michael Brauer does it too, and he's a master of buss compression, so he can't be that wrong
)! Of course that it'll require some tweaking when you actually advance into later stage of the mix, but those tweaks shouldn't be too drastic (in fact, they should be quite minor), if you do it 'right'
I'm just gonna post a quick question and come read the whole thread another day because I'm so tired my eyes are drying up.
But yeah, so if I got a vu+peak meter I should aim between -12 and -6 and then what ? Where can you push the master once your mix is done ?
What is headroom, a technique to avoid clipping ?
But yeah, so if I got a vu+peak meter I should aim between -12 and -6 and then what ? Where can you push the master once your mix is done ?
What is headroom, a technique to avoid clipping ?
headroom refers to the amount of amplitude between the source audio and 0 - and when it's all used up, you're clipping
think of your 2-buss like a big empty water jug, and each individual track as a glass. as you fill each glass and pour it into the jug, it obviously will start to fill up. the more water in each glass, the more full the jug gets. whatever amount of space is left in the jug after all the glasses are poured in would be the equivalent of your headroom. if you pour in too much and you're spilling out the top, that's the equivalent of clipping. the only way to prevent doing so is to lower the level of water that's being poured into each glass.
think of your 2-buss like a big empty water jug, and each individual track as a glass. as you fill each glass and pour it into the jug, it obviously will start to fill up. the more water in each glass, the more full the jug gets. whatever amount of space is left in the jug after all the glasses are poured in would be the equivalent of your headroom. if you pour in too much and you're spilling out the top, that's the equivalent of clipping. the only way to prevent doing so is to lower the level of water that's being poured into each glass.
Nimvi
Member
To sum up in conclusion:
Make sure your balls are bouncing at a decent average height, so your giant jug doesn't run over.
Questions?
Make sure your balls are bouncing at a decent average height, so your giant jug doesn't run over.
Questions?
The-Zeronaut
Mixing..Y U SO DIFFICULT?
- Sep 24, 2007
- 1,794
- 5
- 38
^^hahaha the rubber ball analogy was actually quite a clever way of describing peaks and average levels
Btw I had a private message about this sent to me today, so I'll just copy here what I wrote about how I do it, systematically. It is not the only way of doing it, but my mixes have improved since I adopted it. I don't like that feel of being lost, so that's why try and follow it most of the time. This also answers The-Zeronaut's question.
- trim input to ~0dBVU on every track (again, use Satson's, Klanghelm's or some other VU meter, that's the most intuitive way to do it IMO)
- apply plugins on channel inserts with unity gain in mind
- do desired balancing between tracks via channel faders
- send them to a buss (for ex. send close and OH mics to a drum buss). Repeat the same process on other tracks & busses (sums of vocals, guitars, keys, etc.)
- trim input on busses to ~0dBVU
- apply plugins on group channel inserts with unity gain in mind
- do desired balancing between busses via group channel faders
- send them to 2buss
- trim the 2buss' input to ~0dBVU. Apply plugins with unity gain in mind
If you do it like this, you will get output ~0dBVU on 2buss while using plugins on their operating levels at all times, which can't be a bad thing since most of them are modeled with that in mind (even if some of them are not, it still is a good practice), and you will get good final 2buss level without unnecessary 'wasting' of fader resolution, which you need for automation.
edit:
Ohhhh, and if you mix into a buss compressor, this way you can 'calibrate' it with a ~0dBVU signal. You can use 1kHz test tone for that right at the very beginning, since you already know signal level you'll gonna feed it with even before you start any mixing (Michael Brauer does it too, and he's a master of buss compression, so he can't be that wrong
Of course that it'll require some tweaking when you actually advance into later stage of the mix, but those tweaks shouldn't be too drastic (in fact, they should be quite minor), if you do it 'right'
THANKS A LOT!!
seriously
hey there....
lots of useful info on this forum, though i got a problem/question:
i use satson to trim my tracks... the snare (sd2) comes into the cubase channel peaking at -4.1, after trimming it with satson it peaks at +4, there s no clipping...
huh? is that normal/ok?
chris
lots of useful info on this forum, though i got a problem/question:
i use satson to trim my tracks... the snare (sd2) comes into the cubase channel peaking at -4.1, after trimming it with satson it peaks at +4, there s no clipping...
huh? is that normal/ok?
chris
hurdy
Sup
im guessing satson is a VU meter??
the idea of the VU meter is to get the average peaks around 0 dbVU, it can go over or under that value, but average should be around 0 db VU
the idea of the VU meter is to get the average peaks around 0 dbVU, it can go over or under that value, but average should be around 0 db VU
So do you need a compressor on every track ?
I was just trying a sine wave and while polyphoning it peaked solid.
A single note it hit 0 on the vu meter and -4 on the peak.
When playing a 3 notes chord it tickles the peak 0 and the vu is around +5
Another question; would it be a good idea to, if quad-tracking, arm my 4 tracks and adjust my Guitar Processor'S volume to have the sum of the 4 tracks around 0 on the VU ?
- I used to track each one at 0Vu and it quickly peaked when playing all togetter so I lowered the faders
I'm asking because someone told me I should always set the output of my GT-10 to it's maximum because it's line level or something and adjust my pre-amp to fit because of Signal to Noise ratio.
Sorry for staining the thread with my questions.
I was just trying a sine wave and while polyphoning it peaked solid.
A single note it hit 0 on the vu meter and -4 on the peak.
When playing a 3 notes chord it tickles the peak 0 and the vu is around +5
Another question; would it be a good idea to, if quad-tracking, arm my 4 tracks and adjust my Guitar Processor'S volume to have the sum of the 4 tracks around 0 on the VU ?
- I used to track each one at 0Vu and it quickly peaked when playing all togetter so I lowered the faders
I'm asking because someone told me I should always set the output of my GT-10 to it's maximum because it's line level or something and adjust my pre-amp to fit because of Signal to Noise ratio.
Sorry for staining the thread with my questions.
hurdy
Sup
"So do you need a compressor on every track ?"
I dont think you understand the concept of the compressor. Its main job is to compress the signal to help control the dynamics.
Some elements may not need any compression, E.G. undynamic synth bass or sumthing with no dynamics. So most instruments will need a lil bit of compression
If your talking about mixing (adjusting levels) your tracks into the mastering compressor, then all you will need is the compressor on the master 2 buss to mix your tracks into.
Set up your master track with a compressor with a ratio of 2:1 (commonly used ratio for mastering), with no more than 3 to 6 db gain reduction
Then play a 1Khz tone thru the master to see how much gain reduction is occuring, this is called calibrating the master compressor for mixing (may need further adjustment after recording)
"Another question; would it be a good idea to, if quad-tracking, arm my 4 tracks and adjust my Guitar Processor'S volume to have the sum of the 4 tracks around 0 on the VU?"
Each track should be at 0db VU, then use trim plugin to turn all four tracks down until it reaches 0db VU summed together. Most plugins are designed for 0dbVU so thats why each track needs to be at that level. E.G. one of the four tracks may have a different EQ plugin setting, so that EQ plugin needs to ideally see about 0dbVU. Not everyone does things like this so it comes down to personal preference for some people.
The guitar processor shouldnt be on full volume as this could be distorting the output of the processor. I would set it to half way and then see how it reads thru your sound interface. If its still not loud enough, then turn the GAIN up on your sound interface. Which interface are you using?? You will need to use the instrument input with a high impedence, most likely, unless your using a DI box, which you will plug the XLR into the mic input. I recorded a bands GT-8 I think it was, and the volume was loud enough at half way. If your recording at 24 bit then noise to signal will not be an issue.
I dont think you understand the concept of the compressor. Its main job is to compress the signal to help control the dynamics.
Some elements may not need any compression, E.G. undynamic synth bass or sumthing with no dynamics. So most instruments will need a lil bit of compression
If your talking about mixing (adjusting levels) your tracks into the mastering compressor, then all you will need is the compressor on the master 2 buss to mix your tracks into.
Set up your master track with a compressor with a ratio of 2:1 (commonly used ratio for mastering), with no more than 3 to 6 db gain reduction
Then play a 1Khz tone thru the master to see how much gain reduction is occuring, this is called calibrating the master compressor for mixing (may need further adjustment after recording)
"Another question; would it be a good idea to, if quad-tracking, arm my 4 tracks and adjust my Guitar Processor'S volume to have the sum of the 4 tracks around 0 on the VU?"
Each track should be at 0db VU, then use trim plugin to turn all four tracks down until it reaches 0db VU summed together. Most plugins are designed for 0dbVU so thats why each track needs to be at that level. E.G. one of the four tracks may have a different EQ plugin setting, so that EQ plugin needs to ideally see about 0dbVU. Not everyone does things like this so it comes down to personal preference for some people.
The guitar processor shouldnt be on full volume as this could be distorting the output of the processor. I would set it to half way and then see how it reads thru your sound interface. If its still not loud enough, then turn the GAIN up on your sound interface. Which interface are you using?? You will need to use the instrument input with a high impedence, most likely, unless your using a DI box, which you will plug the XLR into the mic input. I recorded a bands GT-8 I think it was, and the volume was loud enough at half way. If your recording at 24 bit then noise to signal will not be an issue.
hey there,
sorry I guess I wasnt explaining well enough...
the snare hits the cubase channel with ca. - 4 ...
it hits satson at about - 10 ...
so after leveling it in satson at about 0 the cubase channel gives me a + 4
so, is this in accordance to your tips? am I missing something?
chris
sorry I guess I wasnt explaining well enough...
the snare hits the cubase channel with ca. - 4 ...
it hits satson at about - 10 ...
so after leveling it in satson at about 0 the cubase channel gives me a + 4
so, is this in accordance to your tips? am I missing something?
chris
OneDaySky
Clint
I have a question, almost asking about the opposite of trimming to 0dbVU
What if your signal is under that, do you increase the channel gain to 0dbVU?
What if your signal is under that, do you increase the channel gain to 0dbVU?
Nimvi
Member
Jup, especially on verbs and stuff like that, because they often love some saturation from VCC or anything like that.
Is it possible that the Satson meter isn't calibrated to -18dBfs = 0 dBVU? Or that the cubase channel isn't showing dBfs scale?
Snares are often one of the trickiest elements to do this to because they are so incredibly dynamic...but still, that sounds wrong.
OneDaySky
Clint
Cool, thx man, been looking for that answer for a while now. I normally leave levels alone if they under 0dbVU. So ill start trying to get em all at 0dbVU now.
1 more question to clarify. Are these 0dbVu readings Peak or RMS? Most DAW faders are peak meters, so im just curious as how to calculate this if its RMS.
OneDaySky
Clint
I dont think the DI matters much, as long as it doesnt clip. Its only once you reamp or use a ampsim that'll have to check the output of the amp/ampsim to get it to 0dbVU
Uros
Sonic Incision
VU meter is like an average meter, it doesn't measure peaks, because its response time is relatively slow (on purpose, around 300ms). It's built that way.
For a sine-wave tone VU meter gives a true reading of RMS, but with more complex signals (vox, gtrs, drums, etc.), it is less accurate and it will usually read slightly lower than the true RMS value.
Use VST VU meters like Satson or Klanghelm VUMT, which is only 6€, or if you prefer a true RMS meter, FreeG has one built in (so does Voxengo SPAN).
