TranceAddict Forums (www.tranceaddict.com/forums)
- Production Studio
-- Why do people say things like "I'm mixing at -4db"?
Pages (3): « 1 [2] 3 »

quote:

Originally posted by kitphillips ^^^ Its worth the risk because you get more resolution on your meters the higher you mix. There's very little visual difference between -60 and -70, but a whole lot between -3 and -1.

lol that made no sense. How does your meter help you to mix other than tell you that you went over board which you wouldn't have to worry about if you weren't running things so hot.

quote:

Originally posted by Mad for Brad lol that made no sense. How does your meter help you to mix other than tell you that you went over board which you wouldn't have to worry about if you weren't running things so hot.

So you only ever use meters to tell you when your clipping? Why don't you just get a little light that comes on when your clipping? Theres a reason why there are meters rather than just little lights, and its to help you work out what your levels are doing, not just tell you when you when you're clipping.

quote:

Originally posted by G-Con ok cool. Think from now on I will mix as close to 0 as possible then. Think I'd prefer this as any little spikes in volume would cause my master to clip which I can then fix. Where as now, little spikes can go unnoticed when I'm mixing nearer -6. I asked about rendering twice, purely so I could render the file near 0. Then if and when I decided to send to a mastering engineer, I can load up the rendered wav and drop the volume down and render which would be much quicker than loading up and rendering the entire project. However, based on your advice, I will just render the original project again if required. Like you say, its probably all pointless anyway, I'm sure my own production skills will be the cause of any lack of quality to the listener rather than sample rates/dithering/bit-depth etc etc.

Yeah, that was my point above... I find it easier to mix at a higher level because you get more feedback from your meters. It just seems to get me better levels.

So yeah, render your file (with your mastering) at 0 dB, then if a mastering engineer wants something different, just throw a gain plugin on there to get it set right.

I don't honestly think that many mastering engineers ask for -6 dB peak level though, they can set that themselves. What they might ask for is -6 RMS level, which is a very different thing and a bit harder to achieve. Different debate though innit

I mix at everything from +5db to -10db pretty cool eh ?

quote:

Originally posted by kitphillips So you only ever use meters to tell you when your clipping? Why don't you just get a little light that comes on when your clipping? Theres a reason why there are meters rather than just little lights, and its to help you work out what your levels are doing, not just tell you when you when you're clipping.

pretty much. It is an approximation. They aren't exact and to make your mix hot risking clipping to get more meter resolution is rather retarded.

quote:

Originally posted by kitphillips What they might ask for is -6 RMS level, which is a very different thing and a bit harder to achieve.

-6 RMS level is insanely loud. That is louder than Metallica's Death Magnetic, widely considered one of the best (worst) examples of abusive compression.

quote:

Originally posted by kitphillips ^^^ Its worth the risk because you get more resolution on your meters the higher you mix. There's very little visual difference between -60 and -70, but a whole lot between -3 and -1. So it makes sense in terms of being able to see what your mix is doing to just mix at a certain high level and then pull everything down immediately before exporting. I was going to come in here and say exactly this. How many times do we need to have this same thread?? And honestly, all these debates are a bit ridiculous considering the quality of most of the productions in EDM. I can count on one hand the number of producers out there who actually have the mixing skills to even be worried about these sorts of tiny, highly academic nitpicks.

Kit is absolutely right. Firstly, it's not just about meter resolution (which is a good point in itself) but about usable dynamic range.

And as kit pointed out, that last 6dbfs is a big difference in perceived loudness because audio is logarithmic, not linear.

So by not mixing to 0db you're effectively not using a lot of available mixing headroom - essentially your usable range is less where it counts most.

Also, if you record anything at all, you're just introducing noisefloor at a lower relative point (all recordings have some inherent noise so by mixing to -6dbfs peaks, when you then later bring up the signal in mastering to 0dbfs, you are increasing the noise floor as well).

I of course don't use the meters alone, but from an engineering standpoint they are pretty vital in a lot of instances to figure out what's going on so why not make the most of them.

Finally, if you're thinking about having mastering done on your track, then you really shouldn't be rendering, you should only be printing your mixes (as in setting a track up to record the final sum of the mixed track). There's many reasons for this (which I won't go in to on here because it's been done before) but the bottom line is that you get a better result.

This means however all you actually need to do is lower that fader on the record track by 6db but I was desperately trying to avoid giving this info as someone will read it "OK great, I can lower the master", which is bad idea for a number reasons and not what i am saying.

dB's are a logarithmic conversion but they are linear when compared to each other. The difference in energy subtracting 10dB will always be the same so in this regards, it is a form of derivative and is definitely linear. And do you honestly need those 3 dB when your mix is probably so damn loud it won't make one bit of difference. It just isn't worth the headache when working at 24 bit.

And meters might be important but I have yet to meet an engineer that has anything to saw regarding almost every DAW meter. They just aren't accurate. And are they that much accurate from 0 3 dBfs rather than - 6 9 dBfs. That is just pedantic silliness.

And honestly, the notion of noisefloor regarding EDM in itself is kinda worthy of a good laugh. Noisefloor just doesn't matter when you make loud music.

I suppose my end point is that mixing to - 6dBfs makes no difference in sound and anyone that thinks it does is imaging things. So why not ? The AES actually encourages - 18 dBfs. I think they know what they are talking about. And RANN, for video , the convention is it not - 20 dBfs ?

quote:

Originally posted by Mad for Brad And honestly, the notion of noisefloor regarding EDM in itself is kinda worthy of a good laugh. Noisefloor just doesn't matter when you make loud music.

Or highly synthetic music recorded mostly either directly in the computer or though cables, rather than mics in a room.

quote:

Originally posted by Mad for Brad dB's are a logarithmic conversion but they are linear when compared to each other. The difference in energy subtracting 10dB will always be the same so in this regards, it is a form of derivative and is definitely linear. And do you honestly need those 3 dB when your mix is probably so damn loud it won't make one bit of difference. It just isn't worth the headache when working at 24 bit. And meters might be important but I have yet to meet an engineer that has anything to saw regarding almost every DAW meter. They just aren't accurate. And are they that much accurate from 0 3 dBfs rather than - 6 9 dBfs. That is just pedantic silliness. And honestly, the notion of noisefloor regarding EDM in itself is kinda worthy of a good laugh. Noisefloor just doesn't matter when you make loud music. I suppose my end point is that mixing to - 6dBfs makes no difference in sound and anyone that thinks it does is imaging things. So why not ? The AES actually encourages - 18 dBfs. I think they know what they are talking about. And RANN, for video , the convention is it not - 20 dBfs ?

You're forgetting one very important thing. They are linear when compared to each other but not to our hearing: +3db is twice as loud so guess what 6db is? That range of 6db is therefore what you're losing in a digital system,because that 6db is link directly the relative analogue output of your speakers. Sure you can compensate for it, but really why bother.

And comparing DBVU to DBFS doesn't actually indicate anything about the point at hand: why leave (x)db of headroom on a digital system when you mix?

Bear in mind that calibration to the K system (as shown on the diagram) means that you adjust your signal so that you have available headroom between your RMS and available peak space.

That reference there is only if you calibrate to those marks. 0dbfs is whatever you calibrate to, and that diagram is a reference to the K system.
By this I mean if you calibrate your monitors correctly to the K system, then 0dbfs is 83dbspl with your matrial at -20db RMS average.

I can prove this by the fact that if this were not true (and that diagram is an accurate representation of a production system) then -6dbfs would be +14dbvu which is far too hot. But now think about it....

Do you see that the reason we calibrate with a -14dbfs bias (minimum)?

Anyway, this is all a bit pointless as I can probably count on one hand the number of people on here who actually work a fully +4dbvu system.

Furthermore We're not taking analogue in to the digital world as per that diagram - we're producing in a purely digital system, so our available perceived dynamic range on a calibrated system is zero to 0dbfs. You calibrate a digital system to the k-system (and you're right in that at the studio we work to -18dbfs for score) primarily so that you always have a common point of reference with enough dynamic range and your ears will bleed before you clip.

And this is why I say for those working in a a digital system, you are better off using all that is available to you.

My cock is far too small to enter this fight. But it's interesting to watch nevertheless.

i thought all that matters was to avoid clipping/distortion? and to be able to convert the wav into mp3 without that distorting too? -0.2 seems to work for me.

I've read each reply - thanks for the input all.

Just goes to show though that this production lark might be likened to politics - 2 camps with opposing views both of which are CERTAIN they have the right answers.

thats why in most countries except US theres more than two political partys, somewhere in the middle is the right answer.

MadforBrad, I don't really understand your point. RANN pretty much took the words out of my mouth when he pointed out that we hear logarithmically, meters are displayed logarithmically too, so its important to use that top 3 dB of dynamic range well, since its the most sensitive 3 db on the scale.

Whether or not the meters on DAWs are accurate, the fact remains that they ARE MORE accurate between -3 and 0 than between -9 and -3. Its because they're logarithmically scaled, which means that the top 3 dB represents more SPL range (in terms of particle displacement) than does the next 3 dB down.

Further, there is no headache involved in working this way, its easier, thats why we do it. I'm just pointing out that your conventional wisdom that you shouldn't clip channels and plugins, and should mix low to avoid doing so simply doesn't apply in the modern 32/64 bit age.

At the end of the day, my point is that you can work however you like in a 32 bit system, because you have enough resolution that you

1/ won't clip if you go over 0 BEFORE the master
2/ won't lose much resolution if you render 3 dB lower, as long as you render in 32 bit. BUT if you're rendering in 16 bit, you should definately normalise beforehand to make the most out of this slightly sloppy format.

i red on gearslutz that normalizing fucks up the render? why is that?

i was putting my kick at -9db and then everything once mixed would usually peak at -4 but ive noticed over the past 6 months ive been getting it lower and lower so now ive been putting my kick -8 or so and its getting everything at less than -4db which is good.
so although theres ppl on here with more experience than me.. id recommend to try that start with your kick and -8db and try and get your final mix to not peak above -4db (without changing the master fader haha)

quote:

Originally posted by kitphillips MadforBrad, I don't really understand your point. RANN pretty much took the words out of my mouth when he pointed out that we hear logarithmically, meters are displayed logarithmically too, so its important to use that top 3 dB of dynamic range well, since its the most sensitive 3 db on the scale. .

we hear logarithmically but dB is a log based unit which means, unless you've never done any calculus, operating strictly with dBs is a linear matter. Every 10 dB you add or subtract is widely considered a perceived doubling or halving of loudness.

and djRann is only making his points assuming you all don't have calibrated systems and you all are working in a purely digital world. His advice is really for people that don't know what they are talking about as most of you probably don't.

He still didn't address the fact that 3-6dB of headroom in a 24 bit system is not a big deal. There is no difference in sound quality so why try to make things as hot as possible. People that actually know how to use meters are generally using a large mixing board and would be mixing to the references I posted. The visual meter is not so important that you would needlessly make your mix so hot that you will end up with something that might clip. It is an unnecessary risk and there are no benefits sonically speaking.

quote:

Originally posted by Mad for Brad we hear logarithmically but dB is a log based unit which means, unless you've never done any calculus, operating strictly with dBs is a linear matter. Every 10 dB you add or subtract is widely considered a perceived doubling or halving of loudness. and djRann is only making his points assuming you all don't have calibrated systems and you all are working in a purely digital world. His advice is really for people that don't know what they are talking about as most of you probably don't. He still didn't address the fact that 3-6dB of headroom in a 24 bit system is not a big deal. There is no difference in sound quality so why try to make things as hot as possible. People that actually know how to use meters are generally using a large mixing board and would be mixing to the references I posted. The visual meter is not so important that you would needlessly make your mix so hot that you will end up with something that might clip. It is an unnecessary risk and there are no benefits sonically speaking.

I don't know what your point is about log scales? Its irrelevant IMO. I'm aware of how log scales work. This issue is really about whether its sensible to mix to 0 dB or if there is still a reason to mix to -3. I'm claiming that there is no longer a reason to mix to -3 in modern DAW systems. You're saying that there is, but haven't provided a reason. If we both agree that it doesn't sonically matter, then we have no argument. If you find it easier to mix to -3 then fine, but from the perspective of metering etc, I find it easier to mix to 0. I think we both agree that it doesn't matter for any reason other than conveniance though.

I think we'd both also agree that there are very small quality losses when you export at less than 0, so that its better to export at 0 through normalising, especially when exporting to 16 bit.

Do I understand you correctly?

my point is that you should learn the correct way to do things regardless of your current set up as you will might have to work in the analog world someday and it is just a good habit to mix at the appropriate levels.

No there is no difference but why not just learn the right way to do things. Like Rann said, don't touch the master but does it make an audible difference ? no but why not just learn the right way to do things.

Another analogy is orchestration. With the use of sample libraries, the rules of orchestration don't apply as much as the balance can easily be fixed by a fader but I suppose I am someone that prefers to know how to do it the right way from the start and not develop bad habits that will eventually encumber you if you go far enough.

quote:

Originally posted by Mad for Brad my point is that you should learn the correct way to do things regardless of your current set up as you will might have to work in the analog world someday and it is just a good habit to mix at the appropriate levels. No there is no difference but why not just learn the right way to do things. Like Rann said, don't touch the master but does it make an audible difference ? no but why not just learn the right way to do things. Another analogy is orchestration. With the use of sample libraries, the rules of orchestration don't apply as much as the balance can easily be fixed by a fader but I suppose I am someone that prefers to know how to do it the right way from the start and not develop bad habits that will eventually encumber you if you go far enough.

I'm saying the exact same thing and IMO, the right way to do things is to mix to 0dbfs, not -3 or -6.

Why? because if your system should is calibrated to the K system, which already means that at 0dbfs, you are already accounting for the -18db you spoke of earlier. This is what we do at the studio so we aim to get as close to 0dbfs as possible without any clips.

one of the (many) advantages of this is that we mix on real faders, so that means if we were trying to mix to -6db, we would effectively have less space physical to mix with and less resolution in real DB terms.

I also think this is especially important for those reading this that do not have a calibrated system to not mess about with mixing to standards they don't generally understand and just aim for 0dbfs as a way to teach them good practice and to work with levels.

Bob Katz states in his book to stay clear of O dBfs and that is someone who's word I trust. To put it in perspective, A 24 bit mix has 48 more dB headroom than a 16 bit mix. Check out pages 61 where he mentions the practice of safe levels regarding recording and mixing.

quote:

Originally posted by Mad for Brad Bob Katz states in his book to stay clear of O dBfs and that is someone who's word I trust. To put it in perspective, A 24 bit mix has 48 more dB headroom than a 16 bit mix. Check out pages 61 where he mentions the practice of safe levels regarding recording and mixing.

But he's advocating staying below 0dbfs. If I have one hit on 0dbfs it's time to revisit the mix. The problem I have with this is that he never actually explains why to stay at -3dbfs, just to stay away from it.

I know you know what projects I've worked on that were mixed to a fraction below 0dbfs and had no problems whatsoever, I just don't get why Katz is asserting this. Don't get me wrong he's the master and that's why the K system is named after him but he's never clarified this point about -3dbfs apart from esoteric opinion, which I'm not buying until he proves otherwise.

But it is possible to hit 0 and your meter doesn't show it clipping. You have to have several samples at 0 although i'm not quite sure what most daw's use as the window length. I think he is just advocating it because it is better to allow a margin of safety.

quote:

Originally posted by Mad for Brad But it is possible to hit 0 and your meter doesn't show it clipping. You have to have several samples at 0 although i'm not quite sure what most daw's use as the window length. I think he is just advocating it because it is better to be allow a margin of safety.

Well that I get. The amount of times we've pressed play on PTHD and got a different result as to whether were clipping is not funny, and frankly worrying. But 0.1 or 0.2dbfs is enough margin to make sure you're not clipping so again I mix to about that point.

Personally I feel -3dbfs as a margin is excessive and in a calibrated system.

Probably because he has received so many mixes with distortion that he would rather people stay well away. I suppose someone that knows what they are doing can mix at the level you recommend but they are a rare breed.

Powered by: vBulletin
Copyright © 2000-2021, Jelsoft Enterprises Ltd.