Possible Stupid Question Regarding Mix Levels

Once it's in the box, your noise floor is basically set. The digital noise floor is absolutely, without a doubt, lower than what's on the tracks themselves and will be completely masked by that inherent noise floor of the mix itself. As long as the analog staging was reasonable going in, you're as good as you're going to get, and adding a little gain isn't really going to make your s/n noticeably worse. (Unless you use some fancy modeled plugin that adds noise for "authenticity")

7-10db is actually pretty good. As has been mentioned, you don't need a limiter if you're happy with about 19db crest factor. That's way big for most modern music, and will be much quiter than about anything else you'd have in your shuffle. If you go back to the mix, figure out why the right side peaks 3db higher than the left, and fix that[/], then you can get up to about 16db, which is still a bit quiet. Could work for some real dynamic stuff, but I'd want it a bit tighter. I'd be willing to bet that a little more dicking around with the mix could get you closer to 14db which would be pretty ok.
 
4tracker said:
Cool, thanks. I am going to compare the limited and unlimited one tomorrow.

One thing I don't understand: when I was reading up on the best way to gain stage, many people said to turn the trim down on each channel -10db and mix at low levels. Well, I did that, and now at the end of the process I have too low a signal. I am wondering if I would have been better off not doing that and using the faders. Because now I am using limiters to get back the volume I lost from that trim. Or so that is how it feels. Since I have no formal training I don't know if that's accurate.
Click to expand...
That is only good advice if the tracks are too loud to begin with. I end up doing that a lot with other people's tracks because a lot of people normalize them before they send them to me. (Even though I ask them not to)

The thing you have to wrap your head around is, first you need to look at what you have, then decide if you need to turn them down. There is no standard operating procedure that states that all the tracks you will ever get will be too loud and need to be turned down before doing anything else.

The reality is, with that last set of peak levels you gave us, you could simply turn up the volume by 7db to make it louder without a limiter and still not clip. If you want to go louder than that, you need the limiter to shave off the peaks.
 
Farview said:
The reality is, with that last set of peak levels you gave us, you could simply turn up the volume by 7db to make it louder without a limiter and still not clip. If you want to go louder than that, you need the limiter to shave off the peaks.
Click to expand...

Would I be better off doing that, or scrapping this mix and starting over with one that doesn't use a -10db trim from the start? I don't think the tracks were recorded too hot to begin with from what I hear. Maybe just the lead guitar track. I think that track is also causing the right to peak higher since I panned it. But I lowered it quite a bit.

Also, what is the best method for turning up the volume 7db without a limiter? The master fader is already at 0.
 
Last edited:
It may be at 0 - but what are the levels peaking at - I can't see why you don't experiment a little - if all the channels are a bit low, then 0dB is simply a fader position and nothing to do with the actual level, so shove it a bit. If the red light comes on because you hit the actual 0dB, then you overcooked it! During the mix, where abouts are the channel faders sitting as an average?
 
If you aren't using compression on the tracks, you can simply open the session and turn the trim up 7db on all the tracks. Then hit mix.

If you use compression, you would need to also turn the threshold up 7db to compensate.

Or you could take the mix you have, import it into a session, and add 7db of gain.

On the master buss, you can add 7db of trim.

There are probably a dozen other ways to add volume to the mix. Simply adding volume doest change anything, so don't be afraid.
 
Farview said:
If you aren't using compression on the tracks, you can simply open the session and turn the trim up 7db on all the tracks. Then hit mix.

If you use compression, you would need to also turn the threshold up 7db to compensate.

Or you could take the mix you have, import it into a session, and add 7db of gain.

On the master buss, you can add 7db of trim.

There are probably a dozen other ways to add volume to the mix. Simply adding volume doest change anything, so don't be afraid.
Click to expand...

Thanks. Appreciate the help a lot. I've been supplementing this thread with youtube videos and articles to try to get up to speed quickly.

One final question for now: when exporting a wav from the day, is it best to dither it there or export it as-is and dither it during the mastering phase? Since it has so many effects and plugins on it, I would think dithering it upon export is wise, but then I also read that the more you dither the more artifacts can theoretically be created. The same question would apply to limiting -- is it best do to that during mastering or upon the export of the wav file from the daw?
 
When you export the mix, you should do it at 24 bit, no dither. When you 'master' it down to 16 bit, that's when you dither.

You only dither when you go to a lower bit depth, and you should only do it once.
 
ashcat_lt said:
If you go back to the mix, figure out why the right side peaks 3db higher than the left, and fix that[/], then you can get up to about 16db, which is still a bit quiet.
Click to expand...


I've been working on this...it seems it's either stray drum hits on panned overheads, or the backing vocal track being panned. But I'm still not sure. Does the peak difference matter much if it's an intermittent sound?
 
If you want the most volume with the least limiting on the master, then yes. I don't have a problem with mix bus compression/limiting, so I'd probably just let that sort it out, but if you're trying to stay natural, you should probably figure it out.

That said, usually the highest peak in a full mix is actually some place where several tracks all accidentally have their wave forms pushing in the same direction at the same time. That means that if you want to fix it at the track level, you might have to make subtle changes to several tracks. That's probably better anyway compared to making one big change to one track.

Just a kind of wake and bake try it and see kinda thought, but a high pass filter on the master might help. Set it to like 30Hz and see what happens. There are all kinds of ways that subsonics can build up, especially in a DAW, since they're essentially DC coupled. Phase rotation might also help if it really is just a matter of everything hitting at once. You'll get some of that from the HP, but a double pole all-pass somewhere around 100Hz might get your crest factor down without having to touch the dynamics. Yes, it will "smear the transients". That's kinda the point. Is that worse than distorting them? You'd have to try it and see.
 
4tracker said:
I've been working on this...it seems it's either stray drum hits on panned overheads, or the backing vocal track being panned. But I'm still not sure. Does the peak difference matter much if it's an intermittent sound?
Click to expand...
No the peak difference doesn't matter if it is intermittent. But, since the intermittent peak is 3-4db above the rest of the peaks in the song, if you take care of that peak you can turn the song up another 3-4 dB without using a limiter.

It is much easier to just put a limiter on the master buss and just set it to catch those stray peaks. Technically, you are affecting the dynamic range, but not any more than if you went and manually turned down those peaks.

People talk about the evils of changing the dynamic range, but sometimes it's a good thing. For example, if you have a song with a snare hit that is 12db louder than anything else in the song, it would make the song 12dbfs quieter than it could be. Would you fix the snare hit to turn the level up, or wring your hands about changing the natural dynamics of the performance? (However inappropriate it is)
 
Once you have the highest peak up around 0dbfs, the only way to get the mix louder is to lower the peaks and turn up the mix. If you are just lowering random peaks here and there, you won't hear it in the mix and it will not affect the perceived dynamic range. This is because our ears really don't perceive quick transients as 'volume'. Also, if these are just random peaks, their dynamics are inappropriate for the mix anyway.

When people bitch about the dynamic range being squashed, they are mostly referring to the song getting squashed to the point that the quiet parts of the song are just as loud as the loud parts. The songs dynamics are what is important, not necessarily preserving every little variation on every single track.

At the point that you decide a mix is too quiet, even though it peaks at 0dbfs, you have decided that the dynamic range of the song is too big. So you must make it smaller to achieve the loudness you are looking for.
 
If I know a mix is destined to be mastered loud I'm not above limiting individual tracks, usually drums. Let the kick, snare and toms limiters take 3-6dB off those drums at the track level. That way the mastering limiter isn't having to do that reduction on the whole mix where it affects everything.

I'll export a mix and open it in an editor to find the peaks and measure the RMS level. If something is sticking out I can go back to the mix and edit or limit the offender. As mentioned above, sometimes it's an interaction between several tracks.
 
Farview said:
When people bitch about the dynamic range being squashed, they are mostly referring to the song getting squashed to the point that the quiet parts of the song are just as loud as the loud parts.
Click to expand...

Hm, yeah, I am not meaning to bitch about it, but just state a preference for open, dynamic sounds and trying to figure out the best way to get there. At least the best I can with this particular material.

So in this case if it is a stray peak, basically that intermittent peak will cause clipping if I don't find it and try to push the volume, is that accurate? If I find it, I can then turn up the song 3db more without clipping and without a limiter. Is that a good summary?

Would this be an option: turn the gain up 7db since it was ~ -7db on that side, then use a limiter at that point? That might be easier than finding the stray peaks (I don't know how to do this other than trial and error).

The mix sounds much better than the original, which was done by a "pro"...big shout out to the forum for helping.
Today I bought a couple books on mixing and mastering so hopefully I won't have to ask so many questions once they arrive and I get through them.
 
ashcat_lt said:
That said, usually the highest peak in a full mix is actually some place where several tracks all accidentally have their wave forms pushing in the same direction at the same time. That means that if you want to fix it at the track level, you might have to make subtle changes to several tracks. That's probably better anyway compared to making one big change to one track.
Click to expand...

Is there an analyzer that finds the highest peaks and highlights them? That would be awesome.
 
4tracker said:
Hm, yeah, I am not meaning to bitch about it, but just state a preference for open, dynamic sounds and trying to figure out the best way to get there. At least the best I can with this particular material.

So in this case if it is a stray peak, basically that intermittent peak will cause clipping if I don't find it and try to push the volume, is that accurate? If I find it, I can then turn up the song 3db more without clipping and without a limiter. Is that a good summary?

Would this be an option: turn the gain up 7db since it was ~ -7db on that side, then use a limiter at that point? That might be easier than finding the stray peaks (I don't know how to do this other than trial and error).

The mix sounds much better than the original, which was done by a "pro"...big shout out to the forum for helping.
Today I bought a couple books on mixing and mastering so hopefully I won't have to ask so many questions once they arrive and I get through them.
Click to expand...
Yes, you can simply turn up the gain by 7db with a limiter.

If you end up using a mastering limiter, you can use that to turn up the gain. It won't do anything g to the signal until you've turned it up past the point that the peaks would start to clip. (In your case, 7db)

If you want, you can import the mix I to a daw session and edit the mix there. Since the peak is so much older than the rest of the track, you should be able to see it. Then simply select it, and turn it down.
 
4tracker said:
Is there an analyzer that finds the highest peaks and highlights them? That would be awesome.
Click to expand...
What DAW are you in? I know there are a couple of user-built extensions/plugins for Reaper that do something similar to that, but (i think) only on individual tracks, so you might have to do kinda like Farview said and bounce the mix first. If we're still shooting for fixing it at the tracks first, maybe bring that mix file into the original project alongside the tracks. Do your analysis and when you find the peak in the mix, go look at the individual tracks.

I tend to agree that it's not really worth all that for one or two accidental peaks. Just clip em off and be done with it. ;)
 
Why are you so concerned with what levels look like? Your ears are so much better at detecting loudness. You seem very wary of trying things out and experimenting. For instance, if you go above full scale, the waveform flat tops. Sometimes your ears scream at you sometimes you get away with it, buried in the mix, so your final project still works. I found one of my old projects with a few thousand CDs produced had one detected spike when I lost the master and had to redo it, and the software pointed it out? Never noticed five years ago!
 
rob aylestone said:
Why are you so concerned with what levels look like?
Click to expand...

Because i don't really know what I'm doing, so I'm relying on meters for now. I bought some books to help with mixing and mastering concepts. I don't trust my ears. I have good hearing, but I just haven't done this long enough on any serious level to hear subtleties.
 
rob aylestone said:
Why are you so concerned with what levels look like? Your ears are so much better at detecting loudness.
Click to expand...

Transient peaks excepted, which is what he wants to use a graphical display to find so he can address them in the mix so they don't mess with the mastering as much. Are you actually following the thread or just shooting out default responses to certain trigger words and phrases in the most recent post?
 
You must log in or register to reply here.
Back
Top