Don't try to get your final volume until the mix is done.

Headroom is the space between the loudest peak in the signal and 0dBFS. During tracking that should be around 12dB. You should set your levels during tracking so your peaks are not over -12dBFS. The average levels of the tracks will be somewhat lower, especially for very dynamic instruments like drums.

During mixing you also need to leave yourself some headroom. When you put several tracks together the signal level rises. Theoretically you get about 3dB of gain for every doubling of the number of tracks. So remember that turning tracks down in the mix is always an option.

Although you won't hear the clipping while in the project, once you export it the clipping could become very audible.

After you have a good mix it's time to think about volume. Normally this is handled in the mastering stage. Export the mix to a file at the bit depth and sampling frequency of the project and import it to a new project. There you can apply a mastering limiter or other processor to get the final volume you want.

Moonfury, post: 375847 wrote: Hey, folks!

I guess compressing/limiting is like the most discussed topic on this forum. I'm not a total newbie, I won't ask what compressing/limiting/gating is, instead, I want to know why I fail to record really loud tracks. You can check my gear in the profile, I tried to fully decribe it. So I got several questions for you:

1) First of all, I read people talking about headroom a lot. I'm not a native speaker, so this thing is not 100% clear to me. What I think it is - it's the "free room" in dB left after mastering from my track's peak level to 0 dB and it is used to, say, establish a peak level of the track of, say, -5 dB and just to add like 5 or a bit less dB by adding gain to the whole track. Is it so? Or do I have false assumptions about headroom?

In digital recording "headroom" is the distance expressed on the level meters from wherever your recorded signal resides to 0dBFS, which is the point at which the digital signal can go no louder and will distort (known as "clipping"). It can also be described in the analog domain as the distance from whatever point an amplified signal is residing at to the point an amplifier can no longer amplify the signal any further. So for example if you record a track and the meters in your sequencer display that the incoming level of that track is at it's maximum "-10dBFS" then you have "10dB of headroom" before the signal clips digitally. The digital meter scale is marked in levels below "full scale" (0dBFS). Therefore everything has a "-" (negative) sign before the number (except of course zero).

...2) Secondly, I get strange results when mixing. Say, I got two tracks in my composition. When it's only one track playing and the other is muted, I get no clipping. But when they are played alltogether, the meter starts peaking and such horrible sound is produced. I guess that it happens because when multiple tracks are played together, the sound on the same frequencies gets boosted and the level rises beyond the maximum. To avoid it I try to use compressors/limiters, but they are generally not much help, because when I have like 10 tracks in the same song, it's pretty hard to find out which one causes clipping, because when they are played in solo mode, neither clips.. So that's the question: when I have multiple tracks which do not clip if soloed, but which do clip when played together, how do I identify what settings on what track to tweak to get rid of this clipping?

The answer is that you are tracking way too hot.

Your system and all of it's components (preamplifiers, converters, etc) were designed to function optimally at a certain calibrated point that is commonly known as "line level". The term "line level" is relative because the actual point of calibration varies from gear to gear. But in general with a modern system you can expect that line level is roughly an incoming signal into your computer sequencer program that measures at -18dBFS (or on an analog VU meter 0dBVU). That means that if you record with your signals averaging an incoming level into your sequencer of -18dBFS you will have an average of a full 18 decibels of headroom before clipping occurs. Keeping in mind that multiple tracks will increase the overall incoming signal level and you can see how you'll need that headroom to prevent clipping. This is important because clipping and indeed in many cases even coming very close to clipping is one of the main reasons why certain people can't get their recordings to sound good.

When you record too hot (with your signal levels too high) what you do is overdrive the analog and ultimately the digital stages of your equipment. That overdrive sounds terrible. It manifests itself as colder, more congested, more distorted, less realistic, less dimensional sounding tracks. So it's very important to leave as much headroom while tracking AND mixing as you can to prevent even getting close to clipping. Keeping in mind that music varies in it's level and intensity you can imagine that it's therefore important to leave more than enough headroom so that when you get occasional, momentary peaks in level they also don't clip your gear.

When recording it's not at all unusual for someone experience and skilled to leave as much as 20 to 24dB of headroom, depending on the dynamic range of the source being recorded. That would mean tracking with an average incoming signal level of -20dBFS to as low as -24dBFS. That might seem quiet to some. But notice that most people these days are recording at 24bit digital, which is capable of providing more than enough signal level above the noise floor of the system. So there is no need to track as hot as you can as was formerly the case in the days of analog tape being the storage medium. Those who still think that you need to track as hot as you can are either completely clueless or they are applying old school analog tape thinking to the digital realm and the two are not the same medium.

...3) Thirdly, I'd like to know if multiband compressors are a better choice over simple compressors. If so, then I need a tool to analyze the frequencies of the track in each moment of time to find out what band to compress.

Multiband compressors are a specialist tool that make more sense to be used very sparingly and usually only during the mastering stage of audio production. You being as obviously inexperienced as you are have no good reason to entertain the idea of using a multiband compressor. They are very powerful sound shaping tools that are easy to use accidentally and screw up the sound of a track. I am an experienced, professional recording and mix engineer and I have been working for over 10 years and I have never used a multiband compressor for anything at all. There are those that use them. Most don't know what they're doing. There are a few that do know what they're doing. But in general it's wise to avoid them.

...4) Fourthly, is there like a simple way to make a loud song if you got a lot of tracks in it? :)

Yes there is a simple way. But in the words of Bob Katz: "It's not how loud you make it. It's how you make it loud." If you don't yet know enough about preserving headroom during tracking and mixing then you will not be able to make something sound loud and still sound good. So at this point there is no reason for you to be concerned with how loud a track is. Focus on it sounding good first and when you can consistently do that while preserving headroom throughout the tracking and mixing you'll be in a much better position to consider loudness. Any idiot can slap on a limiter and turn it up. But if you don't know what you're doing you'll just screw up the sound by squashing the dynamics of the track.

5) Fifthly, there is such a thing in Cubase as Audio > Spectrum Analyzer. What does it show? It shows like average level of each frequency for the whole audio file?

Something like that. Yes I think Cubase has one. As far as how useful it is for tracking and mixing goes, I think it's not really all that useful. You mix with your ears, not your eyes. Seeing the frequency spectrum is like dancing about painting. They are not the same thing. Until you know what something is supposed to sound like, seeing it displayed over various metering tools will do you no practical good.

...6) Sixthly, when recording vocals I get an unwanted hiss due to amplifying the signal. I try to gate it out or to manually add silence between phrases, but it turns out that it is silent when it should be, but when the person speaks you still can hear the hiss, because it appears after total silence with the voice and you easily notice the contrast between total silence with no hiss and voice with hiss-over. Can I do something to get rid of it? De-noising with Cubase's de-noiser heavily affects the quality of the vocal tracks :( Or another question: Is there a thing to gate something out with fade-in/fade-out settings, so I can make a gradual increase in the volume after silence and so that this way the hiss is not highlighted after the total silence?

Again: YOU ARE TRACKING ENTIRELY TOO HOT. Unless you have a broken (or very shoddily designed) piece of gear or you're trying to use a dynamic mic to amplify a mouse fart from 10 meters away there is no reason for you to be hearing the kind of hiss you're describing due to preamp gain. That's a sign that you're tracking too hot. Turn down the preamps until the incoming signal averages around -18dBFS and then turn up your monitors so you can hear things. No gate or other piece of gear can mitigate poor tracking skill. You shouldn't need a gate at all, unless you're in some kind of very loud environment that is constantly affecting your tracks.

7) Lastly, Cubase sometimes shows red marker in the output channels, which refer to clipping. But I can't hear it. Is it a bad thing to leave it with these clipping markers, although I can't really hear clipping? Or should I try to get rid of any indication of clipping, even if it's hardly audible?

I'll say it again: YOU'RE TRACKING TOO HOT. There should be no clipping, ever. That should never become a problem ever. If there is clipping then it's a problem. That you "can't hear it" says that you don't know what to listen for.

You seem VERY inexperienced and unschooled at this stuff, which is fine. But you need to get this concept right now or you'll NEVER succeed at recording good sounding tracks.

-Stop tracking too hot.
-The answer to poor sounding tracks is NEVER another signal processor or plug-in
-Multiband compressors are not for your use right now (probably not for the next few years)
-Learn about the INCREDIBLE importance of preserving headroom. HEADROOM IS GOOD ALL OF THE TIME.

Man.. You are my god now.. Really, put so much effort to explain all the stuff to such a newbie.. I really appreciate it. Long live your family. Really.

:)

Say I got a nice mix of ten tracks with -18 dBFS headroom. What do I do after? I mean, I tried to do stuff this way, but if I add gain on the output channel after the mixing is done I still get really quiet tracks compared to professional audio. Yes, I get that their equipment is like 10-100 times more expensive than mine one, but still.. And another question - what is an easy way to monitor level of the signal in dBFSes? Since Cubase's meters are all numberless, I have to manually add an insert to the track, like iZotope Alloy, which has numbered scale, to monitor the signal level..

Cubase meters have scales on them just click on the full scale "up" arrow icon on the top left side of the mixer and select the meter icon on the left hand bar! The control room mixers also have full scale meter with level indicators. Just enable Control Room.

Moonfury, post: 375900 wrote: Say I got a nice mix of ten tracks with -18 dBFS headroom. What do I do after? I mean, I tried to do stuff this way, but if I add gain on the output channel after the mixing is done I still get really quiet tracks compared to professional audio.

So long as you have been tracking so that your incoming signals are averaging with enough headroom you can mix and aim for keeping your output level at the master bus averaging as high as -10dBFS or even higher at say -6dBFS if you so desire (I personally mix so that the overall average output signal level is around anywhere between -15 and -10dBFS. But occasionally I'll mix a little hotter and average around no higher than -8dBFS). The point with the whole "-18dBFS" thing is about average input signal levels while tracking, so you'll be making an estimation of that average based on where you see the meters reside most of the time as the incoming signal is playing while you set the gain at the preamp. That allows you to preserve your headroom for the next stage, which is mixing.

As you then go from tracking to mixing you will have an increase in overall output level with all tracks just simply playing back with their faders set to unity gain. That's normal and expected. All you need to worry about as you mix is that when you finally do finish mixing you are being certain that you are leaving enough overall headroom at the main output bus so there is room for the mastering engineer to do their thing without squashing your tracks. They don't need quite as much headroom for mastering as you need for mixing. So you can get away with letting the average mix output level be higher than your average input signal level while tracking.

Keep in mind that mixing and tracking are separate things. When you mix at some point you'll need to pull all of the recorded channel faders down and push them up just enough so that the overall average output level leaves you with enough headroom. That's not what you do when tracking. When tracking you leave the channel input faders at unity (I also generally leave the channel output faders at unity unless I'm doing a cue mix).

You're not even using the input faders as you mix. Also, you need to be aware that occasional transient peaks in output level are fine, so long as they don't clip. It's the average level that matters the most when regarding the question of headroom at the output bus. So long as the average level leaves enough headroom for dynamic shifts and transient peaks to not clip (even thinking as far forward as what they might do during the mastering stage) you'll be fine.

...Yes, I get that their equipment is like 10-100 times more expensive than mine one, but still..

For the most part perceived loudness has little to do with that.

...And another question - what is an easy way to monitor level of the signal in dBFSes? Since Cubase's meters are all numberless, I have to manually add an insert to the track, like iZotope Alloy, which has numbered scale, to monitor the signal level..

Cubase's meters are not numberless. The numeric display is at the bottom of the meter.

thanks, djmukilteo, thanks, Mirrormix!

Mirrormix, post: 375904 wrote: [...] you'll be making an estimation of that average based on where you see the meters reside most of the time as the incoming signal is playing [...]

or I could use the descriptive audio statistics from Cubase's Audio menu.

Professional songs have their average volume level floating just below 0 dB according to Cubase's meter. So mine one after mixing will be floating at around -10 dB. I didn't catch what happens after I mix the composition properly..

Mirrormix, post: 375904 wrote: [...] so there is room for the mastering engineer to do their thing [...]

you mean I have to hand it over to some third party to do the loudness thing for me after I completed mixing and preserved like 10 dB of headroom? That's kinda.. Sad :(

In Cubase you can insert the Maximizer/Limiter plugin on the Master bus. Put that in the last two plugin slots (yellow colored slot #17). This will get you to -3 to 0dbfs. Play with the settings or use some of the presets to get what you like. Then use that UV dithering plugin in #18 slot as your final plugin before conversion to MP3.
You can always add compressors or verb up in the top 1,2,3 slots if you like on the master bus as well.
Having that -10dbfs headroom is a good to have to give you some room to push the final mixdown...
Hope that hepls

Moonfury, post: 375900 wrote: Say I got a nice mix of ten tracks with -18 dBFS headroom. What do I do after? I mean, I tried to do stuff this way, but if I add gain on the output channel after the mixing is done I still get really quiet tracks compared to professional audio. Yes, I get that their equipment is like 10-100 times more expensive than mine one, but still..

Reread my first reply.

Moonfury, post: 375907 wrote:

Professional songs have their average volume level floating just below 0 dB.

I think redbook specifies -0.02 db. Someone please correct me if I'm wrong.

Moonfury, post: 375907 wrote:
you mean I have to hand it over to some third party to do the loudness thing for me after I completed mixing and preserved like 10 dB of headroom? That's kinda.. Sad :(

Some third party? Mastering engineers are your friend; most are highly respected by their peers and industry and there is nothing sad about what you described. Mastering is a different science/art than mixing. You should be thinking in terms of collaboration. What is there to be sad about?

Ripeart, post: 375911 wrote: I think redbook specifies -0.02 db. Someone please correct me if I'm wrong.

There could be a specification, but I've never heard of one. Early on it was unnecessary because you only have to worry about intersample peaks when there's been processing in the digital domain. If you simply convert a stereo signal to digital and don't process it there will be no intersample peaks. So I suspect there isn't a standard for headroom in Redbook.

I have heard two suggestions regarding headroom to prevent intersample peaks, -0.1dBFS and -0.3dBFS. I use -0.3dBFS just to be cautious.

So it turns out, that I should rely on digital amplification, moreover - the most basic one, included into Cubase? I mean, after I have a nice mix with all the volumes and say 10 dB of headroom, I have to digitally boost my composition?

Ripeart, post: 375911 wrote: [...] Mastering engineers are your friend [...] You should be thinking in terms of collaboration. What is there to be sad about?

I'm not saying it's bad or anything.. The sad thing is that, first of all, if handing my composition to a matering engineer is an essential thing to do, then I fail here, because I don't have such friends :) Secondly, it is sad, because I want to do it myself with descent quality, and it turns out that it's so complicated :) Thanks for your replies, Ripeart and bouldersound!

Moonfury, post: 375941 wrote: So it turns out, that I should rely on digital amplification, moreover - the most basic one, included into Cubase? I mean, after I have a nice mix with all the volumes and say 10 dB of headroom, I have to digitally boost my composition?

Yes, get the mix right with headroom to spare and export it to a high quality stereo audio file. Then bring that file into a new project and start messing with it. While you may be able to use a generic dynamics processor successfully, it may be worthwhile to find a mastering-specific limiter. The first reason is that it will probably sound better in that application. The second reason is that often they have automatic makeup gain. Automatic makeup gain is a terrible feature on a compressor for tracks in a mix, but for mastering it's quite handy.

Up to a point all you really have to do to get loud is push the signal into a limiter. But as you push the level the sound will start to degrade. That's when the the difference between "how loud you make it" and "how you make it loud" becomes painfully obvious.

While I am in basic agreement with the statements above, I don't think the ones about mastering are relevant to your situation. Remember, until the current "volume wars" there was a lot less obsession with mastering. Mastering engineers were well known in the business, but it was really inside baseball. "You need to send a disk out for mastering" is good advice for a commercial recording, but it is a waste of money on most music you will make as you are learning to record.

At this stage you just want to concentrate on getting a good mix with a lot of headroom. No, it will not sound as loud as commercial recordings - turn up the volume on your stereo. If you listen to music from the pre-digital era there are actually a lot of them that were mastered with a lot of headroom. (The original digital master of Abbey Road is about -4dBFS. Only a few kabillion people bought it.) As you get better at compression and limiting you can apply those to the final mix and raise the average volume a bit, and if you get better at it you squash it a bit more and call it mastering (if you don't mind the pros yelling at you) or finalizing (if you do). When you get better at mixing you can send it out to someone with a fresh set of ears (that have not heard the same song 1,000 times) and equipment designed for two tracks instead of two hundred.

To answer your point 3), I do not think that multiband is better than single band.
it really is different.
And is sounds different.

When in single band mode, you have to bear in mind that signals with a lot of energy will trigger the gain reduction. Like a kick a snare or a bass. If might feel like the so called "glue" in the mix. And might get pumping.
If the kick triggers the compression then the whole signal will be reduced.
You have to experiment. This might be a sound that you like.

when doing multiband... you can have the kick "stay" where it should be. The vocal "stay" where it should be. The snare as well etc. Because these instruments do not share the same frequencies. it is much more subtle. Much more complicated to use. But you can achieve something that is louder without sounding too compresses.

Hope that helps