Sounding good at the right levels

  • Thread starter Thread starter sjoko2
  • Start date Start date
sjoko2

sjoko2

New member
In recent threads I have often encountered questions and opinions on things like mastering, getting “radio level” (whatever that might mean) etc. As a result I thought I’d write a post for “homers” on, well, mainly on the basics of getting something to sound good at the right volume levels

In order to come to an objective point of view, I have over the last couple of days spend some time listening to MP3 files, here, on MP3 and on record company websites, in order to listen if there were area’s of common “mistakes”. I also converted some of my tracks to MP3 files, something I had never done before. Three tracks, a rough mix / demo, a mixed track, and a finished, mastered track, so as to get a good comparison.

My observations were that there is an enormous difference in quality and volume on MP3 files presented, as well as some very noisy tracks. Quality is to be expected, but there is no excuse for excessively noisy or clipping tracks.

I will try to address this in two sections. The first being some tips to get your mix “up-to-the-level”, and the second a brief overview on how I would master a track. As far as the latter is concerned, please note that there are many ways; each engineer has their own way of doing things and their own favorite tools to use. Please don’t come back to me and say; “Alright for you to say, I haven’t got those tools”, that’s not why I write it; it is to give you a better idea. Most important – if you get the first section right, you’re 90% there.

Tracking

I’ll use drums as an example. The most common mistake “homers” seem to make is getting a good balance on their monitors, and then record that to tape / disk. Wrong! Your balance should be for your monitors only. Look at each individual channel as a separate project. Start with your kick, and make it sound right. Take your time! If you record at home, it does not matter, you don’t pay studio time! If it takes you a week of experimenting with microphone placements, tuning, pillows etc., who cares? Don’t stop until you are happy with the results.

When you have the sound you want on every single channel, take time to evaluate the complete kit’s sound, and make adjustments to ensure all tracks together sound as a complete instrument, without for instance one tom resonating in the exact frequency of another. Make adjustments accordingly, until you are happy with the result.

Time to record. Record each channel to the maximum level. Don’t clip out on it, don’t overdo it, make sure you leave room for dynamics, but aim for a level which hovers just below 0dB.

Do that with all your tracks, and you will have the basics right to start a mix.



Metering

How can you tell you are at the right levels when you only have tiny little meters? Metering is very important. Unfortunately only high-end equipment seems to have accurate, large metering. If it had large meters, it would be a lot more expensive.
But, there are cheap solutions. If you work on an analogue board, go to an army surplus place, or even radio shack, buy some meters cheap, and put them on the left and right bus of your console. A little bit of wiring, a couple of bucks, and you’ll be forever grateful you did it. If you work on a DAW, get some metering software. There are cheap ones and expensive ones; everything is better than none at all. (I use Spectrafoo software).

Mixing

Having finished tracking, now you come to the point where you will start to “set-out” your mix. In an ideal scenario, which never ever happens, you would have all your faders, including your masters, at 0dB. Your objective is to turn out a mix with 0dB levels. Obviously you cannot do that, but always keep in mind that it is the ideal world objective.

What happens if you loose track of all the above?
First of all you have to realize that your equipment is designed for optimum performance up to 0dB. If you tracked something at a low level and you have to push that track way up in the mix, you will introduce noise with it. You might be able to live with it on one track (not acceptable as far as I’m concerned), but 2, 3 or more tracks? A LOT of noise.
Second thing you could do – and many people do – start mixing at a low level on your faders. Then you are happy with your mix, but your overall output levels are well below the “ideal” – near a 0dB level. So, easy solution, you did have your masters on 0dB (if you didn’t, that is where they should have been), and you get the levels up by pushing your masters up. Same results – you introduce a lot of noise, you are exceeding the optimum levels for which your equipment has been designed.

If you have done the basics described above right, you will end up with a good volume, full sounding piece of work. If you ten want to print it to a CD, or convert it to an MP3 file, the same principles go. Get it as close to 0dB as possible, but don’t push it, leave room for dynamics.

Mastering

Some basics, like I said, just from my perspective. Once again, if you get the stuff above right, you’re almost there. I assume that you have mixed your track(s), so now you have a stereo track / file. You feel, after having listened to your mix on many different systems, it needs a bit of a “tweak” here and there. You won’t send it to be mastered, and you don’t have a lot of tools you can do mastering with. So, what to do? I can’t tell you. because I don’t know what type of equipment you have got, and I don’t know what kind of adjustments you want to make. What I can do is tell you how I would go about it with the “right” tools, and perhaps you have a piece of equipment available with which you can at least have a go at something.

1). Like I said, listen to your mix critically, on different systems. Ask people whose opinion you value to listen as well. Write down their comments. When you listen yourself, have a pad and pen ready, write your own feelings / comments down. Compare it to stuff you like, write comments down. Do you see a common thread? Now you’ll know what to do. First thing I do: Listen and make notes.

2). I use a spectrum analyzer. This allows me to see the frequency range of the track. It has “peak hold” facilities, so I can see where a track peaks out of proportion. Cool tool! I use it everywhere, especially for mixing and mastering, also for live gigs, which unfortunately I don’t do a lot anymore.

3). EQ – If for any reason you feel the track lacks something in one frequency, or needs something in another, You’d carefully apply some EQ. I use TC tools for that, which allow me to adjust frequencies in an almost surgical manner. If you don’t have anything like that, put it through 2 channels on a console, and see if you can make an improvement. Keep both, play it on your grannies boom box, in your car etc. See if you are happy with it.

4). Peaks. Something peaks wildly? First of all find out what it is. Is it a certain frequency that pops out? Can you fix it with an EQ? If not, apply some limiting, carefully, not too much. If you have not got access to a limiter, you might have to lower the overall volume of your track …. just enough. If you have access to a compressor, you might be able to squeeze it just a little, again, just enough.

5) Low level. If there is a low at a certain place in a mix, you could use an expander. If you haven’t got that, you’ll have to try and ride faders. If you work on a DAW you’d be able to do it very precisely, if not, hands and practice will do the job. Make sure you don’t overdo it and loose the feel, or dynamics, of the mix.

For 4 and 5 I also use TC, TCMasterX, which has the following features:
3 band expander / compressor / limiter
+12 to –25 dB gain control per band
Adjustable band crossovers
High precision ppm meters with consecutive clippings counter
Dynamics processing curve display per band
Target control per band, for module specific processing adjustment (hi/lo)
Look ahead delay of up to 10 milliseconds
Digital ceiling with an accuracy of up to 0.01dB
Separate L/R input level controls for post-balancing mixes
Soft clipping with analogue emulation for limiter stage


Most important, as you can see from 3, 4 and 5 above; all you are doing is fixing things which should have been addressed in the tracking and mixing stages. Once again, get your tracking and mixing right, which takes time, patience and practice. That way you can avoid most of the things which a mastering engineer has to do, namely ensuring the track is up to a good level, sounding good and faultless.

I hope the above will help someone, have fun!

PS I didn’t address things like compression in detail, Shailat has done an excellent piece on that, I suggest you read it.
 
Christ! And I thought my "remote drum recording in 2 days" post was wordy!!!

:D :D

Great post SJ2!!! I bet it helps a lot of people out quite a bit!

Bruce :)
 
it helps ME out quite a bit.

Do you eq your speakers or do you leave them flat as they came from the factory?

A recording book said to eq speakers with a vocal recording, if one does not have access to a frequency analyzer.
 
one more thing sjoko2,



Your reaction to noise is a new point of view for me, and I am sure there are many people in the world who cannot stand that noise.

I must confess that before now, I have been guilty of having excess noise in my recordings and other times allowing some tracks to clip, in the quest for the hottest signal possible.

I can totally see how that will let the cat out of the bag about a home recording.

thanks for that information
 
Good advise Sjoko.

Metering is something That I have struggled with. Sometimes I feel like I can't trust the meters in my software. In Cubase it seems I can't get the sound right unless at least a few of the tracks are clipping. BUt are they really clipping or is that red light just coming on? I think most software lies about their levels, mostly I think there is more headroom above that litte red light. I've used lot's of software and I find that metering is one of the biggest differences to them. Like you have to learn the meters, and disregard them sometimes. I use the meter's on my mixer as another reference, I set the levels at unity. But wait... can I be sure that the D/A converters on my card are putting out levels at unity? IT DRIVES ME NUTS!!!!!

For a lot of us "homers" (I really like that phrase sjoko ;)) It bolis down to guess work and experimentation, but we have time to do that don't we?

It's funny with your drum example... Last time I tracked some drums, I did exactly as you said was the wrong thing to do. I got a good balance on the kit, it sounded real nice, I'm happy with it. Then later I was like "hey, these levels aren't very loud at all" -really though it was the best drum sound I've gotten, but that's cuz I nailed the mic placement for once, except for the ride, had that too upfront, wasen't cutting so I kept changing it and changing it, till finally it cut, but in the mix a little too much... it was a submix situation so I can't do anything about it now!

So we know that increasing a fader above 0db introduces noise... What about attenuation, aren't we losing some depth there? (especially in the digital realm) This is where I have to disagree to some extent about getting all the levels hot during tracking. (man oh man it's funny that little old me seems to often have a counterpoint for YOU! :D) If there is something that you know you'll want "back" in the mix. Don't record it so damn hot! I've done this lots of times where I get some screaming distorted guitar part, track it close to 0db, then have to suck the life out of it by turning the meters waayyyyy down. This is an exception though. Something like distorted guitar isn't going to be very dynamic, it's by nature a very compressed signal, so recording it hot as hell actually gives you less room to work with in the end sometimes.

-jhe
 
sjoko-
thanks very much for your post. I am right in the middle of tracking, and I'm going to go back and redo some stuff with this help. I think this will be my best track ever!

H2H
 
CyanJaguar, I think the book you read, correct me if I’m wrong, is either quite old, or they are talking about live sound reinforcement. Modern monitors are made to be used “as is” especially bi-amped ones. Placement is important; as is the acoustic environment you are listening in. I am not saying you have to build a whole studio, use your ears and common sense; you can do a lot with for instance blankets draped over boom stands behind your monitors or in front of reflective surfaces.
EQ’ing monitors would be the last thing I’d like to get involved in, that’s another job for another bunch of specialist pro’s.
Even when I do live sound with an existing system, first thing I do is take out any graphic EQ’s and things like that. Why? Example – lets say they have taken out a certain frequency, but I want to boost that frequency on a channel. I change the channel eq, nothing audible happens, so I have to boost it fully to get the response I want – introducing noise again.

James, I’m sure you’d be able to get a VST metering plug-in. There are loads of VST sites, this is just one of them http://driene.student.utwente.nl/e.m.szwajcer/crm/fx/plugs.html
The one thing is, if you work in Cubase, you’ll be able to get hold of plug ins for everything I describe in my post, for free of very little money.
Next part of your question; an essential part of every engineer’s “toolbox” should be a reference disk, one you can use to calibrate your system. With such a disk you can run tones, use the tones to see your signal flow, set is at 0dB (without your monitors on!) and see if all meters in the path are at odB. If not, make the necessary adjustments.

The final part – and it is VERY IMPORTANT to understand this clearly. When I said always track to a maximum level, I mean ALWAYS. It is (or it damn well should be) a golden rule.
Let me give you an extreme example, and then you just think about it:

Say you have a signal path. It starts with a microphone, and a microphone makes its own noise. Cable, more noise, pre amp, more, your console, more. You measure this noise from beginning to tape / disk, and say you have 1,5dB of noise in total in your signal path. Now you grab your fader, which is all the way down. You open her up for 1,5dB, and what will you hear? ½ noise, ½ signal, a ratio of 1:1, nothing you could use. Now you double your fader level up to 3 dB, your signal will become clearer, but ½ of it will still be noise. Increase the level another 15 dB, and your noise is virtually gone compared to your signal. And now for the key problem. Say you record the track in question at its 3dB level, its noisy as hell, but hey! It was a nice balance when you listen to “the whole”.
Now you come to mix the track, and you find that signal needs a little boost in the mix. What do you boost? The signal, yes …. but with it a load of unwanted noise.

For the same reason, and I think this is also something very important. If you use a pre amp to go to tape, or a pre amp converter to go to disk / digital tape, why on earth would you put that through a console as well? By all means use the console, but just to listen to your returns. If you get into the habit of tracking direct, you will end up with infinitely better, cleaner sounding tracks.
 
thank for the link sjoko.

about using my mixer to double check meters, I'm speaking only at the mixing stage, to see where my master fader is. sorry if that got confused.

I agree that as a golden rule, you should track to a maximum level. BUt if I can record a part and not have to move the fader on mix down (I'm talking software faders) shouldn't I theoretically be getting the best performance out of it being that it was designed for optimum performance at that level? The less "math" I have to do on the signal, the "truer" it will be. If I have to pull that fader down over 6db I'm losing something am I not? In the analog realm I agree with you 1000 percent. And I'm not disagreing with you, I'm just saying that in some cases, breaking the rule can be a good thing. I could be wrong though. It's just that I trust those little virtual faders less than those friggin virtual meters! At 24 bit 48K or 96K do we really need to be so nitpicky over a few db? Can the softwares virtual master buss handle 24 traks of audio coming in, all just under 0db, without hitting the cieling and sacrificing headroom? Please, tell me where I'm wrong. I try not to be so stubborn! :D

-jhe

-jhe
 
James, the noise is in your signal path, not in your virtual mixer. Does that make it clearer?
 
I came back here cuz i wanted to add to my post and I see you've already responded. Thank you for this, this dailog is very good! Yes the noise is in my signal path. What I don't get (sorry I have to be stubborn here) is exactly that. Why turn the level of that signal path up (ie, a nosy pre-amp, noisy mic, noisy cable) just to get that signal up to maximum level to my A/D converters, just so that I can turn them down agian with my virtual mixer?

I used a VERY LOUD guitar example before, now I'll use a quiet example. A triangle hit. just one. It's not going to be exceptionally loud in the song. I could, wanting to get the loudest signal, crank up the gain of my pre-amp, introduce all kinds of noise from my crappy starved-plate tube pre amp and my poorly designed chineese capsule Large diaphragm mic, and hit that disk hard! Seems like overkill with such gear. (which is what a lot of us have) yes I understand that even at quieter levels, the noise from my signal chain will be amplified by any increases in gain. noise is noise, yes, and it not going to get any better no matter what or where at what stage i increse or decrese the gain. I'll not push the weaker link.

The only reason I bring any of this up, is that I think saying "track at maximum levels" is too easily misinterpreted by stupid "homers" like myself. So now we are obsessed with making eveything as loud as possible, and then we'll freak out. "OH no! the snare drum peaks a bit, so I better compress and limit it so that I can get my average levels up- otherwise i'll have all kinds of noise in my signal!!" That defintly isn't what you are saying, but, throw a qualifyer out there like "maximum" and some people wont know what to do with it. I was going to say that it would be better to say "track at appropriate levels" but then, what's appropriate? I'm glad I didn't say that. ;)

hmm.. what's gotten into me? :D nit-picky I be tonite!

(once again, James talkes himself into a corner, over explains himself, and really proves nothing. :D)

-jhe
 
james? I'm talking about your signal to tape? The input to your recording system, yes????
 
I read the Dithering article at www.Digido.com , it basically states that any change to a digital signal, eq, gain changes, compression etc results in an increased digital 'wordlength'.

When mixing we are obviously going to need to manipulate the signal and therefore increase the number of bits used to represent it.

How much digital headroom should we leave to avoid bit truncation during mixing?

Am I even understanding this correctly? :(

Thanks.
 
wordlength

Well, its doubtful that you could really measure digital overhead as far as wordlength goes. Incorporating a packet sniffer somewhere inline might work, but I dunno how feasible that would be. A bitmap maybe.
In 24 bit digital, all 0's is the lowest number, all 1's being the highest. Anything over 24 consecutive ones would be considered an over, and i guess distortion or clipping of some type would occur.
I'd like to know how moving faders actually affects the word lenght, and if wordlenght somehow "returns to normal" if you undo something, of if its just made that much bigger.
I come from a computer networking background, but audio, supposedly, operates a little differently than what I'm accustomed to.
 
Tubedude, I am trying to keep this as non-technical as possible, as it really does not have to go into those things.
Doesn't matter if you are a total novice or a pro, neither do you need to know anything about the technical detail of digital audio. I know quite a few top engineers who haven't got a clue how a computer works, but they can read and write on Pro Tools. What you do need to know is;
"I have to get the levels right, to much and I clip out, and screw up, to low and I'm introducing noise to my recording".
 
Sjoko2,

This is the most beautiful, best fuck of my life, kick-ass batch of info
I have every seen at this web-site.

God bless you (if there is one) for spreading the word on how to
achieve a better sound for one's songs & taking the time to point
out what a lot of us are doing wrong at home.


You taught me something today. Thank you.


Love,
SEAN
 
Thanks smelly, thats what I'm here for, so I'm just glad its not in vein
 
Sjoko2:

Great stuff. You appear to have the knowledge AND the ability to explain it to those of us aren't rocket scientists.

Here's one for you:

Why is it that I can't feed anything into my line in (sound card) and get a hot enough signal to even approach the "reds" (just under 0 db). For example, when using my DR-660 drum machine, I have to run it through my tascam's mixer to get sufficient signal boost. What you said seemed to confirm my suspicion that running the DR-660 through a mixer would create noise. Preferably, I'd like to bypass the mixer. I have the same problem with the bass guitar, and to some degree with the electric guitars (in that case, boosting the POD output is sufficient). The mic in is fine, and I've checked the line in input on the soundcard (actually it's an onboard soundchipset - C-Media).

Any insight to offer (anyone)?

Thanks
 
Well ps'66 - first things first. I know NOTHING about soundcards, so I cannot make any comments on them, I have never measured their input sensitivity etc.

Assuming your soundcard is configured for line level input (if there is anyone here who knows about such things.... HELP!) everything you will try and run into the card direct has to be at line level. You won't be able to plug instruments direct into it, without going through a pre amp stage to boost the signal level. In case of your guitar / POD combo, the pod will function as a pre amp an boost the signal. By the way, ever tried the bass pod? Its really cool!

One way to overcome this problem, go direct and bi-pass the mixer, would be to get some good direct boxes.
Another option would be to adjust the input level of your soundcard, but then, I have no clue if you can do that or not.

Sorry I can't be of more help.
 
I just finishing reading an article, prior to reading your response (thanks). It emphasized exactly what you said (line input only accepts line level). Your suggestion regarding a direct box is quite valid I believe (maybe I'll even try running the drum machine through the "tube preamp" setting of the POD and send it directly to the computer - might be "interesting"). Also, it's probably a good time to buy a "real" mixer as opposed to using the Tascam. Seems as though many users dislike Behringer mixers. What's the best mixer bang for $200?

Thanks for the reply (that damn learning curve........).
 
Back
Top