can anyone explain how to get the highest possible potential volume?

[acorec]

I'd like to know what techniques you use on drums, guitars, guitar amps, bass guitars, and vocals as far as setting them up in the room. I'd also like to know how you set up the microphones and what levels you set them at when recording. I use an Aardvark Direct Pro 24/96 and it says when recording, the levels should "be in the yellow"... slightly above green and definitely not peaking/distorted. It's all relative, I know. Anyway, what do you do to get the highest potential volume? What do you do when mixing to get highest potential volume? Then what do you use to compress/limit the hell out of it?

I'd rather get a great recording without compress/limiting... and by great, I mean bad, because I mean loud.

The idea of getting the highest volume per track goes back to analog tape and 16 bit recording. Without getting too far into it, 24 bit recording allows much more volume data than 16 bit recording. You don't have to record "hot" at all. With analog tape,the idea is to push the noise floor down by recording at 0 or above without audable distortion. The 16 bit data scheme had to be recorded hot because of the available bits that set the volume of the data. The 24 it system has an additional 8 bits of data that specifically relates to the volume of the recorded content.

Now, in a mix, it is *relative* volume that you are concerned with. Most all recordings are vastly out of proportion if you think about it. When was the last time you could hear an acoustic guitar at the same volume as a drum kit or distorted guitar? Multitrack recording is by definition out of balance in the extreme. The loudness you might be concerned with is the final mix to a CD. That is the current craze and has more to do with mastering than anything else. What kills this aspect of mixing to CD at a pro volume is that you have transients that send you into the "red". In the digital world, going into the "red" is not like analog at all. The effect is serious distortion so bad that it will scare you. Mastering is just a way of being able to hear the content and adjusting out any transients in the mix. Then the level can be run to the top with no distortion. Recording each track at the high end of the volume scale is not a great method of achieving a good mix as you run the risk of killing the summing bus. The real solution is to get a balanced mix and mastering the result.

 

[acorec]

"...Lastly, if you want something to sound loud ... then don't try to play or sing "louder." That's the last thing you want to do. For example: the bassist who picks really hard during "loud" parts of the song, or the guitarist who strums extra hard, or the vocalist who tightens his vocal chords and the drummer who smashes the daylights out of his loud cymbals.

Instead, try playing more deliberately and with controlled intensity. Think of it like this: What do you really hear when a bassist plays "louder?" I know what I hear ... a bunch of "PINK!" "TING" "POP!" crap. Pure muddled crap, basically. Actually sounds a lot weaker than when he was playing 'quieter.'...."

I forgot this, but Chessrock hit the nail on the head. When I record the bass as an example, I put the volume way up and actually play *lighter*. The loudness and smoothness comes from the string and not the "attack". The best bass players play in a very controlled manner. I also use a limiter just in case I play a little too hard. You must pay attention to playing very lightly and it takes alot of practice. Guitar is the same way. Attacking the instument just introduces sounds that take away what you are trying to record in the first place. Smashing on these insruments actually does not allow the sound to "bloom".

 

[FALKEN]

from my experience, I would have to disagree about not worrying about recording levels hot. Any volume increase after the sound has been put on hard disk lowers your sonic quality by bringing lower bit distortion into the higher bits. at least that is what I hear. you will get a much better final product by recording your tracks as loud as possible without clipping and then lowering the volume to the appropriate level in mixing. I try to avoid any increase in volume digitally. Especially if in 16 bit. any digital volume increase will lower you to say, 14 bits, depending on how much increase.

I would have to agree on the way it is played and dynamics though.

Jimi used to use that trick. The first track on his album would be some quiet, noisy, ambient thing. then, track 2 would come in, and sound loud as shit, when actually, compared to say tracks 3 and 4, it is not that loud at all. it is all in the contrast.

But the original question is on final CD mix volume at home. Compress on the way in. Compress the vocals, the bass drum and snare drum (gently), compress anything panned center. leave the guitars and overheads un-compressed. compression will stop the transients, therefore allowing you to get a louder mix without clipping. compressing on the way in will allow you to do it without applying digital makeup gain.

if you disagree with any of this, that is fine. I am not an engineer and this is only what I have learned through my own experience, trying to acheive the same goal myself.

 

[Blue Bear Sound]

if you disagree with any of this, that is fine. I am not an engineer and this is only what I have learned through my own experience, trying to acheive the same goal myself.

I definitely disagree with most of it.... there are some valid point with respect to 16-bit digital, but with high-quality 24-bit converters, getting "hot" levels simply doesn't matter AT ALL. Period.

 

[FALKEN]

bruce.

how many bits would you calculate are lost by a 10db increase? or lets call it a 12 db increase to make the math easier.


just occured to me: every bit = 6db ?

 

[Blue Bear Sound]

The answer is ZERO bits are lost. The only thing you change is the relative proximity of the signal to the noise floor, and at 24-bits, there is A LOT of room before that occurs.

Before trying to argue the point further with me, I suggest you read Bob Katz' articles on his site......

http://www.digido.com/portal/pmodule_id=11/pmdmode=fullscreen/pageadder_page_id=18/

But these 2 in particular:
http://www.digido.com/portal/pmodule_id=11/pmdmode=fullscreen/pageadder_page_id=35/
http://www.digido.com/portal/pmodule_id=11/pmdmode=fullscreen/pageadder_page_id=27/

 

[acorec]

The answer is ZERO bits are lost. The only thing you change is the relative proximity of the signal to the noise floor, and at 24-bits, there is A LOT of room before that occurs.

Before trying to argue the point further with me, I suggest you read Bob Katz' articles on his site......

http://www.digido.com/portal/pmodule_id=11/pmdmode=fullscreen/pageadder_page_id=18/

But these 2 in particular:
http://www.digido.com/portal/pmodule_id=11/pmdmode=fullscreen/pageadder_page_id=35/
http://www.digido.com/portal/pmodule_id=11/pmdmode=fullscreen/pageadder_page_id=27/

I would argue with you, but then I would be wrong since you are 100% correct.

 

[FALKEN]

preface -

I am not trying to be argumentative, and I truly respect everybody who posts on here. But this is shattering my world. Here is a quote from the first article -

"Just start with a 16-bit or 20-bit source file, drop the gain a tiny amount and capture the result to a 24-bit file. All 24 of the new bits will be significant, the product of a gain multiplication that is chopped off at the 24th bit. You'll see the new lower bit activity on the Bitscope."

I am going to lunch and will read the 2nd article when I return.

 

[Reggie]

I saw somewhere where someone did the math on this stuff, and it was something rediculous like -48db in 24 bit was equivalent to 0db in 16 bit. If that is right, it makes it almost retarded to try getting right near 0db as you record the signals in 24 bit. You will have to monkey around with volumes at some point when you mix anyway, so there isn't much use worrying about the recorded volume unless you are recording something REALLY quiet that will need to be REALLY loud later in the mix.

 

[giraffe]

yea, it's not a one to one thing, it's a logarythmic increase

(spelling????)

 

[JazzMang]

so there isn't much use worrying about the recorded volume unless you are recording something REALLY quiet that will need to be REALLY loud later in the mix.

i dont know about you, but i value lower noise floors altogether.

 

[Massive Master]

That is correct. A -48dBfs PEAK in 24 bit is basically equal to the effective resolution of a 0dBfs 16 bit signal.

If more people got that through the first layer of gray matter, recordings would start to sound much better.

And of course, I'm not one to shoot for "weak" signals myself - But I don't ever get within 6dB of full-scale during tracking or mixing. There's no good reason for it. There isn't even a decent bad reason for it. It's certainly not worth stopping an otherwise perfectly good take because the drummer thought a nice POP would be cool -

Headroom is good room.

 

[FALKEN]

I saw somewhere where someone did the math on this stuff, and it was something rediculous like -48db in 24 bit was equivalent to 0db in 16 bit.

exactly as I thought - each bit = 6db.

here is a quote from the second article:

"Do not change gain (changing gain deteriorates sound by forcing truncation of extra wordlengths in a 16-bit workstation)"

I rest my case.

However, I think what Bruce means is that since the final product is 16-bit anyway, losing the last bits won't matter because they get lopped off anyway. Unless you dither, which I hope you do.

Thanks for the links!!!

 

[Reggie]

i dont know about you, but i value lower noise floors altogether.

What noise floor? The one hovering around -110db in a decent digital system? Like I was trying to explain, maybe if you record a kick drum that only peaks at like -65 db (give or take) and then later raise the volume to 0db would you have to worry much about the digital noise floor. But in real world practice, peaks at -10 or -12 db are perfectly fine.

 

[FALKEN]

What noise floor? The one hovering around -110db in a decent digital system? Like I was trying to explain, maybe if you record a kick drum that only peaks at like -65 db (give or take) and then later raise the volume to 0db would you have to worry much about the digital noise floor. But in real world practice, peaks at -10 or -12 db are perfectly fine.

A peak at -12 will require a loss of 2 bits to get it to 0. then, when you limit your final mix up another 12 db, your 24 bit file will be only 20 bits, and your digital noise will be 4x as loud. on just that one track. plus, your mixer noise.

I hate to be causing a ruckus, but I guess if you have fantastic gear it probably doesn't matter. Most of my gear is total crap so it matters.

 

[giraffe]

in a 24 bit digital system any noisefloor worth worrying about comes from my shitty pre-amps

 

[Reggie]

A peak at -12 will require a loss of 2 bits to get it to 0. then, when you limit your final mix up another 12 db, your 24 bit file will be only 20 bits, and your digital noise will be 4x as loud. on just that one track. plus, your mixer noise.

I hate to be causing a ruckus, but I guess if you have fantastic gear it probably doesn't matter. Most of my gear is total crap so it matters.

Don't forget about the difference between significant bits and least significant bits. And the fact that you will have to dither down to 16 bits for the CD. You could turn the digital noise up around 15x as loud and still not have it be in the way of a mix. The musical noises, amp hums, room decay, and mic & preamp self noise will still cover it up easily.


And no ruckus caused at all; just having a friendly discussion.

 

[grn]

you guys kind of went off on a tangent without answering my question. oh well.

It starts at the mics you use up until the final cut hits the CD for burning. The lower you can get your noisefloors, the higher dynamic range you can achieve. Good gear means more clairty for you.

yes, but how do you personally do this. just as an example. do you record at very high levels with the mics (as high as you can without distortion)? do you compress certain instruments and not others? do you compress/limit the whole mix?

yeah I know you can't make crap sound good louder, it's just louder crap. that's an obvious statement. I didn't want to know that. I want to know, if you have something good, how you can record it properly to be able to turn up the whole mix at the end fairly loud and get a decent sound. then maybe on top of that limit/compress it if you want to match the current sound of the major record labels. sheesh.

what do YOU do? do use any specific mic techniques for certain instruments? do you put someone record acoustic on a piece of plywood when recording? anything you do to get that mix loud AND clear, without compress/limit at the end I want to know. I'm rambling again.

 

[FALKEN]

try mic'ing "off-phase". they say if you use a 57 from the side instead of the front that it creates "natural" compression. like I said, compress on the way in. others have said play evenly. do not overlook a vocalist "working" the mic. create a more even track by backing off on loud parts. get good levels. cut out lows on instruments that dont need them (on the way in!). the typical rock song on the radio today has most of the lows cut out of the guitars. and i think the most important one, which was said at the very beginning, is to play evenly and not too loud! create separation by using different gear on different tracks.

obviously you dont mean "highest possible potential volume" because you know how to use a limiter. what you mean is "highest possible potential volume without sounding like shit" .......right?

 

[scottboyher]

What are you trying to record? (Acoustic Guitar, Vocals, Electric Guitar, Mic'd Amp, etc)

What is your signal chain going to the Ardvark?