Volume question.

badgas

New member
I'm finally doing some recording.
I've read in my VS840 manual that the volume should be set to just below the peak.

Am I to understand that as I lay my drum, bass, rythme, lead, vocals, harp that my first basic tracks should be as loud as the recorder can go without the 'peak' light coming on?

The reason I'm asking this, it at the end of two verses, I want a low volume, far in the back harmonica playing.

Does the above make sense?
Then when I get to the mixing portion, I adjust the volume there for what I want?

Thanks for any replies on this. I've read some posts about this but they are talking about other things, not just one thing, 'harp' to be specific.
 
The farther you get away from "nominal" levels in digital, the more "bits" you lose (resolution). Record everything as close to peak (without EVER clipping), then adjust relative volumes in the mix. As for the harp, record it hot as well, pan it to either side and add a bit of verb. This will place it further back in the mix.

Bob
 
Ooo a question I can answer :eek: If you were to record the harp at the level you wanted it in the final mix you would be wasting headroom (wow, I used a real recording word), loose resolution (as bob said) and also reduce your S/N ratio (I think). Do as bob says. Plus if you decided later that u wanted to bring it up louder in the mix, then u would end up amping any noise recorded as well as the harp.
 
Sorry Sonusman for cramming this in the wrong forum. I'm a cherry at recording and techie talk yet. :(

Thanks a bunch Buffalo. That's what I thought. I appreciate your time. :)
 
Ohh. I understand Grandalf.
Resolution means nothing to me. I have over forty Gigs of drive space.
I appreciate the 'lingo'. Headroom. Now I know what it means. ;)
 
No prob badgas. Just though I would put it where it is pertinent.

You SHOULD be worried about resolution, unless you are recording at 20 bits or higher.

Some misconfusion here about things. You aren't "wasting" "headroom" by tracking at a lower volume. There was a thread concerning headroom recently (is the search funtion still working?) and in fact, there are MANY engineers who feel that when tracking in digital that it is best to track to levels that the instrument will wind up in the final mix so that DSP doesn't have to be applied to that instrument. An interesting question is whether the DSP of a volume change, or a few bits or resolution loss is worse. Having worked quite a bit recently on Sonar, I would rather lose a few bits of resolution (of course, I record in 24 bit anyway) than to have to apply a volume change to the audio. I can STILL hear the difference in the original audio track when ANY DSP is applied. The track lose's it depth and ear friendliness. It is like I can just hear the sound starting to pull away from me. I have found that the more DSP applied, or the more extreme the DSP is (like a 12dB reduction in volume) the worse the effect is.

I think what you should play around with is tracking something at the volume you want it at in the mix, then track the same thing at as high of a level as you can and apply DSP to reduce it's volume and hear for yourself if there is a difference in the sound between them.

I can tell you this. High resolution benefits really do depend on the type of source you are recording. Take keyboards, or distorted guitar. These sounds have either been sampled already (keyboard sounds) or are square waves. Increasing bit depth for them really probably won't pay off, UNLESS, they have a lot of trail off's in the volume (like holding out decaying notes). So, with these types of sounds, I doubt that you are going to hear that much difference between tracking them as close to digital full scale and say -30dB down.

BUT, take something very percussive, well, like percussion! :) Now, these are very short lived notes that DO have a resonance that fades out very quickly. The sound ALWAYS fades out after each hit and that fade out is a very BIG part of it's sound. In this case, increased bit resolution becomes very important and more bits = lower noise floor and more bits to represent low level sound. The fade out of the resonance will sound smooth and more accurate. Cool!!!

So, with percussive parts, I might tend to go for very high resolution, and on more sustained types of sounds, not worry about it. Get it?

You guys don't understand headroom yet. Keep bringing it up and I might attempt to explain it! :)

Eddie
 
In one of those threads it was explained that if the audio is averaging around say -10dB, you would have 10dB of "headroom". Simply, that is it! You CREATE headroom on your system. It doesn't make much sense to apply the term headroom and sound to noise ratio the same thing, because while in a sense, you COULD say that your sound to noise ratio, and more importantly, your DYNAMIC range is headroom, you are overly complicating what headroom is referring to. Dynamic Range vs. Sound To Noise Ratio is that more complicated thing. I will attempt to illustrate.

Music usually WANTS to room to get louder. Since ALL recording mediums have a maximum voltage that can be stored (yes, even analog....don't confuse the highest voltage value that tape can handle with tape saturation and PERCIEVED volume) you have a "loudest" volume that can be achieved. So, in a way, we have a fixed ceiling to loudness in electronics. If your audio was at full volume 100% of the time, no dynamics would be possible. That would suck. It is not ear friendly at all, even though it is possible to create clean sounding audio with no dynamics (or barely any), it is just tiring to listen to. So, we NEED dynamics. How much dynamics does you song need? Here is the question, and that has do with "headroom". If you need your music to be able to get 10dB louder in certain parts, then you need 10dB of headroom possible. Period! There is nothing else to it. I have looked at audio recordings that don't even allow that much headroom. On certain types of music, it is not terribly offensive to only have 4-6dB of headroom (although, I only say terribly offensive based upon current average volume standards of today....I like having more headroom...I don't like making tonal sacrifices that are neccesary to have such high average volumes...).

Can music still SOUND like it retains dynamics with very little headroom? Sure can. An understanding of the Fletcher/Munson Relative Loudness Curves and multi-band compression will help you more or less do that if you like that kind of sound. 80% of the voltage in your average volume is comprised of frequencies 200Hz and below. It is possible to just compress the hell out of those frequencies so that they STAY at about 80% of your voltage at all times, and allow midrange frequencies to grow. You could make this ratio even higher, like 90% and let the midrange make up the rest if you want REALLY LOUD.

Anyway, think of headroom as the amount of voltage you set aside for the audio to get louder. That is ALL it means. If you don't require much headroom, then you don't. If you do, you need to CREATE it. You have to keep average volumes LOWER to create more headroom. Get it?

Experience will tell just how much headroom you actually need. Like I said earlier, some genre's of music just don't require that much headroom. Others, like orchestral recordings require more of it.

Sound to noise ratio vs. dynamic range are two different things. Sound to noise ratio has to do with the volume that electronic noise is at. But, audio can heard that is actually at a lower voltage than the equipments self noise is at (dithering uses this concept! :)). So, DYNAMIC RANGE in gear has MORE of a bearing on minimum to maximum voltages that can be stored than sound to noise ratios do. But, a really crappy sound to noise ratio DOES have a bearing too, just not as much. Think of the s/n ratio as the level at which things start deteriorated really fast. The more distance between the lowest volume you really need recorded and the self noise level is of concerning. It is possible to overcome s/n ratio problems with compression! (this is what noise reduction like dolby do).

So, we know that our ears are sensitive to the 2-5KHz region. Audio that is possibly up to 20dB LOWER than the equipments self noise level would probably be heard! Cool. But, other frequencies that low might not be heard because the broadband noise the electronics make would "mask" it. Too bad...:( So, s/n ratio's are considered "nominal" when listed (usually they use 1kHz as the tone used for the displayed spec for s/n ratio. Other frequencies it could be worse, some better.

But, dynamic range is usually FIXED on any device. The unit can only work with a range of voltages. Good circuit designs allow more range, poor designs usually allow for less range.

So, s/n noise ratio/dynamic range aren't really what are being talked about when "headroom" is being talked about. Headroom is just how much voltage ABOVE average levels is set aside for the audio to "grow". You CREATE it and can have as little as NONE, or as much as the total dynamic range possible in the signal chain.

Man, I haven't talked about this kind of shit directly for a long time. Thanks for the opportunity to write it out. I am not sure yet if I have explained it very well. If you don't understand something, or need a clarification, please ask as this helps me sharpen how I describe stuff.

Eddie
 
Actually Gandalf, you were very close!!!

I do agree though that you will be increasing the noise floor if you have to boost low level audio later on. In the digital realm though (pure digital I mean...meaning, once it is digital it STAYS digital until the final D/A conversion of the final mix in the CD player) a case could be made that having to lower the volume of something is also creating noise because of quantiniztion errors. Of course, then you could say that having to raise the volume will have those errors too, PLUS, increasing the noise that was contained in the recording.

A lot of times, the noise is eliminated adequately by applying a High Pass Filter to the track, which has other benefits too! :) Most of that low level noise is 60Hz hum. High Pass set at around 70Hz assures it is out of the equasion at the master buss of the mixer. Setting it even higher on instruments that don't need low frequency content helps create more "depth" for the instruments that DO need low end.

Eddie
 
Ed- Do you really hear a noticeable differance in audio quality if you apply ANY volume change in DAW? I could understand boosting or normalizing the signal causing artifacts but to just lower the volume a little?

Not argueing with you here, just trying to make sure that is what I am reading. I would think if you lowered the volume it would also lower the perception of any artifacts. What type of changes do you notice in a lowered volume?
 
I happens more when the track count is high. The track seems to lose it's roundness and depth. Becomes a tad more harsh. It is subtle, but the more DSP applied, the less roundness and depth it seems to have. I just don't hear the same artifacts when mixing on analog mixers, with analog dynamic processors.

Eddie
 
alright ed...

you've really got me thinking now on two points...

what I thought before (i was the one who "explained" the -10db average = 10db headroom thingy in the other thread) - was that recording at lower levels was all well and good if that was what you wanted..

my only concern with this was that I had heard that "cheaper" preamps...like XDRs, or my Aardvarks...for example....did not put out their best sound when recording at lower levels. That you really needed to drive a Mackie, to make it sound good.

So, I figure sure...record as loud as you can - and, i think that's the only sensible rule for this stuff. "can" being defined by how much headroom you need and such.

The two points I am now wondering about -

In terms of volume adjustment in the mix - you say that this degrades the quality? So, you mean just turning the level up or down...in say, Logic...will hurt the quality? (it makes sense..but, it sucks! heh). My question then is how do you mix? If something is too loud....it's going to be impossible to go retrack that...just becuase you feel it needs to come down 1 or 2 db. Whats the solution? You are telling me you should just eliminate level adjustment as part of your mixing process? Whoa, thats going to be a hard one for this cat. heh. Also..."DSP" in general...hurting the quality. Basically that would eliminate doing any EQ...or effects or anything else? Maybe I am misunderstanding this a bit...or misunderstanding the impact? Becuase if you can't use any DSP - you can't mix, essentially.

The other concern is on bit loss. I don't see the logic of why you lose bits of data on a quieter source? I mean...the only thing I can think of is "there isn't as much there". It's like, if I am recording a really quiet acoustic guitar....vs. a really loud one - yes I am going to have more 0's in the quiet one...but that isn't a *bad* thing? It just means it's a quieter passage...the quality of the instrument could still be as good fidelity wise as the loud one...

Or am I way off on that? I think "losing bits of resolution" sounds really scary...becuase...everyone is so concerned with "yes! I have 24 bits!" not knowing what that means really.

And I am not claiming that I do....all of this digital stuff is weirding me out ;). So, please explain those concerns further if you can...much appreciated.

Last week I was fretting about "summing", now this...sheesh. heh.

Thanks,
Wes
 
I have a question that I think has something to do with this as well.
I've been programming drum parts in Logic and using velocity to give the hits the human feel. In Battery, I have modulation set; velocity-to-volume. I've been running all percussion sounds (except kicks) on one channel because I only have 16 tracks available. So when I am finished- many samples with volumes that were originally near 0db have been attenuated very much. And then I have to go back and crank the volume on the entire instrument track- and even add a gain plugin as well. Now I'm thinking that I must be really skrewing things up since I'm taking these sounds and attenuating the hell out of them and then turning around to crank the overall volume as well. Should I possibly reduce the range of velocity variation and crank all the note velocities until the hardest hit is at 127 so that I will not have to make up so much gain in the end result? Or maybe separate the louder sounds from the softer ones onto another channel? Looks like this whole situation is really just a give and take sort of thing. I guess I should just find a balance that sounds good. i dono.
I am producing dance music- would you suggest that this 'loss of resolution' even matters when much of it sounds synthetic anyway? Sometimes I wonder if by listening to you guys- I'm really trying to apply techniques that really benefit acoustic recordings of live instrumentalists much more. Your thoughts would be greatly appreciated.
 
24 bits gives u 2 to the power 24 discreate levels that can be recorded, that means that each little "jump" from one level to another is quite small. If u record the same signal at 1/2 the volume, then the jumps are effectively twice as big compared to the signal. I think this is what Eddie ;) means by loosing bits. then again I could just be talking out of my arse again :rolleyes:
 
dru said:

I am producing dance music- would you suggest that this 'loss of resolution' even matters when much of it sounds synthetic anyway? Sometimes I wonder if by listening to you guys- I'm really trying to apply techniques that really benefit acoustic recordings of live instrumentalists much more. Your thoughts would be greatly appreciated.

If you are using soft synths...and the sound is being created digitally and stays in the digital domain the whole time - then no the output level you use for it INITIALLY is not going to degrade the sound quality. I am not sure on where that falls into acoustic recordings...yet. Though...like I said, I think the idea of "losing resolution" isn't like, "losing resolution" on an image. Where you are getting a worse looking picture...

To me the fact that you "lose" bits is obvious. If something is quieter,you will have more 0's - right? I think the whole "losing information/resolution/bits" idea is somewhat of a circular argument. Follow me here: "I am recording something at a lower volume...therefore I am losing bits" - would be the same as saying "I am recording something at a lower volume therefore my end result has a lower volume" .

I could be way off...but, I am not thinking *recording* at a low volume and losing bits that way is a bad thing for your sound. I think it should sound exactly like what you put in...

Judging by what Ed is saying though, adjusting the volume *after* you have the track done can really hurt the overall sound. And the more I think about it the more I agree.

I hate to think that the "digital revolution" will never quite be completed. Like, if we have to record to digital...mix in analog realm...and then send it back to digital....to get the absolute best sound, then that sucks to me. I guess there is always the point of looking into better DSPs - and I think I am going to post another topic about that in "the rack", since this has really spiked my interest.
 
Ahhhhh...finally someone is seeing the light!!! :)

Yup. Usually when you see that someone "produced" their project in ProTools, they probably also sent it off to Mr.Bigtimedood mixing engineer who will mix it on an ANALOG console BACK INTO a digital system. Interesting eh?

Why would they do that? Well, because DSP sucks ass in comparison to analog for many functions in mixing. That is why. Big time, good sounding analog consoles with very nice automation packages already exist. For manufactures to come out with digital consoles that every START to rival the big time analog consoles, the digital console would be VERY expensive. Bit time studios are NOT going to sell off a nice SSL or Amek console to buy a digital console that STILL doesn't sound as good! That is the bottom line.

It is a hard reality for home rec'ers to find out that their little Wintel boxes WILL NOT rival good analog gear.

You wanna start talking about GOOD high resolution audio, like Sonic Solutions, ProTools, Sadie, and Soundscape, well hell yes!!! They will deliver very good results (MUCH better than you are going to achieve in Logic, Sonar, etc....). But look into what one of those systems is gonna cost you!!!

I keep saying it, and here and there somebody actually realizes it, you gotta PAY for good quality. Buy cheap buy twice.

This whole silly PC recording thing is what demo's are made of, NOT serious audio production. But it IS cool for demo production, and as a way that many of you can start to learn about audio production.

Eddie
 
I forgot to add though that there is some very exciting digital stuff in developement right now. I have read parts of white papers on some of it. VERY COOL. I know I sound like an idiot bringing it up because there is no WAY I can share the info with getting people in trouble, so no names, no mention to the technology, etc....But man, some of the stuff that will be coming out in a couple few years IS exciting. It will be some time though before this stuff makes it down to the price range that many of us can afford for a home studio. BUT, when it does come out, and big time studios jump on the bandwagon to buy it, nice little ProTools type system will flood the market for relatively cheap prices!!!

Eddie
 
Of all the cheaper DAW systems is one considered to be better than the others for sound quality? Is CoolEditPro's 32bit processing going to make it sound better overall than some of the others?
 
Back
Top