Recording Levels

Recording Techniques, People Skills, Gear, Recording Spaces, Computers, and DIY

Moderators: drumsound, tomb

Post Reply
User avatar
farview
tinnitus
Posts: 1204
Joined: Tue Aug 31, 2004 1:42 pm
Location: St. Charles (chicago) IL
Contact:

Post by farview » Thu Sep 28, 2006 11:27 am

There seems to be a lot of confusion between average level and peak level. I don't think anyone has said anything about peaking at -18db. That isn't it. You are supposed to have your signal averaging (hovering around, the meat of the signal, etc...) at about -18db. That leaves you 18db worth of headroom for the transients. If a snare hit or two peak at -1dbfs, you are still good. This would be much easier for everyone to get if everything had VU meters instead of (or in addition to) peak meters. VU meters being slower than peak meters, they give you an idea of what your average level is. That average level is supposed to be at 0dbVU. 0dbVU will equal somewhere around -18dbFS on your DAWs meters. Transients will excede that, they are supposed to.

adhooker
alignin' 24-trk
Posts: 63
Joined: Fri Jun 16, 2006 7:46 am
Location: Brooklyn, NY

Post by adhooker » Thu Sep 28, 2006 11:41 am

farview wrote:There seems to be a lot of confusion between average level and peak level. I don't think anyone has said anything about peaking at -18db. That isn't it. You are supposed to have your signal averaging (hovering around, the meat of the signal, etc...) at about -18db. That leaves you 18db worth of headroom for the transients. If a snare hit or two peak at -1dbfs, you are still good. This would be much easier for everyone to get if everything had VU meters instead of (or in addition to) peak meters. VU meters being slower than peak meters, they give you an idea of what your average level is. That average level is supposed to be at 0dbVU. 0dbVU will equal somewhere around -18dbFS on your DAWs meters. Transients will excede that, they are supposed to.
Ok, that makes sense to me. I think I'm in the clear then. I was just using this plugin to check RMS and peaks:
http://www.plugorama.com/customer/produ ... ductid=156

User avatar
farview
tinnitus
Posts: 1204
Joined: Tue Aug 31, 2004 1:42 pm
Location: St. Charles (chicago) IL
Contact:

Post by farview » Thu Sep 28, 2006 12:46 pm

RodC wrote:Look a the add for every DAW software package out there, they sho almost every signal peaking as high as possible. When you zoom in enough to make the waveform useful you can fit about 3 tracks on a 19" LCD !
There should be another control that just makes the wavform bigger without making the track taller.

User avatar
Mark Alan Miller
dead but not forgotten
Posts: 2097
Joined: Wed Apr 14, 2004 6:58 pm
Location: Western MA
Contact:

Post by Mark Alan Miller » Thu Sep 28, 2006 12:51 pm

faairview: yes! it's all about average levels in this context, being at -18 or whatever reference level we're talking about. Leaving room for transients and the occasionally over excited musical moment.
Looking over at the meters right now, I see a range of peak levels from -15 to smacking up close to 0dbfs. Average levels from -20 on up to about -10. All sounds great. :)
he took a duck in the face at two and hundred fifty knots.

http://www.radio-valkyrie.com/ao/aoindex.htm - download the new record (free is an option!) or get it on CD.

User avatar
jayf
gettin' sounds
Posts: 113
Joined: Thu Sep 07, 2006 11:13 am
Location: Seattle, WA
Contact:

Post by jayf » Thu Sep 28, 2006 12:56 pm

I wanted to read-up on analog / digital gain matching with specifics about Digital Performer / MOTU interfaces, and I found this informative post from a MOTU tech on Digital Performer / MOTU interface levels and gain matching (from 2004, on the bigbluelounge.com forum).
For your reference, current MOTU interfaces have 14db of headroom above the +4db analog 0VU reference. What that means is if you send a +4db test tone into an analog mixer, you'll see a 0VU reading on that mixer (if it's properly calibrated for +4db reference). We all know that analog zero is not considered a top end. You plug your mic in and peg the meters, right? The question is, how far past that analog zero can you go? The answer with MOTU interfaces is another 14db. More than that and you get input clipping.
So, with my own quick experiment, it looks like 0 VU ~= -14db on my MOTU 828mkII interface.

User avatar
farview
tinnitus
Posts: 1204
Joined: Tue Aug 31, 2004 1:42 pm
Location: St. Charles (chicago) IL
Contact:

Post by farview » Thu Sep 28, 2006 12:57 pm

Just to further elaborate: Some sources have a higher peak to average range than others.

Drums, for example, have a large transient and a quick decay. Havings something like that peaking above -6dbfs is not a problem.

Distorted guitars, square wave keyboard lead sounds, etc... have almost no transients. If you make these signals peak at -1dbfs, you are really running your preamps 17db hotter than you should.

Obviously, sometimes it's good to hit the level hard to get a certain sound out of a piece of equipment, but it is important to understand how it was meant to be used. That way, you will have a better understanding of why things happen the way they do.

User avatar
Mark Alan Miller
dead but not forgotten
Posts: 2097
Joined: Wed Apr 14, 2004 6:58 pm
Location: Western MA
Contact:

Post by Mark Alan Miller » Thu Sep 28, 2006 1:36 pm

farview wrote:Distorted guitars, square wave keyboard lead sounds, etc... have almost no transients. If you make these signals peak at -1dbfs, you are really running your preamps 17db hotter than you should.
Or 13 db, if you're using MOTU stuff, apparently. :)
he took a duck in the face at two and hundred fifty knots.

http://www.radio-valkyrie.com/ao/aoindex.htm - download the new record (free is an option!) or get it on CD.

User avatar
farview
tinnitus
Posts: 1204
Joined: Tue Aug 31, 2004 1:42 pm
Location: St. Charles (chicago) IL
Contact:

Post by farview » Thu Sep 28, 2006 1:42 pm

Mark Alan Miller wrote:
farview wrote:Distorted guitars, square wave keyboard lead sounds, etc... have almost no transients. If you make these signals peak at -1dbfs, you are really running your preamps 17db hotter than you should.
Or 13 db, if you're using MOTU stuff, apparently. :)
Correct.

User avatar
RodC
dead but not forgotten
Posts: 2039
Joined: Thu Dec 30, 2004 8:53 pm
Location: Right outside the door
Contact:

Re: link broke

Post by RodC » Thu Sep 28, 2006 5:05 pm

kweis7 wrote:http://www.soundonsound.com/sos/may00/a ... igital.htm

this link was just working, now it is not??? sorry bout that if it is broke for you.
Seems this guy suggests to push the -10 switch if you can. I think I'll try some of all the suggestions combined.
'Well, I've been to one world fair, a picnic, and a rodeo, and that's the stupidest thing I ever heard come over a set of earphones'

http://www.beyondsanityproductions.com
http://www.myspace.com/beyondsanity

User avatar
RodC
dead but not forgotten
Posts: 2039
Joined: Thu Dec 30, 2004 8:53 pm
Location: Right outside the door
Contact:

Post by RodC » Thu Sep 28, 2006 5:27 pm

In cakewalk SPE you can change the zoom factor to chop off a few db on either side of the wave form. Just click on the DB scale on the left of the track waveform and move it up and down. You can make a very small wave form fill up the whole track view.

Makes the new smaller waves easier to view.
'Well, I've been to one world fair, a picnic, and a rodeo, and that's the stupidest thing I ever heard come over a set of earphones'

http://www.beyondsanityproductions.com
http://www.myspace.com/beyondsanity

User avatar
Mark Alan Miller
dead but not forgotten
Posts: 2097
Joined: Wed Apr 14, 2004 6:58 pm
Location: Western MA
Contact:

Post by Mark Alan Miller » Thu Sep 28, 2006 8:13 pm

farview wrote:
Mark Alan Miller wrote:
farview wrote:Distorted guitars, square wave keyboard lead sounds, etc... have almost no transients. If you make these signals peak at -1dbfs, you are really running your preamps 17db hotter than you should.
Or 13 db, if you're using MOTU stuff, apparently. :)
Correct.
Still, if a preamp has 28 db, or even 18 db, headroom above nominal operating levels at +4 (which is quite common), in either scenario it should be able to do it relatively cleanly. Borderline, perhaps, but...

Sorry, don't mean to split hairs so.
he took a duck in the face at two and hundred fifty knots.

http://www.radio-valkyrie.com/ao/aoindex.htm - download the new record (free is an option!) or get it on CD.

User avatar
RodC
dead but not forgotten
Posts: 2039
Joined: Thu Dec 30, 2004 8:53 pm
Location: Right outside the door
Contact:

Post by RodC » Fri Sep 29, 2006 5:22 am

Mark Alan Miller wrote:
farview wrote:
Mark Alan Miller wrote:
farview wrote:Distorted guitars, square wave keyboard lead sounds, etc... have almost no transients. If you make these signals peak at -1dbfs, you are really running your preamps 17db hotter than you should.
Or 13 db, if you're using MOTU stuff, apparently. :)
Correct.
Still, if a preamp has 28 db, or even 18 db, headroom above nominal operating levels at +4 (which is quite common), in either scenario it should be able to do it relatively cleanly. Borderline, perhaps, but...

Sorry, don't mean to split hairs so.
Exactly what I was thinking. What I have came away with is:

If you think about it, im sure there are plenty of analog ppl who record at transient levels greater than 0 on their vu. (They would never see it on their slow meters) Where do you think their peaks are? They may never know if they are recording to tape.

Im no longer going to worry if I get around -18 if it sounds good.

Im going to feel OK about recording multiple guitar tracks at lower volume if I know Im going to have to turn them down later.

Im always looking for ways to prevent edits, destructive or not.
'Well, I've been to one world fair, a picnic, and a rodeo, and that's the stupidest thing I ever heard come over a set of earphones'

http://www.beyondsanityproductions.com
http://www.myspace.com/beyondsanity

User avatar
farview
tinnitus
Posts: 1204
Joined: Tue Aug 31, 2004 1:42 pm
Location: St. Charles (chicago) IL
Contact:

Post by farview » Fri Sep 29, 2006 9:52 am

Mark Alan Miller wrote: Still, if a preamp has 28 db, or even 18 db, headroom above nominal operating levels at +4 (which is quite common), in either scenario it should be able to do it relatively cleanly. Borderline, perhaps, but...

Sorry, don't mean to split hairs so.
There is a big difference between handling transients at +10 (what it was designed to do) and having to maintain the bulk of the signal at +10. (not what it was designed to do)

The power supplies of the equipment will easily handle random spikes at +20, but will not necessarily be able to keep up with a full on signal anywhere near it.

The specs that are quoted for a piece of equipment are only true at and around line level, that includes things like self noise.

jeddypoo
re-cappin' neve
Posts: 603
Joined: Thu Oct 07, 2004 10:24 am
Location: brooklyn
Contact:

k

Post by jeddypoo » Fri Sep 29, 2006 9:59 am

I'll just say this again: I really think that a lot of the "mud" in people's mixes has to do with not making them quiet enough. And I think that's even more important than the specific maths involved in gain structure when printing. I mean, I always print quietly because it just seems to make sense working in digital mediums, but I really feel like mixing too loudly is more of an issue than this is, at the end of it.

However, I still think this is a really valuable debate. If it can stay polite- which it's gotten close to not doing already.
I find adherence to fantasy troubling and unreasonable.

User avatar
Mark Alan Miller
dead but not forgotten
Posts: 2097
Joined: Wed Apr 14, 2004 6:58 pm
Location: Western MA
Contact:

Post by Mark Alan Miller » Fri Sep 29, 2006 12:59 pm

farview wrote:
Mark Alan Miller wrote: Still, if a preamp has 28 db, or even 18 db, headroom above nominal operating levels at +4 (which is quite common), in either scenario it should be able to do it relatively cleanly. Borderline, perhaps, but...

Sorry, don't mean to split hairs so.
There is a big difference between handling transients at +10 (what it was designed to do) and having to maintain the bulk of the signal at +10. (not what it was designed to do)

The power supplies of the equipment will easily handle random spikes at +20, but will not necessarily be able to keep up with a full on signal anywhere near it.

The specs that are quoted for a piece of equipment are only true at and around line level, that includes things like self noise.
It's true... headroom for transient peaks vs rms level is somewhat different. Hence the example of drums peaking higher than distorted guitar being so apt. Your clarification is appreciated. I didn't state my thoughts clearly enough...

Of course, the ability to handle higher rms levels above nominal operating level will vary from unit to unit. Ofteh, cheaper gear will be the less forgiving in this regard than more expensive (and likely more robustly designed and conservatively tested and rated) will be able to be 'pushed' more forgivingly. Some really well designed gear is overengineered in the power supply department for just this reason. And it usually shows in the price tag (wall warts are signs of this not being the case, at least often.)
jeddypoo wrote:I'll just say this again: I really think that a lot of the "mud" in people's mixes has to do with not making them quiet enough. And I think that's even more important than the specific maths involved in gain structure when printing. I mean, I always print quietly because it just seems to make sense working in digital mediums, but I really feel like mixing too loudly is more of an issue than this is, at the end of it.

However, I still think this is a really valuable debate. If it can stay polite- which it's gotten close to not doing already.
Agreed - and not being quiet enough can also very often be caused by overcompression, both on the channel-by-channel front and on the L/R buss.

And I hope I haven't said anything that offended anyone or was taken as anything impolite. I certainly didn't mean to...
he took a duck in the face at two and hundred fifty knots.

http://www.radio-valkyrie.com/ao/aoindex.htm - download the new record (free is an option!) or get it on CD.

Post Reply

Who is online

Users browsing this forum: No registered users and 43 guests