peak + rms question

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

Moderators: drumsound, tomb

Post Reply
User avatar
2121TrumbullAve
gimme a little kick & snare
Posts: 77
Joined: Sat Aug 12, 2006 8:44 pm
Location: Denver

peak + rms question

Post by 2121TrumbullAve » Sun Nov 04, 2007 9:42 am

so in sonar, if i have peak + rms on my record meters, when i record drums, say kick for example, the gap between the peak and rms is huge, like over 20 db.

so is the object to keep the peak from exceeding 0db, always? because in order to prevent peaks from reaching 0db, my signals are reading -30 to -24 rms, and i'm concerned that this is not quite hot enough.

i understand that a good tracking level is like -18, but if i do this the peaks are way over 0db.

any insights are appreciated.

- J
*insert pricey DAW specs here

User avatar
Jeff White
ghost haunting audio students
Posts: 3259
Joined: Sun Feb 26, 2006 6:15 pm
Location: Philadelphia, PA
Contact:

Post by Jeff White » Sun Nov 04, 2007 12:08 pm

Compression will squash that dynamic range to a user friendly level. Limiting will keep peaks in check without bringing up RMS to the extent that straight-up Compression will. Also, micing techniques should be investigated. And remember that most instrument tracks can and should be high-passed. I bet that the low energy in those tracks is driving Peaks, and without anything below 80Hz-100Hz in certain tracks you may notice the gap closing a bit.

Also, check out this meter and make sure that Sonar's meter is totally accurate:
http://www.sonalksis.com/index.php?section_id=99

Oh, and remember, that -18dB reference isn't a reference unless you mean -18dBfs, meaning -18dB below absolute digital zero. Some DAW meters are set up like analog (Digital Performer is like this) and allow 6dB of level above "zero". In this case I shoot for -12dB on DP's meters. Some folks say -14dBfs is what to shoot for. I basically say anything between -12dBfs and -18dBfs RMS is safe if the peaks are in check. And remember... This is only relevant for 24-bit recordings.

Jeff
I record, mix, and master in my Philly-based home studio, the Spacement. http://jeffwhiteaudio.com/

philbo
suffering 'studio suck'
Posts: 469
Joined: Fri Feb 13, 2004 11:43 pm
Contact:

Post by philbo » Sun Nov 04, 2007 5:51 pm

It is totally common for percussion to have a high peak to RMS ratio - - that is the nature of drum sounds. If you are recording in 24 bit, you should be able to run with RMS levels down to -35 dBFS without any issues, assuming your signal chain isn't noisy. If you're using 16 bit recording, yes, you have a problem. Switch to 24 bit if possible, or use compressors on the signal if you can't.

BTW, the ratio of peak to RMS is called 'crest factor' - a high crest factor in audio generally means that the attack of the sound is hundreds of times louder than the sustain and decay. (Which kinda sounds like what drums do...)

A lot depends on the sound you're going for - - compressing drums has been known to infuriate people who do jazz, classical, and audiophile grade acoustic music; these are people who want the sound of their instrument faithfully recorded and reproduced, and compressing them will get you fired or at least set you up for a thorough ass-chewing. For rock, pop, hiphop, country & similar, though, it's standard practice.
________
Freexxx
Last edited by philbo on Sun Mar 20, 2011 12:18 am, edited 1 time in total.

User avatar
2121TrumbullAve
gimme a little kick & snare
Posts: 77
Joined: Sat Aug 12, 2006 8:44 pm
Location: Denver

Post by 2121TrumbullAve » Sun Nov 04, 2007 6:45 pm

thanks for the replies.

i do it all ITB, so i don't have any compressors on during tracking, and i record at 24/88.

i've read in sonar forums that you can hit 'em hard, and when i have overs i do not hear anything nasty, so sonar must have some cushion, perhaps 6 dbfs.

i can see how the crest factor of drums would be large now, and i think i'll just not fret about rms levels during tracking being down around -24dbfs.
*insert pricey DAW specs here

river
pushin' record
Posts: 221
Joined: Wed Jan 25, 2006 8:05 am
Location: Nashville, TN
Contact:

Post by river » Sun Nov 04, 2007 7:38 pm

Why not just set meters to read peak when tracking drums? Shoot for the occasional peak of -6dbfs. That will leave you plenty of headroom to work with when mixing. Using RMS was a fact of life back in the analog days, and a safety margin was necessary to factor in because of the slow response of the meters themselves, but with accurate peak meters these days, why bother?
"Madam, tomorrow I will be sober, but you'll still be ugly" Winston Churchill

rwc
resurrected
Posts: 2333
Joined: Sun Sep 09, 2007 8:21 pm
Location: Bed Stuy, Brooklyn

Post by rwc » Sun Nov 04, 2007 10:16 pm

ignore RMS and look at peak for setting record levels.

digital overs sound like ass.

I like to set it when recording drums so the peaks don't go over -10. Because if he decides to hit other drums at the same time + the drum mic I am looking at on the level meter harder.. it may go up quite a bit.

You've got 144 dB of dynamic range. Who cares if the RMS is -24? I'd rather record less hot than risk a digital over in my recording.
Real friends stab you in the front.

Oscar Wilde

Failed audio engineer & pro studio tech turned Component level motherboard repair store in New York

MoreSpaceEcho
zen recordist
Posts: 6463
Joined: Wed May 07, 2003 11:15 am

Post by MoreSpaceEcho » Mon Nov 05, 2007 8:45 am

RWC wrote:I like to set it when recording drums so the peaks don't go over -10.
me too. and that's kick and snare. overheads and everything else i track lower. to me there's no point in recording any hotter than that because you're just going to have to turn everything down again when you mix...

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

Post by farview » Mon Nov 05, 2007 9:08 am

For drums and other percussion, ignore the RMS and worry about peak. I normally shoot for -10 to -6dbfs just in case the drummer gets excited.

The whole -18dbfs thing is all about not distorting the analog chain (preamp, compressor, etc...) on the way to the converters. Preamps can handle loud transients without a problem. The distortion happens when you have a signal that has no dynamic range (distorted guitar, sine wav synth patches, etc...) and you are pushing the level 12db above nominal (line level)

User avatar
Jeff White
ghost haunting audio students
Posts: 3259
Joined: Sun Feb 26, 2006 6:15 pm
Location: Philadelphia, PA
Contact:

Post by Jeff White » Mon Nov 05, 2007 9:30 am

farview wrote: The whole -18dbfs thing is all about not distorting the analog chain (preamp, compressor, etc...) on the way to the converters.
It's also all about not clipping plug-ins like Waves RenComp or RenEQ once the audio is in the computer, and mixing your 24-48 track mix so that all tracks output on the Master Fader around -6dB to -3dB, fit for mastering. I have noticed a big big difference in sonic quality when my tracks are individually around "-18dBfs", which of course allows peaks to be much higher. Obviously percussion is going to have dynamic range. Obviously you don't want to squash the life out of mix.

Waves plug-ins clip in Digital Performer. It is nasty sounding and totally obvious. Keeping individual tracks' RMS around -18dBfs is a good solution to this, and across the board ITB mixes benefit from mixing like this. I apply this same thing to my signal chain coming into the converter, allowing enough headroom.

Jeff
I record, mix, and master in my Philly-based home studio, the Spacement. http://jeffwhiteaudio.com/

User avatar
JohnDavisNYC
ghost haunting audio students
Posts: 3035
Joined: Fri Oct 03, 2003 2:43 pm
Location: crooklyn, ny
Contact:

Post by JohnDavisNYC » Tue Nov 06, 2007 6:18 am

it's great that in the last year or so it seems like everyone on the interweb has finally started to take some interest in gain staging. even more amazing is the fact that in order for digital to work, you have to treat it like analog. imagine that...


the person (most likely an employee of Avid) who decided that meters in a DAW shouldn't have some sort of '0vu' reference at -18 or -12 (go DAT!!!) should be taken into the street and beaten with his own shoes. or even just making the meters turn yellow when you hit -18, then orange at -6, and red at -3.... would have probably saved us from a ton of crappy sounding digital recordings.


it is really all about just recording properly. 0vu is the standard for normal operating level... and for good reason. if everything is recorded with the RMS levels at 0vu (-18dbfs) it will sound better.


i can only hope that mastering engineers will start reading the internet and realize this.

john
i like to make music with music and stuff and things.

http://www.thebunkerstudio.com/

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

Post by farview » Tue Nov 06, 2007 10:52 am

toaster3000 wrote:i can only hope that mastering engineers will start reading the internet and realize this.

john
Mastering engineers (the good ones) always knew this. Unfortunately, their customers aren't quite up to speed and insist on having the loudest CD in the world. If they want to stay in business, they have to comply.

User avatar
JohnDavisNYC
ghost haunting audio students
Posts: 3035
Joined: Fri Oct 03, 2003 2:43 pm
Location: crooklyn, ny
Contact:

Post by JohnDavisNYC » Tue Nov 06, 2007 10:57 am

i know... just a joke.

john
i like to make music with music and stuff and things.

http://www.thebunkerstudio.com/

Fletcher Munson
audio school
Posts: 6
Joined: Tue Jun 01, 2004 9:57 pm

Post by Fletcher Munson » Thu Dec 27, 2007 12:18 am

I've found that an occasional clip of transients is not a problem.....I try to avoid them, but if something clips, I don't sweat it, as long as it's occasional. Three or four times per tune....occasional......three or four times per measure....NOT.

Post Reply

Who is online

Users browsing this forum: Snarl 12/8 and 38 guests