RMS Issues -Sound Forge - PC - MAGIX Community Forums Archive

Subject:RMS Issues Posted by: jlim Date:2/11/2000 5:02:00 PM Hello, I was wondering if anybody could help clarify exactly what RMS is and it's applications? I read the SoundForge help page on it but don't quite understand how it works with normalizing. I've been using dynamic compression instead of normailizing and was also wondering how RMS relates with that. I apologize for being vague. Any insight would help. Thanks in Advance Jerry Lim MADMedia Wilmer Eye Institute Johns Hopkins Hospital http://www.madlab.jhu.edu
Subject:Re: RMS Issues Reply by: Dale_Maxfield Date:2/11/2000 7:12:00 PM Jerry, RMS is basically the "continuos peak load" of the sound. If a speaker can handle a peak of 100watts, it's RMS is usually around 70Watts or so. In this way, you can have a continuos guitar roaring at 70Watts and the drums smacking up to 100Watts on the beat without overdriving the speaker. So normalizing the RMS lower or higher will make the file get louder overall (much like turning up the volume) where normalizing to a peak lowers the amount of headroom between the "loudest smack" of the drums and 0db. Oddly enough, some of the best discussion I've seen on the subject is at Crutchfield.com in their car audio speaker section. They define both quite succintly. Regards, Dale Maxfield Secret Government Records scarecrow@hyperchat.com Jerry Lim wrote: >>Hello, >> >>I was wondering if anybody could help clarify exactly what >>RMS is and it's applications? I read the SoundForge help >>page on it but don't quite understand how it works with >>normalizing. I've been using dynamic compression instead of >>normailizing and was also wondering how RMS relates with >>that. I apologize for being vague. Any insight would >>help. Thanks in Advance >> >>Jerry Lim >>MADMedia >>Wilmer Eye Institute >>Johns Hopkins Hospital >>http://www.madlab.jhu.edu
Subject:Re: RMS Issues Reply by: O_G_Killa Date:3/10/2000 1:34:00 PM 1006 Just to re-iterate what Dale wrote in different words..... Normalizing is a type of volume control of which there are two types, Peak and RMS. Peak- Peak scans through the file and finds the absolute loudest spot (a.k.a loudest peak). Ususally if you zoom out so you can see the whole audio file you can usually spot the loudest peaks throughout a file. After it finds the loudest one, it then takes that and makes it as loud as possible without clipping (going over digital 0dB), and then scales everything louder in proportion to that. So for example if you had a file that had a peak value of -6dB (peak value is the loudest peak within the file), the normalizer would then increase the volume of the whole file buy 6dB. Thus making what was once a -6dB peak now a 0dB peak and everything else gets 6dB louder along with it. Peak is really good on percussive instruments and/or a single instrument at a time. But it is not as effective on a whole mix, or an instrument that has a sustaining sound (like strings) or sharp attacks with sustain (like voice, or guitar). RMS- Literally translated means Root Mean Squared. It is the mathmatical equation used to quantify the percieved volume. What the heck does that mean?!?! Well I'll tell you.... Notice how sometimes you can listen to something and it sounds really quiet and every so often it has these sharp volume spikes (listening to a John Coltrane Sax solo should give you an idea of what I am talking about)? The song seems quiet to you even though there are sudden spikes in volume because the overall volume is lower. Whereas when you listen to modern music there are not that many spikes but it just seems to be louder as a whole. This is what RMS is. It is the Average power (loudness, volume, whatever you want to call it) of a recording. What it does, is it looks at all of the peaks and valleys withing the audio recording and averages them out to come up with an average volume rating. So while you may already have a file hitting 0dB when it spikes, it may still have an RMS level that is really low. This is where RMS normalizing (which in a way is like compression) comes in.... Instead of making the loudest volume the loudest possible volume like peak does, it makes the overall volume louder. But then what happens to the peaks, do they just clip? There is a section at the bottom of Sound Forge's normalize window that says, "If clipping occurs:" and then gives you several choices. By playing with these options you can find what sounds best for the audio you are editing. Most professional CDs try to have their RMS at around -6dB to -8dB but I have seen some that are higher. Try playing with it to see what gives you the effect you want. Also in the Normalize window there is a button in the lower left corner called "Scan"...or something like that. Use this to find out what the peak and RMS values of the file are before you start playing with the normalize settings. But in general, you usually use peak on percussive instruments, or all of the tracks in general just to get them to the loudest volume before peaking. and then you usually use RMS when Mixing, or on vocals, guitar, and keys. But honestly, you can use whichever for whatever you want! :-) Hope that helps shed some light on the mystery of RMS and normalizing for you. Jerry Lim wrote: >>Hello, >> >>I was wondering if anybody could help clarify exactly what >>RMS is and it's applications? I read the SoundForge help >>page on it but don't quite understand how it works with >>normalizing. I've been using dynamic compression instead of >>normailizing and was also wondering how RMS relates with >>that. I apologize for being vague. Any insight would >>help. Thanks in Advance >> >>Jerry Lim >>MADMedia >>Wilmer Eye Institute >>Johns Hopkins Hospital >>http://www.madlab.jhu.edu

1006

Just to re-iterate what Dale wrote in different words.....

Normalizing is a type of volume control of which there are two types,
Peak and RMS.

Peak-
Peak scans through the file and finds the absolute loudest spot
(a.k.a loudest peak). Ususally if you zoom out so you can see the
whole audio file you can usually spot the loudest peaks throughout a
file. After it finds the loudest one, it then takes that and makes
it as loud as possible without clipping (going over digital 0dB), and
then scales everything louder in proportion to that. So for example
if you had a file that had a peak value of -6dB (peak value is the
loudest peak within the file), the normalizer would then increase the
volume of the whole file buy 6dB. Thus making what was once a -6dB
peak now a 0dB peak and everything else gets 6dB louder along with it.

Peak is really good on percussive instruments and/or a single
instrument at a time. But it is not as effective on a whole mix, or
an instrument that has a sustaining sound (like strings) or sharp
attacks with sustain (like voice, or guitar).

RMS-
Literally translated means Root Mean Squared. It is the mathmatical
equation used to quantify the percieved volume. What the heck does
that mean?!?! Well I'll tell you....

Notice how sometimes you can listen to something and it sounds really
quiet and every so often it has these sharp volume spikes (listening
to a John Coltrane Sax solo should give you an idea of what I am
talking about)? The song seems quiet to you even though there are
sudden spikes in volume because the overall volume is lower. Whereas
when you listen to modern music there are not that many spikes but it
just seems to be louder as a whole. This is what RMS is. It is the
Average power (loudness, volume, whatever you want to call it) of a
recording. What it does, is it looks at all of the peaks and valleys
withing the audio recording and averages them out to come up with an
average volume rating. So while you may already have a file hitting
0dB when it spikes, it may still have an RMS level that is really
low.

This is where RMS normalizing (which in a way is like compression)
comes in.... Instead of making the loudest volume the loudest
possible volume like peak does, it makes the overall volume louder.
But then what happens to the peaks, do they just clip? There is a
section at the bottom of Sound Forge's normalize window that
says, "If clipping occurs:" and then gives you several choices. By
playing with these options you can find what sounds best for the
audio you are editing. Most professional CDs try to have their RMS
at around -6dB to -8dB but I have seen some that are higher. Try
playing with it to see what gives you the effect you want.

Also in the Normalize window there is a button in the lower left
corner called "Scan"...or something like that. Use this to find out
what the peak and RMS values of the file are before you start playing
with the normalize settings.

But in general, you usually use peak on percussive instruments, or
all of the tracks in general just to get them to the loudest volume
before peaking. and then you usually use RMS when Mixing, or on
vocals, guitar, and keys. But honestly, you can use whichever for
whatever you want! :-)

Hope that helps shed some light on the mystery of RMS and normalizing
for you.

Jerry Lim wrote:
>>Hello,
>>
>>I was wondering if anybody could help clarify exactly what
>>RMS is and it's applications? I read the SoundForge help
>>page on it but don't quite understand how it works with
>>normalizing. I've been using dynamic compression instead of
>>normailizing and was also wondering how RMS relates with
>>that. I apologize for being vague. Any insight would
>>help. Thanks in Advance
>>
>>Jerry Lim
>>MADMedia
>>Wilmer Eye Institute
>>Johns Hopkins Hospital
>>http://www.madlab.jhu.edu