Mixing Question - Levels

[Len911]

I'm gonna leave the digital theory stuff to the Dan Lavry's of the world,lol! He has an a/d converter that even has a feature called "soft saturation" that uses digital compression before the digital conversion. What you can and can't do really depends on the design and calibration of whatever product you are using. The meters in daws may be highly accurate, but can you really tell they are 0db when they are red, or do they conservatively calibrate them? Maybe the red begins to show at -1db? Although it's not scientific, you can always make multiple mixes using different settings and compare them. That probably gives a truer picture than speculation on what either happens, doesn't or even makes a difference.
Honestly, I have run many circles in some of this theory stuff, and all I get is dizzy. Wavelab 6 has the Bob Katz meters, on top of all the other meters they have, and the Bob Katz or any other meters make a lot of sense if you use an spl meter and find and mark your volume settings on your monitor amp and headphone amp. That's where I have been having a lot of trouble, determining a point of reference with the gain settings on my monitor amps. Of course if you do that, the Bob Katz meters aren't really necessary, because they just give you a meter where 0db is either -12,-14, or -20, which you would use if you used a regular meter. Of course, you don't really need an spl meter, if you just take mixes you find ideal, and mark the settings for those mixes and use as a guide for your own.

[jdhogg]

I took a class in digital recording at UCLA, which was waaay over my head, but one thing I believe I remember: When recording tracks in digital, it's similar to signal-to-noise ratio in analog. The idea is to fill up all the "bits" (?) to get the beef on the tracks, just like getting into the red in analog on each track.

If I'm wrong about this, clue me in! But if that is correct, that means that the tracks are recorded to digital capacity, so lowering levels should not degrade the sound, just lower the volume of the sound.

Is this right? If not, I wasted my $600 because I didn't learn anything !

Thanks in advance,

Ern

You might want to get your money back

There is no such thing in digital..........because it's digital. Bits have nothing to do with saturation. The "beef" in digital has more to do with sample and hold fidelity, power and low jitter, not at what level your input is to your DAW.

What happens in volume after you record has to do with the program algorithm, not the record process.

Bits in digital refers to how often a snap shot of the wave is taken. 24 bits takes twice as many snap shots of the wave there for it is more a true picture of the wave. But how high you drive an input for recording will not change the fidelity whether it be 16 or 24 bit rates.

Bits has nothing to do with "how often a snap shot" is taken that is the sampling frequency. The number of bits that the sampler uses sets effectively the number of volume levels, the more bits the more levels the greater the dynamic range.

Ern you are right! Record at a good level and then dont worry about the fader levels.

To the original poster :- Some of the posts here are misinformed.

GOODBYE

[billg1]

I have a very unscientific method that seems to work for me. My recording process always starts with drums which usually peak the "highest" in my mix. I make sure the drums peak at -3db & then I track everything so it sounds right relative to the drums. When I mix & pan/eq, etc. I keep everything (when it's summed) to -3db on the master buss. Sometimes this means there has to be slight limiting on individual tracks (usually the snare gets loud when everything is summed) . . . never more that a db or so. If I have to limit more than that I know I just need to reduce some volume or change eq settings. When the mix sounds right like this, peaking at -3db on the master buss, everything seems to sound better after mastering, which normally just means a little master buss eq, a couple db of compression & then a good limiter to bring things up to -.2 db. I allow the limiter to "eat" no more that 1db & only of the loudest peaks. I use the opto compressor in T-racks for this on it "fast 2db" preset, it seems very transparent to me. I use elephant for limiting because it shows me dynamic range (and is also very transparent imo). At that point if it needs to be louder I can squash it an additional 1-2db with a low ratio compressor (2 to 1 max) keeping things at -.2db. I use the stock compressor (wave Hammer) in Sound Forge 10 for this. This usually yields a mix with 14-16 db of dynamic range but stills sounds comparable in volume to my reference mix ()commercial mixes (alt country/Americana, etc.).

I never pay any attention to how much I have to reduce individual tracks when things are summed as long as the initial mix sounds good at -3db. I have no idea if this is a right or wrong approach. Mastering engineers seem to always suggest mixes that peak at -3db, I just figured it made sense totry to start mixing with that in mind and it seems to help a lot.

[eeoo]

Man, I'm so clueless about this stuff. I never know what a meter is actually telling me, i just try and get my mixes to sound good when they peak around -3db. Basically I just use my ears, I have no "system", every mix is different beast for me...

eo

[Dwayne Russell]

Bits has nothing to do with "how often a snap shot"...)

YEs it does. People commonly get confused about it. Levels of samples is probably a better term. At 24 bits you are getting allot more of the picture. Put it that way.

Go ahead and keep on thinking you get to make your level "hot".

This post was made because they were having a problem with that. and you are telling him to keep doing it.

[Dwayne Russell]

Man, I'm so clueless about this stuff. I never know what a meter is actually telling me, i just try and get my mixes to sound good when they peak around -3db. Basically I just use my ears, I have no "system", every mix is different beast for me...

eo

Good idea!!

[Casey H]

Bits has nothing to do with "how often a snap shot"...)

YEs it does. People commonly get confused about it. Levels of samples is probably a better term. At 24 bits you are getting allot more of the picture. Put it that way.

Go ahead and keep on thinking you get to make your level "hot".

This post was made because they were having a problem with that. and you are telling him to keep doing it.

Bits DO NOT indicate how often a digital snapshot is taken of an analog waveform. They indicate how much resolution in each sample. With 16 bits, the resolution is 1/65,536. That means every sample is effectively rounded to the nearest 1/65,536. At 24 bits, the resolution is 1/16,777,216. Higher bits give you better granularity (e.g. accuracy) in measurements.

How often is determined by the sample rate such as 44.1K, 48K, etc. At 48K, the waveform is sampled 48,000 times each second.

Casey

[Dwayne Russell]

They indicate how much resolution in each sample.

Casey

And how do you suppose that happens?

[Casey H]

They indicate how much resolution in each sample.

Casey

And how do you suppose that happens?

Little munchkins run up and down the waveform to take samples?

[matto]

Bits has ABSOLUTELY NOTHING to do with how often a snapshot is taken....as Casey points out, it is the sampleRATE or sampleFREQUENCY that would indicateds the "how often". The words RATE and FREQUENCY are a pretty good clue... ...and so is the fact that the unit of measurement is Hertz...defind as "the unit of frequency of a periodic phenomenon".
As Casey points out, Bits indicate the resolution of AMPLITUDE, in other words how many ones and zeroes, i.e. how many discrete steps, are available to express the amplitude of the waveform at each point in time that a snapshot is taken.

matto