Micing Technique, or bit depth???

[SecondSon]

So I just got a bunch of RAM put in my computer which solved my problem of not doing 16 bit audio. I am now doing 24 bit with the same sampling rate. I have been having trouble with drums, so I tried only four mics. kick, snare, and two overheads spaced about 4 inches from each other above the drummers head facing the toms. We spent alot of time moving these around, but when we finally did a take, it sounded friggin great. What I am curious, is how much a difference 16 to 24 bit is going to make? Did we do a great job micing, or should the bit rate take more credit??? I am sure it is a little of both, i just can't tell because we never used this mic technique when we were in 16 bit.
Haha, <<that dude is buggin me out!!!

[Randy]

One way to find out is to keep the setup the same and switch back to 16 bit and give it a try- preferably in the same session, so you know that the drums didn't go out of tune in the mean time.

My guess is that it is the mic'ing. At least on my setup, the biggest difference between 16 and 24 is that the cymbals sounded a little better.

[cgarges]

There were perfectly good recordings done back in the days of "evil sixteen-bit technology." If you got good sounds, it was because of the sound of the instrument, the player, the mics, their positioning and phase relationships, the barometric pressure in the room, etc. What they drummer had for breakfast probably had something to do with it, too. In any case, 24-bit is a big improvement over 16-bit (to me, at least), but I doubt highly that it was the cause of your ability to get good sounds. That was you, man.

Congratulations!

Chris Garges
Charlotte, NC

[wing]

i can't record in 24-bit because my software (adobe audition, formerly cool edit pro) only lets me select between 16-bit and 32-bit... though my interface allows 24-bit, i've been dealing with 16-bit as i couldn't do 32-bit... so is there anything i can do about this?

[chetatkinsdiet]

you lost me wing....
You might want to look into that. Most interfaces, well, just about all that I know of are 24 bit. The software that says 32 bit, is actually 24 bit fitted into a 32 bit word length. Does this make sense? So, you should be recording in that 32 bit mode on your software and then dithering down to 16 when you are ready to burn to a CDR, or whatever you do when you completely finish a song.
I'm sure there's another post or thread that explains this better than I can. I'm trying to type and play a really loud guitar in a really small room before my wife gets home and makes me turn it down....
later,
m

[SecondSon]

Awesome, I think it was the micing too. Anyway, yeh, go 32, if your computer can't handle it and you have fast processing speed, then you need RAM. I thought I was going to have to buy a faster writing hard drive, but then a gig of RAM solved my prob.

[jajjguy]

I think if your software is recording at a higher bitrate than your interface is giving it, it'll just add on the extra zeros. Meaning your "32 bit" recording will really just have whatever bitrate your interface gave it, in this case 24. In your case, though, it should be easy enough to compare them, just do a short session at 16 bit and another at 32. If you don't hear a difference, then nevermind.

[philbo]

So I just got a bunch of RAM put in my computer which solved my problem of not doing 16 bit audio. I am now doing 24 bit with the same sampling rate. I have been having trouble with drums, so I tried only four mics. kick, snare, and two overheads spaced about 4 inches from each other above the drummers head facing the toms. We spent alot of time moving these around, but when we finally did a take, it sounded friggin great. What I am curious, is how much a difference 16 to 24 bit is going to make? Did we do a great job micing, or should the bit rate take more credit??? I am sure it is a little of both, i just can't tell because we never used this mic technique when we were in 16 bit.
Haha, <<that dude is buggin me out!!!

One thing that 24 bit recording does for you is that you don't have to worry so much about running the signal into the DAW 'hot' (peaks hitting close to 0 dB).

With 24 bit, you can ease off a bit, record with the peaks hitting -12 to -15 dB and still have it sound great, with a lot less danger of clipping.

Phil

[EasyGo]

i can't record in 24-bit because my software (adobe audition, formerly cool edit pro) only lets me select between 16-bit and 32-bit... though my interface allows 24-bit, i've been dealing with 16-bit as i couldn't do 32-bit... so is there anything i can do about this?

(at least in CEP 2.0) when you mix down, there's a 'custom' button (or some such name) on the dialog box, where you can select 16 or 24 bits for the mixdown. My sound card can't do 32 bits either, but I render the mixes as 24 bit.

[ottokbre]

Bit Rate = Dynamic Range
Sample Rate = Frequency Range

In rough terms. This may be obvious to many peeps here, but I didnt know this for the longest time. So if ya don't know, nah ya knows!

[greatmagnet]

I did NOT know that and thank you VERY much! That is the most helpful sentence I've read in about two months.

So basic and yet...

[SecondSon]

Bit Rate = Dynamic Range
Sample Rate = Frequency Range

In rough terms. This may be obvious to many peeps here, but I didnt know this for the longest time. So if ya don't know, nah ya knows!

I didn't know that either. So if my Sampling rate is 41000, then I think it would follow that two samples would be needed for each complete wave, so then 41000/2 gives me 20500hz as the maximum frequency I can record. Does that sound right to anyone???? None the less, what is so special about recording at 96000 sampling rate anyway, I have heard a lot of people say it doesn't really make a difference.

[wing]

i can't record in 24-bit because my software (adobe audition, formerly cool edit pro) only lets me select between 16-bit and 32-bit... though my interface allows 24-bit, i've been dealing with 16-bit as i couldn't do 32-bit... so is there anything i can do about this?

(at least in CEP 2.0) when you mix down, there's a 'custom' button (or some such name) on the dialog box, where you can select 16 or 24 bits for the mixdown. My sound card can't do 32 bits either, but I render the mixes as 24 bit.

yea, but what do you set it to record in as? 32-bit?

[EasyGo]

yea, but what do you set it to record in as? 32-bit?

I set it for 32 bits, but it's really working at 24 bits. It confused me too, cuz it never gives you 24 bits as an option when you start a new file, just 16 or 32.

[wayne kerr]

I didn't know that either. So if my Sampling rate is 41000, then I think it would follow that two samples would be needed for each complete wave, so then 41000/2 gives me 20500hz as the maximum frequency I can record. Does that sound right to anyone???? None the less, what is so special about recording at 96000 sampling rate anyway, I have heard a lot of people say it doesn't really make a difference.

It's called the Nyquist Theorem. It states that the highest frequency you can sample will be half of the sample rate. 44.1 = 22.5K, 48 = 24K, 88.2 = 44.1K, 96 = 48K, 192 = 96K. You are correct, that it takes that many samples to account for both the positive and the negative going portion of the waveform. Why do you want more samples if the human ear can not hear above 20K? (for most of us, it's more like 16K and for our parents, it's closer to 10K!)

One word: Harmonics

Harmonics are whole number multiples of the fundamental and determine the timbre of the instrument. In fact, here is where bit depth becomes so important. Not only does bit depth (or word length) determine the overall dynamic range of your system, but also the resolution of your system's ability to reproduce relative amplitude. This sounds like it's getting complicated, but it's not.

Let's say you play middle-C on a piano (261.62Hz) and on a guitar. You are playing the exact same note, but the two sound radically different.
Why? Harmonics, of course. For the sake of math, let's say our C is 260Hz.

The second harmonic is the next whole number multiple of the fundamental:

F = 260
H2 = 520 (octave)
H3 = 780
H4 = 1.04 (octave)
H5 = 1.3
H6 = 1.56
H7 = 1.82
H8 = 2.08 (octave)
H9 = 2.34
H10 = 2.6
H11 = 2.86
H12 = 3.12
H13 = 3.38
H14 = 3.64
H15 = 3.9
H16 = 4.16 (octave)
H17 = 4.42
H18 = 4.68
H19 = 4.94
H20 = 5.2


and so on...

So, what's the big deal? Well, each time the hammer strikes the strings on the piano, it produces all those frequencies - the fundamental, which is the note you played and all the harmonics (also called partials and sometimes overtones). The harmonics all play at various relative amplitudes (comparitive volume levels). That is why your guitar sounds different than your piano - differing relative amplitude of the harmonics. Well, guess what? Even though 20K is the limit of human hearing, sound in nature can go on forever. We have sophisticated oscilliscopes than can measure oscillations well into the GHz! Even though you can't actually hear 52.32K (which also happens to be the 200th harmonic and the 20th octave of middle C!), the sound wave still enters your ear and is processed by your brain - it gives clues to help the brain interpret the sound and distinguish between a violin, a viola, and the clacking of a computer keyboatrd.

Most analog recorders are capable of reproducing a continuously variable analog waveform of up to about 80K - some as high as 100K. Ever see theat spec called THD? It stands for total harmonic distortion and measures how accurately these frequencies are reproduced - the lower the number, the better the distortion and the more natural the recording sounds. This is a radical simplificaiton, but I think you get the point.

So, at 44.1, your highest frequency sample is a paltry 22.05K. Ever wonder why CDs sound so shitty? 48 is scantly better. As processors got better (faster) and hard drives got bigger, engineers realized they could do better sample rates - the cost per bit had come down to the point where consumers could afford to do 96/24! So, a sample rate of 96K will get you to a little over half of the frequency response of most analog decks, and a 192K sample rate puts you squarely in the ballpark with a frequency response of 96K.

But remember earlier I said bit depth was equally important? Some engineers actually think it's more important. Not only is the frequency of the harmonic important, but so is its relative amplitude. In an analog system, you may have a lower overall dynamic range (60 -70 dB w/o noise reduction vs about 140 dB at 24 bits), but you've also got infinitely variable amplitude, whereas with digital, you don't - you have finite voltage steps which determine the amplitude of each sample.

The formula is 2n - at 16 bits, you have 2(16) or 65,536 discrete voltage steps and at 24 bits, you have 2(24) or 16,777,216 discrete voltage steps between total silence and clipping. Or, at 16 bits, your amplitude resolution is 0.0000873779296875V and at 24 bits, it is 0.0000000731945037841797V! That is a measurement of hoe small a voltage difference the two systems can distunguish.

Those extra 8 bits really add up! So, 24 bit audio begins to approach analog's TRILLIONS of relative amplitudes and more accurately reproduces both the frequency of the all important upper harmonics, AND the amplitude of those notes relative to one another, thus making your french horn sound different than your xylophone and your snare drum different than Sammy Hagar's falsetto.

In the end, it all comes down to what you hope to accomplish. I still record on tape because I like the sound. I know why I like the sound too. It seems more natural to me, and I love what I can do with tape compression. Tape seems to respond more like nature - with soft boundaries and asymmetry as opposed to finite variables and rigid order. But then I can't take 15 different vocal takes and comp them in five minutes on tape either, so like I said, It's all about what you want to acomplish.



CC
PP