12 bit v. 16 bit setting at DVinfo.net

Sam Shore · November 9th, 2005, 04:20 PM

If someone could provide me any links to discussion related to these questions, I'd really appreciate it. Searching doesn't work on my computer.

I don't know if this is general to many cameras, or specific to the Sony VX2100; if it's specific, let me know, I'll post in that forum.

I'm trying to get a basic understanding of audio. In which cases should I record audio on the 12-bit setting, in which cases 16-bit?

Why is the factory setting 12-bit?

My manual says 12-bit records two stereo sounds, 16-bit records one stereo sound with high quality. Huh?

Is it a bad idea to change audio settings on one tape? on one project?

Any other basic info would be greatly appreciated.

David Ennis · November 9th, 2005, 04:27 PM

If you record with 12 bits, it leaves enough room on the tape to dub on another stereo track. But who wants to do that? No one I've come across so far (some of us would be tempted if you could record 4 channels at at time, but that's not how it works). Most people here will tell you record in 16 bit and forget about 12 bit. That's been my policy.

Glenn Chan · November 9th, 2005, 06:02 PM

I would try to stay 16bit... it makes your life easier.

A few particular cameras can record 4 channels of sound at once (i.e. XL1) at reduced quality. 12bit instead of 16bit depth, 32khz sampling instead of 48khz.

2- On many cameras, you can later add narration onto those two other tracks. But since everyone has an editing system nowadays, that's pretty useless.

Adam Keen · November 9th, 2005, 06:06 PM

If you're doing linear editing and dubbing over, you'd want to use 12 bit.

When mixing audio on a computer, go with the 16 bit.

Can you really hear a difference between the two?

George Ellis · November 9th, 2005, 09:26 PM

I have wondered lately if using 12-bit would reduce audio dropouts. On about 6 tapes (out of 18 - mixed cameras), I had a single frame audio dropout. Annoying to fix, but OK compared to tales I hear. But, I did wonder if 12-bit would have fewer chances.

Ty Ford · November 10th, 2005, 06:01 AM

Quote:

Originally Posted by George Ellis

I have wondered lately if using 12-bit would reduce audio dropouts. On about 6 tapes (out of 18 - mixed cameras), I had a single frame audio dropout. Annoying to fix, but OK compared to tales I hear. But, I did wonder if 12-bit would have fewer chances.

No and the frequency response and noise figures for 32 khz 12-bit audio are not as good as 44.1 kHz 16-bit.

Although if you were feeding the same mono audio to all four channels you might not have drop out on a four at the same time.

Try having your heads and trsansport cleaned and/or using better tape.

Ty Ford

George Ellis · November 10th, 2005, 07:08 AM

Thanks Ty. And it was cheaper tape, and a rental that needed service (in-cam battery dead = lost settings (I did not want DVCAM), noisy tape drive (the mic was 100' away).)

Chris Luker · November 10th, 2005, 11:23 AM

Bits are the numbers used to describe the amplitude of the audio waveform or how high it is. kHz is the frequency of those numbers (or snapshots of the waveform) or how many numbers taken per second. The more bits and kHz the better. Cd audio is 16 bit 44.1kHz and is the minimum if you are doing anything other than straight voice over. You can get away with 12 bit 32kHz for voice only, but why not go 16 bit 44.1kHz?
Also, if you need to edit your sound on a computer, every fade, compressor, reverb etc. fx you put on the audio will cause a calculation to be done to the file. The math needed is complex and if you don't have a big enough number to hold the results, the computer will round it off.
I do all of my audio recording in my studio at 24 bits and 44.1kHz (48kHz for video post) so that all of the effects I put on any file will have a big number (24 bits worth) to write it's calculations to. Then at the end, I can dither down (a whole 'nuther topic) to 16 bits for a nice CD ready track.
But, eh, ya... go with 16 bit, it's more better!

Noah Hayes · November 10th, 2005, 12:10 PM

Are there any cameras that record in higher than 16-bit 44.1KHz? I haven't seen any, but maybe I'm just not looking in the right place. For audio recording voice/ambient noise for video, is 24-bit 48KHz significatly better than 16-bit 44.1Khz?

Ty Ford · November 10th, 2005, 12:40 PM

Quote:

Originally Posted by Noah Hayes

Are there any cameras that record in higher than 16-bit 44.1KHz? I haven't seen any, but maybe I'm just not looking in the right place. For audio recording voice/ambient noise for video, is 24-bit 48KHz significatly better than 16-bit 44.1Khz?

As a specification yes, it's better. If it's not implemented right, then maybe not.

Ty

Stephen Finton · November 10th, 2005, 06:15 PM

You guys keep on saying 16bit 44.1kHz. It's 16bit 48kHz.

Not being a cybersnob, just worried you might end up with a nonstandard DV video.

A. J. deLange · November 10th, 2005, 10:14 PM

The "NPR" (Noise Power Ratio i.e. the ratio of the power of the maximum signal which will not overload the converter to the quantizing noise power) of an A/D converter is approximately 6*(n-2) + 3.6 dB. Thus for a 12 bit converter you have about 63.6 dB NPR while for 16 it is 87.6 and for 24 a whopping 135.6. Note that while a 12 bit A/D has very nearly 12 effective bits a 16 bit device may only deliver 15 effective bits and a 24 bit as many as a couple of bits fewer than 24 but still the principal that more bits yields more dynamic range follows. Higher dynamic range means less work for the ALC to do and less liklihood that quantizing noise will be heard. Higher sampling rates mean that the hardware antialiasing filter which precedes the A/D converter can be pretty sloppy because it can be followed by a very sharp digital antialiasing filter after the A/D.

November 9th, 2005, 04:20 PM	#1
Sam Shore Regular Crew Join Date: Oct 2005 Posts: 58	12 bit v. 16 bit setting If someone could provide me any links to discussion related to these questions, I'd really appreciate it. Searching doesn't work on my computer. I don't know if this is general to many cameras, or specific to the Sony VX2100; if it's specific, let me know, I'll post in that forum. I'm trying to get a basic understanding of audio. In which cases should I record audio on the 12-bit setting, in which cases 16-bit? Why is the factory setting 12-bit? My manual says 12-bit records two stereo sounds, 16-bit records one stereo sound with high quality. Huh? Is it a bad idea to change audio settings on one tape? on one project? Any other basic info would be greatly appreciated.

November 9th, 2005, 04:27 PM	#2
David Ennis Fred Retread Join Date: Jul 2004 Location: Hartford, CT Posts: 1,227	If you record with 12 bits, it leaves enough room on the tape to dub on another stereo track. But who wants to do that? No one I've come across so far (some of us would be tempted if you could record 4 channels at at time, but that's not how it works). Most people here will tell you record in 16 bit and forget about 12 bit. That's been my policy. __________________ "Nothing in the world can take the place of persistence..." - Calvin Coolidge "My brain is wired to want to know how other things are wired." - Me

November 9th, 2005, 06:02 PM	#3
Glenn Chan Inner Circle Join Date: Jun 2003 Location: Toronto, Canada Posts: 4,750	I would try to stay 16bit... it makes your life easier. A few particular cameras can record 4 channels of sound at once (i.e. XL1) at reduced quality. 12bit instead of 16bit depth, 32khz sampling instead of 48khz. 2- On many cameras, you can later add narration onto those two other tracks. But since everyone has an editing system nowadays, that's pretty useless.

November 9th, 2005, 06:06 PM	#4
Adam Keen Regular Crew Join Date: Nov 2005 Posts: 49	If you're doing linear editing and dubbing over, you'd want to use 12 bit. When mixing audio on a computer, go with the 16 bit. Can you really hear a difference between the two?

November 9th, 2005, 09:26 PM	#5
George Ellis Trustee Join Date: Feb 2004 Location: Suwanee, GA Posts: 1,241	I have wondered lately if using 12-bit would reduce audio dropouts. On about 6 tapes (out of 18 - mixed cameras), I had a single frame audio dropout. Annoying to fix, but OK compared to tales I hear. But, I did wonder if 12-bit would have fewer chances.

November 10th, 2005, 07:08 AM	#7
George Ellis Trustee Join Date: Feb 2004 Location: Suwanee, GA Posts: 1,241	Thanks Ty. And it was cheaper tape, and a rental that needed service (in-cam battery dead = lost settings (I did not want DVCAM), noisy tape drive (the mic was 100' away).)

November 10th, 2005, 11:23 AM	#8
Chris Luker Major Player Join Date: Oct 2005 Location: Woods Cross, Utah Posts: 310	Bits are the numbers used to describe the amplitude of the audio waveform or how high it is. kHz is the frequency of those numbers (or snapshots of the waveform) or how many numbers taken per second. The more bits and kHz the better. Cd audio is 16 bit 44.1kHz and is the minimum if you are doing anything other than straight voice over. You can get away with 12 bit 32kHz for voice only, but why not go 16 bit 44.1kHz? Also, if you need to edit your sound on a computer, every fade, compressor, reverb etc. fx you put on the audio will cause a calculation to be done to the file. The math needed is complex and if you don't have a big enough number to hold the results, the computer will round it off. I do all of my audio recording in my studio at 24 bits and 44.1kHz (48kHz for video post) so that all of the effects I put on any file will have a big number (24 bits worth) to write it's calculations to. Then at the end, I can dither down (a whole 'nuther topic) to 16 bits for a nice CD ready track. But, eh, ya... go with 16 bit, it's more better!

November 10th, 2005, 12:10 PM	#9
Noah Hayes Major Player Join Date: Jun 2005 Location: Sterling, Virginia Posts: 226	Are there any cameras that record in higher than 16-bit 44.1KHz? I haven't seen any, but maybe I'm just not looking in the right place. For audio recording voice/ambient noise for video, is 24-bit 48KHz significatly better than 16-bit 44.1Khz?

November 10th, 2005, 06:15 PM	#11
Stephen Finton Major Player Join Date: Jun 2005 Location: Houston, TX Posts: 344	You guys keep on saying 16bit 44.1kHz. It's 16bit 48kHz. Not being a cybersnob, just worried you might end up with a nonstandard DV video.

November 10th, 2005, 10:14 PM	#12
A. J. deLange Major Player Join Date: Jun 2004 Location: McLean, VA United States Posts: 749	The "NPR" (Noise Power Ratio i.e. the ratio of the power of the maximum signal which will not overload the converter to the quantizing noise power) of an A/D converter is approximately 6*(n-2) + 3.6 dB. Thus for a 12 bit converter you have about 63.6 dB NPR while for 16 it is 87.6 and for 24 a whopping 135.6. Note that while a 12 bit A/D has very nearly 12 effective bits a 16 bit device may only deliver 15 effective bits and a 24 bit as many as a couple of bits fewer than 24 but still the principal that more bits yields more dynamic range follows. Higher dynamic range means less work for the ALC to do and less liklihood that quantizing noise will be heard. Higher sampling rates mean that the hardware antialiasing filter which precedes the A/D converter can be pretty sloppy because it can be followed by a very sharp digital antialiasing filter after the A/D.