44100 at 16 bit limit?

This section is now closed.
Forum rules
Audacity 1.3.x is now obsolete. Please use the current Audacity 2.1.x version.

The final version of Audacity for Windows 98/ME is the legacy 2.0.0 version.
Locked
polkat
Posts: 54
Joined: Fri May 20, 2011 9:00 pm
Operating System: Please select

44100 at 16 bit limit?

Post by polkat » Wed Dec 21, 2011 1:15 am

This might be a silly question (Lord knows I've asked them before), but if CD's are all recorded at 44100 baud at 16 bits (as I understand it), and your intention in using Audacity is to eventually burn CD's, what is the point of setting Audacity record rate higher...say to something like 48000 at 24? Is it just wasting space, or is the idea to get the highest quality possible before burning to CD?

kozikowski
Forum Staff
Posts: 69367
Joined: Thu Aug 02, 2007 5:57 pm
Operating System: macOS 10.13 High Sierra

Re: 44100 at 16 bit limit?

Post by kozikowski » Wed Dec 21, 2011 3:23 am

The Audacity default of 44100, 32-bit floating is not a bad way to start the process because nobody can hear anything up beyond 18KHz anyway, and the extra bits let you (within reason) overload and do complex editing, filtering and processing with no significant sound damage.

For just one example, if you apply a filter that accidentally increases the volume of a recording beyond "zero" creating clipping overload damage, you better hope you're in 32-floating and not 16-bit. 16-bit will create permanent distortion and harsh clipping damage. 32-floating just creates an opportunity to reduce the volume of the show and keep cranking with little or no permanent damage.

Koz

polkat
Posts: 54
Joined: Fri May 20, 2011 9:00 pm
Operating System: Please select

Re: 44100 at 16 bit limit?

Post by polkat » Wed Dec 21, 2011 4:34 am

Okay, I understand that...sort of. So you are suggesting that 44100 at 32 bit float is a good overall setting? When I installed 1.3.14 the default was 44100 at 16 bit, not 32 float, but I'll give 32 bit a try.

kozikowski
Forum Staff
Posts: 69367
Joined: Thu Aug 02, 2007 5:57 pm
Operating System: macOS 10.13 High Sierra

Re: 44100 at 16 bit limit?

Post by kozikowski » Wed Dec 21, 2011 6:48 am

There is a tiny bit of damage when Audacity converts from 32-float to 16-bit when you export for the music CD. So if you know you will never be filtering or adding significant effects, 16-bit works fine. 16-bit is good for something like 96dB between the noise floor and overload.

That's 20dB headroom for loud sounds plus 76dB noise floor. Smashing Loud and Dead Quiet. I bet none of your analog equipment can do that.


If you installed 1.3.14 after already installing earlier versions, it would have "assumed" the settings of the older Audacity. If you return Audacity preferences to First Birthday, it should go up to 32-float.

http://manual.audacityteam.org/man/Preferences

Koz

DVDdoug
Forum Crew
Posts: 9419
Joined: Fri Sep 10, 2010 11:30 pm
Operating System: Windows 10

Re: 44100 at 16 bit limit?

Post by DVDdoug » Wed Dec 21, 2011 6:40 pm

Almost all audio editors work in floating-point (32-bits or 64-bits) internally, even though you may be importing/exporting 16-bit (or 24-bit) integer files . (Except, I think Pro Tools was integer based until recently.)

But, there are no floating-point ADCs (analog-to-digital converters) or DACs (digital-to-analog converters). So, if you record to a floating-point format, or to a bit-depth higher than your ADC, the driver/software is making a conversion and you are just wasting space. Similarly, if you play-back a floating-point format, or higher bit-depth than your DAC, the driver will make a conversion.

In addition to what Koz mentioned about signal levels, DSP (Digital Signal Processing) is almost always done in floating-point. There is lots of math involved in DSP, and when you add two or more 16-bit integers together, you need more than 16-bits to hold the result. If you divide, you might need to handle fractions.
There is a tiny bit of damage when Audacity converts from 32-float to 16-bit when you export for the music CD.
I believe it's mathematically possible to convert from 16-bit integer to 32-bit floating-point, and back to 16-bits and you get back the exact-original 16-bit number. As I understand it, Audacity won't change anything (except for you editing processing) unless you are dithering. And, the "damage" from dithering is intentional and when used correctly it's supposed to be beneficial (assuming you can even hear dithering :P ).

Of course, the opposite is not always true. If you convert from 32-bit floating-point to 16-bit integer, and back to 32-bit, you may get distortion or loss of resolution.

steve
Site Admin
Posts: 81627
Joined: Sat Dec 01, 2007 11:43 am
Operating System: Linux *buntu

Re: 44100 at 16 bit limit?

Post by steve » Wed Dec 21, 2011 11:59 pm

polkat wrote:Okay, I understand that...sort of. So you are suggesting that 44100 at 32 bit float is a good overall setting? When I installed 1.3.14 the default was 44100 at 16 bit, not 32 float, but I'll give 32 bit a try.
Yes, 44100 at 32 bit is a good setting overall setting. It is the default for Audacity 1.3.x, but Audacity will "inherit" many settings (including bit rate and bit format) from previous versions of Audacity that have been installed on the machine.
9/10 questions are answered in the FREQUENTLY ASKED QUESTIONS (FAQ)

polkat
Posts: 54
Joined: Fri May 20, 2011 9:00 pm
Operating System: Please select

Re: 44100 at 16 bit limit?

Post by polkat » Thu Dec 22, 2011 12:10 am

Yes, I am aware of that, so I deleted the .cfg file in Documents and Settings before I installed 1.3.14 It's retaining 32 bit now, so I guess I'm okay.

Locked