24 bit vs. 16 bit when it comes to effects -Sound Forge - PC

Subject:24 bit vs. 16 bit when it comes to effects Posted by: pup Date:2/27/2002 9:57:27 AM Hello, all. I was wondering if I'd get better results adding Sound Forge effects to 24-bit files rather than 16-bit files. Like using Time Stretch, for example...would it sound better if I had recorded the files in 24 bit first? What's the connection behind bit rate and post effects? Thanks. - pup
Subject:RE: 24 bit vs. 16 bit when it comes to effects Reply by: Iacobus Date:2/27/2002 1:43:46 PM Using 24-bit gives more processing headroom to work with. Using 24-bit effects even on a 16-bit file makes the effect that much more faithful and accurate. You'd especially notice the difference on acoustic recordings. It's subtle to some people, but it's there. I say always go for 24-bit all the way if you can. (I'm no expert. I just sound like one. :) ) Iacobus
Subject:RE: 24 bit vs. 16 bit when it comes to effects Reply by: rjt Date:2/27/2002 8:42:00 PM I agree with using 24 bits.... it also makes a hugh difference when you process a file several times.... compression, eq, chorus, reverb.... then you will really hear it.
Subject:RE: 24 bit vs. 16 bit when it comes to effects Reply by: dbarry Date:2/28/2002 8:40:30 PM The world is moving to 24 bits for a good reason. The old 16 bit is really pretty close to unacceptable. As soon as you manipulate the data you get degradation. The way is works is this: Digitizing means converting the analog signal to a number. 16 bits really means the signal is digitized to 15 bits, plus a sign bit for polarity. This means that the smallest voltage measured is given the number 1. The largest voltage measured is given the number 32767 (or 2 raised to the 15th power). So the dynamic range for "16" bit conversion is only 10log(32767/1) = 45 db Volts. So much for your 100 db preamps! If you add 2 signals that are over half of the 32767, then you must scale them back to avoid producting a number that will not fit into the 15 bits plus sign. When you go to 24 bits, the largest number is 2^23 or 8,388,607 plus sign, about 70 db (voltage). So you can add a lot of smaller numbers together without messing up the precision. At the end, sadly, you must convert back to old 16 bit to make CDs but at least you did not have rounding errors during all those cool effects that you did. The 24 bit thing is very real, but the effect of 96 KHz sampling is much less important.

Subject:24 bit vs. 16 bit when it comes to effects

Posted by: pup

Date:2/27/2002 9:57:27 AM

Hello, all. I was wondering if I'd get better results adding Sound Forge effects to 24-bit files rather than 16-bit files. Like using Time Stretch, for example...would it sound better if I had recorded the files in 24 bit first? What's the connection behind bit rate and post effects? Thanks.

- pup

Subject:RE: 24 bit vs. 16 bit when it comes to effects

Reply by: Iacobus

Date:2/27/2002 1:43:46 PM

Using 24-bit gives more processing headroom to work with. Using 24-bit effects even on a 16-bit file makes the effect that much more faithful and accurate. You'd especially notice the difference on acoustic recordings. It's subtle to some people, but it's there.

I say always go for 24-bit all the way if you can.

(I'm no expert. I just sound like one. :) )

Iacobus

Subject:RE: 24 bit vs. 16 bit when it comes to effects

Reply by: rjt

Date:2/27/2002 8:42:00 PM

I agree with using 24 bits.... it also makes a hugh difference when you process a file several times.... compression, eq, chorus, reverb.... then you will really hear it.

Subject:RE: 24 bit vs. 16 bit when it comes to effects

Reply by: dbarry

Date:2/28/2002 8:40:30 PM

The world is moving to 24 bits for a good reason. The old 16 bit is really pretty close to unacceptable. As soon as you manipulate the data you get degradation. The way is works is this: Digitizing means converting the analog signal to a number. 16 bits really means the signal is digitized to 15 bits, plus a sign bit for polarity. This means that the smallest voltage measured is given the number 1. The largest voltage measured is given the number 32767 (or 2 raised to the 15th power). So the dynamic range for "16" bit conversion is only 10log(32767/1) = 45 db Volts. So much for your 100 db preamps! If you add 2 signals that are over half of the 32767, then you must scale them back to avoid producting a number that will not fit into the 15 bits plus sign.
When you go to 24 bits, the largest number is 2^23 or 8,388,607 plus sign, about 70 db (voltage). So you can add a lot of smaller numbers together without messing up the precision. At the end, sadly, you must convert back to old 16 bit to make CDs but at least you did not have rounding errors during all those cool effects that you did.
The 24 bit thing is very real, but the effect of 96 KHz sampling is much less important.