my thought on sample rates

J

jho1986

New member
There are a lot of people who are always wondering what sample rates to record at (and i was once one of them) and this is my answer... 88.2 KHz. Most likely you are eventually going to put your recorded files onto a CD. Therefore you are going to have to bounce your mix to 44.1 KHz and 16 bit at least if you want to listen to it. So why would anyone ever use 96 KHz. Lets say you took a 100 foot line and placed 960 marbles on that line from start to finish evenly spaced. Now what if I told you to take away 519 marbles while trying to make the remaining 441 marbles stay evenly spaced. That is what has to be done to convert 96K to 44.1K. It is a very messy conversion and unless you have top of the line gear to do that conversion, you simply shouldn't. Now take that same example and instead have 882 marbles in a 100 foot line, and now turn that into 441 marbles. I can do that without even thinking. All you have to do is take every other marble out. Same thing goes for your computer. Your mixes will convert way cleaner if you started at 88.2 and go to 44.1 then if you started at 96 and went to 44.1.

The point of all of this is that if you are going to convert to 44.1 at some point, record at 88.2 KHz.
 
This has been discussed before, and while it may be a great idea if you were doing the calculations yourself, your computer really doesn't mind the extra math. :) More the point, it doesn't necessarily yield a less lossy SRC result. The SRC algorithm isn't going to just throw out every other sample.
 
wrong.
http://www.lavryengineering.com/documents/Sampling_Theory.pdf

jho1986 said:
There are a lot of people who are always wondering what sample rates to record at (and i was once one of them) and this is my answer... 88.2 KHz. Most likely you are eventually going to put your recorded files onto a CD. Therefore you are going to have to bounce your mix to 44.1 KHz and 16 bit at least if you want to listen to it. So why would anyone ever use 96 KHz. Lets say you took a 100 foot line and placed 960 marbles on that line from start to finish evenly spaced. Now what if I told you to take away 519 marbles while trying to make the remaining 441 marbles stay evenly spaced. That is what has to be done to convert 96K to 44.1K. It is a very messy conversion and unless you have top of the line gear to do that conversion, you simply shouldn't. Now take that same example and instead have 882 marbles in a 100 foot line, and now turn that into 441 marbles. I can do that without even thinking. All you have to do is take every other marble out. Same thing goes for your computer. Your mixes will convert way cleaner if you started at 88.2 and go to 44.1 then if you started at 96 and went to 44.1.

The point of all of this is that if you are going to convert to 44.1 at some point, record at 88.2 KHz.
Click to expand...
 
There is no value whatsoever in recording at sample rates higher than your final product. For pert near all of us, that's 44.1.
 
By far, the greatest benefits of a high sample rate are when you originally encode at that rate. Up sampling from a lower to higher rate will not create anymore information than was encoded at the original rate. It will however, spread that same originally encoded data out into more discrete steps, which can be very usefull in applying dynamic processes which use look ahead in there side chain detector.

You math is not applicable to whats going on inside a floating point processor. todays algorithums ( like Rbrain pro from voxegno)can convert those rates with great accuracy.

How do I know this , you ask? Because I slept at a holiday inn last night!!!!! :p :p
 
flatfinger said:
It will however, spread that same originally encoded data out into more discrete steps, which can be very usefull in applying dynamic processes which use look ahead in there side chain detector.
Click to expand...
WTF?? Sample rate has nothing to do with dynamics........
 
While I track at 44.1 mostly because it's less of a drain on my computer and it sounds pretty darn good to me, I could easily see an advantage to tracking at the highest rate possible. Even though I'm editing in the box, I don't mix there. I use an analog console to mix through and the wider frequency response of the higher sampling rate would be cool. And god only knows what will be "Red Book" in 5 years.
 
Track Rat said:
And god only knows what will be "Red Book" in 5 years.
Click to expand...


"Broadband quality" music on your phone, that's what. CDs won't even be around.

(i don't really believe that, I wish everyone would jump on DVD-A or SACD, but it doesn't look like that's gonna happen anytime soon, sigh......)
 
jho1986 said:
There are a lot of people who are always wondering what sample rates to record at (and i was once one of them) and this is my answer... 88.2 KHz. Most likely you are eventually going to put your recorded files onto a CD. Therefore you are going to have to bounce your mix to 44.1 KHz and 16 bit at least if you want to listen to it. So why would anyone ever use 96 KHz. Lets say you took a 100 foot line and placed 960 marbles on that line from start to finish evenly spaced. Now what if I told you to take away 519 marbles while trying to make the remaining 441 marbles stay evenly spaced. That is what has to be done to convert 96K to 44.1K. It is a very messy conversion and unless you have top of the line gear to do that conversion, you simply shouldn't. Now take that same example and instead have 882 marbles in a 100 foot line, and now turn that into 441 marbles. I can do that without even thinking. All you have to do is take every other marble out. Same thing goes for your computer. Your mixes will convert way cleaner if you started at 88.2 and go to 44.1 then if you started at 96 and went to 44.1.

The point of all of this is that if you are going to convert to 44.1 at some point, record at 88.2 KHz.
Click to expand...
Is this why you're not an engineer til you lose all your marbles?
 
My prediction... In the future people will realize that digital is inherently limited in it's scope and switch back to Vinyl. But I might just be crazy.
 
nah, this whole aural thing is just a fad.
I think we'll start to see the disappearance of music and sound all together fairly soon. Soon evolution will just take our ears away and replace them with another set of eyes.

Yeah, yeah, that's what's gonna happen.
 
Well the smithsonian is putting all their recordings on Shellac because they're concerned about the life of CDs. As one engineer once said to me, "vinyl wants to return to it's natural form" I like my records though.
 
RAK said:
Well the smithsonian is putting all their recordings on Shellac because they're concerned about the life of CDs. As one engineer once said to me, "vinyl wants to return to it's natural form" I like my records though.
Click to expand...

I got fooled by that April Fool's story on NPR, too. It was brilliantly executed.
 
apl said:
I got fooled by that April Fool's story on NPR, too. It was brilliantly executed.
Click to expand...

The funny part when I was writing that I was thinking, "well maybe it was the Library of Congress, not the Smithsonian." Turns out I was right about that part, the article was about the Library of Congress, however...

I never knew it was a hoax, someone just told me they were doing that (must have been back in '03) and I had no idea it wasn't true, because I never looked into it. Now that I'm reading the article, it's pretty rediculous. haha, thanks for pointing that out to me. That'll teach you to listen to public radio.

Nice job, All Things Considered.
 
Track Rat said:
I never miss that or the Car Guys or Prairie Home Companion.
Click to expand...

Did you see the movie? I really enjoyed that, Might have to get up to MN for a tapeing some time, that would be fun.
 
You must log in or register to reply here.
Back
Top