KVR Audio

Stupid American Pig · Post by **Stupid American Pig** » Mon Feb 20, 2006 8:38 pm

I definitely prefer using the 64 bit engine, I think there may be something to the 64 bit math and large numbers of tracks. In my experience, I notice less stray clipped peaks with the DP turned on. I have no proof tho.

bduffy · Post by **bduffy** » Mon Feb 20, 2006 8:56 pm

Stupid American Pig wrote:I definitely prefer using the 64 bit engine, I think there may be something to the 64 bit math and large numbers of tracks. In my experience, I notice less stray clipped peaks with the DP turned on. I have no proof tho.

Well, there really should be something better going on. I just recall, when it came out, that KVR had a mini-flood of posts saying, "OMG, it sounds so much better", and I can't imagine it just sounding better to the naked ear, right away. I could see your mix sounding strong after lots of editing, fx, etc. Might be a bit louder, perhaps?

You know, we have Sonar 5 here at work, I guess I could try it out!

I just recall having a rough time with it, and I don't use it that much anymore. But when I get a sec I'll re-install it and see if I notice anything.

Anyway, I don't want to ruffle any feathers. I dig Sonar, that was just bugging me.

kp · Post by kp » Mon Feb 20, 2006 9:02 pm

I'll love it if this ends up showing that digital "distortion" through truncation is pleasing to the ear in the same way valve "distortion" is. That'll show the audiophiles :-)

bduffy · Post by **bduffy** » Mon Feb 20, 2006 9:07 pm

kp wrote:I'll love it if this ends up showing that digital "distortion" through truncation is pleasing to the ear in the same way valve "distortion" is. That'll show the audiophiles

You know what's funny? I used to record my guitar HOT into Cakewalk Pro Audio 7 all the time, and it sounded great. Don't know what that was...

Ron Kuper [Cakewalk] · Mon Feb 20, 2006 11:42 pm

bduffy wrote:So do you think there is an audible difference in the sound quality, if one just opens an old mix with 64-bit enabled, say?

I have lousy ears, but I'm too bad at math.

What that means is, I would probably be hard pressed to tell myself (though others have claimed 64-bit sounds better). But I can show you some math that demonstrates that processing using 32-bit floats introduces calculation errors, and that these errors go away if you use 64-bit double precision instead.

You can read about it here:
http://www.cakewalk.com/x64/

There's a lot information about x64 and about 64-bit double precision on that page. If you want focus on 64-bit double precision, please read the section about it in the whitepaper:
http://www.cakewalk.com/Press/Cakewalk_ ... ations.pdf

There is source code to demonstrate the math problems discussed in the whitepaper:
http://www.cakewalk.com/Events/AES_2005/FloatTest.zip

runagate · Post by **runagate** » Mon Feb 20, 2006 11:57 pm

These discussions always seem strange to me because it's 1/2 people who know what bitdepth means and 1/2 who don't know but don't realize it.

Bitdepth = number of steps of volume created in digital

Samplerate = number of times that the waveform is sampled per time unit

Just thought I'd contribute that so that this discussion doesn't bog down.
I don't use Sonar so I was just curious.

Stupid American Pig · Post by **Stupid American Pig** » Tue Feb 21, 2006 1:04 am

Ron Kuper [Cakewalk] wrote:
bduffy wrote:So do you think there is an audible difference in the sound quality, if one just opens an old mix with 64-bit enabled, say?
I have lousy ears, but I'm too bad at math. What that means is, I would probably be hard pressed to tell myself (though others have claimed 64-bit sounds better). But I can show you some math that demonstrates that processing using 32-bit floats introduces calculation errors, and that these errors go away if you use 64-bit double precision instead.

You can read about it here:
http://www.cakewalk.com/x64/

There's a lot information about x64 and about 64-bit double precision on that page. If you want focus on 64-bit double precision, please read the section about it in the whitepaper:
http://www.cakewalk.com/Press/Cakewalk_ ... ations.pdf

There is source code to demonstrate the math problems discussed in the whitepaper:
http://www.cakewalk.com/Events/AES_2005/FloatTest.zip

I actually just compiled and ran that project, I noticed that as you increase the # of bits per sample the number of errors calculated increase for the 32 bit calcs. I dont know enough C to really dig, but I would get compilation errors if the Bits per sample was increased to > 32. Unless there is a math error in this app, it would seem that the deeper the bit depth the more important it would be to use the 64 bit mixer?

spaceman · Post by **spaceman** » Tue Feb 21, 2006 10:32 am

has anyone mentioned 20-40% yet?

Ron Kuper [Cakewalk] · Tue Feb 21, 2006 2:06 pm

Stupid American Pig wrote:I actually just compiled and ran that project, I noticed that as you increase the # of bits per sample the number of errors calculated increase for the 32 bit calcs. I dont know enough C to really dig, but I would get compilation errors if the Bits per sample was increased to > 32. Unless there is a math error in this app, it would seem that the deeper the bit depth the more important it would be to use the 64 bit mixer?

That's exactly the point. If you use 24-bits per sample the errors start showing up almost immediately. BTW if you are using floats for processing, than you really can't have sample size greater than 24-bits anyway, because all those bits won't fit into the float's mantissa.

jasonsantiago · Post by **jasonsantiago** » Thu Feb 23, 2006 9:04 pm

Ron Kuper [Cakewalk] wrote: Our customers who can hear a difference between the two engines are using real world songs, where IMO the difference is that 64-bit retains more accuracy with lots of processing.

Hahahahahahahhahah

did you actually do a double blind study where people picked which sounded better?

I doubt it -- I'm guessing people just think 64 bit will sound better, so they claim it does.

Thought Sonar 5's 64-bit engine sounded "better?"