Hi guys. Going through the info within the Kingston drums download, I picked up on Voxengo's 'R8 brain' for the first time. Apparently this gives very high quality results with wav conversions. So, I wondered if anyone else is using this, and, the main question:- I use VST 5.1, currently dither using the inhost VST plug, it's now occurred to me that maybe I could just save my finished song as a WAV, then dither through the R8 Brain. Will this be a better option OR .... would this just be adding yet another degrading conversion to the finished wav? ? Anyone else doing it?

r8brain is for samplerate conversions not dithering, although it'll do this as it's final step if you selected anything below 32bits.

Unless you're converting your song from Cubase to higher or lower samplerates, there's no use for it.

For example, your song in cubase is exported as 48khz 24bit, and you want to convert to 44.1khz 16bit.

The dithering in r8brain is simple gaussian dither. Cubase can do this, too. (if memory serves me correctly).

Could somebody post that Ozone dithering link again?

Very good info there.

(edit)

Nevermind, here it is:
http://www.izotope.com/products/audio/ozone/guides.html

Kingston wrote:

R8brain is for samplerate conversions not dithering, although it'll do this as it's final step if you selected anything below 32bits. Unless you're converting your song from Cubase to higher or lower samplerates, there's no use for it. For example, your song in cubase is exported as 48khz 24bit, and you want to convert to 44.1khz 16bit.

Yes, exactly Kingston. I record at 48/24 and currently use the host to dither. Am now thinking of turning OFF the host dither plug, and using R8brain instead. Just wondered if anyone thought this a better quality solution?

Well you pretty much have to dither when going from the hosts internal 32bit float to the 24bit format.

I'm guessing you want to export the file for mastering and to maintain as high quality as possible.

In that case, just simple dither will do, turn off the noise shaping. Use noise shaping only as the final step going to 16bit before burning.


Here is what I did yesterday for example:

1. Export song from logic at 48khz 24bit (simple dither plug at the output, no noise shaping)

2. Convert 48khz 24bit to 96khz 24bit for mastering. Even in mastering, still no noise shaping.

3. Final convert to 44.1 24bit, then a high quality noise shaped dither to 16bits. (for example the MBIT+ one in Ozone).

4. BURN

The cleanest signal path I can think of! Works for a hardcore goldenear purist like me.

Kingston wrote:

Well you pretty much have to dither when going from the hosts internal 32bit float to the 24bit format. I'm guessing you want to export the file for mastering and to maintain as high quality as possible. In that case, just simple dither will do, turn off the noise shaping. Use noise shaping only as the final step going to 16bit before burning. Here is what I did yesterday for example: 1. Export song from logic at 48khz 24bit (simple dither plug at the output, no noise shaping) 2. Convert 48khz 24bit to 96khz 24bit for mastering. Even in mastering, still no noise shaping. 3. Final convert to 44.1 24bit, then a high quality noise shaped dither to 16bits. (for example the MBIT+ one in Ozone). 4. BURN The cleanest signal path I can think of! Works for a hardcore goldenear purist like me.

hehe
All I do is to export from within cubase at 88.2khz 32 bit and then I use r8brain in order to convert it to 44.1khz 16bit.

Quote:

All I do is to export from within cubase at 88.2khz 32 bit and then I use r8brain in order to convert it to 44.1khz 16bit.

suppose I'd do that, too. But alas! This box is too slow!

Then there's the mastering to take into account as well.

Funny thing - I just tried doing a sample convert operation on Kingston's Easy Rider kit using R8Brain.

I backed up the original 48kHz 24 bit sample set and then converted the first version to 44.1 at max quality using R8Brain 1.7's handy little batch converter.

I did a bounce with the sample rate at 44.1kHz of a 14 bar midi file through the EXS24 with Easyrider loaded - first with the original 48kHz samples which had to be sample rate shifted by EXS24 and then again with the 44.1kHz files I made using R8Brain. Audibly better with the 44.1 set - things like a ringing ride bell sample sound better - general bashing around on kick and snare is not so obviously different.

One thing is for sure - R8Brain at max quality will eat the EXS24's sample conversion for lunch.

The only issue you will find in switching is that you will lose any sample offsets when you switch the sample files.

Eg

Quote:

One thing is for sure - R8Brain at max quality will eat the EXS24's sample conversion for lunch.

Why do you think I suggested using it in the manual?

It'll eat any other samplers' conversion, too, by the way (and spit on them just to make the point clear).

It works the other way, too. Convert them to higher samplerate and expect subtler, but still positive results.