Getting rid of distortion in my mix?

Mistral

RyosaMusic
Here is a short clip of a song I'm working on in Cubase.. It is a kind of doomy rock metal song, and I'm going for a pretty "in your face" mix. It is far from complete - cuts to just one guitar at the end.. and no vocals yet. This clip does not have any mastering applied but is a representation of the material I'm working with. I mixed it down from a 32 bit session to 16 bit using Cubase's UV22HR dithering, then I believe I applied some limiting afterwards before converting it to mp3. I realize now that this was a mistake (dither should always be done LAST.)



I am working largely with samples (drums, bass) as well as vsti synths, and guitars are recorded clean and processed real-time with NI Guitar Rig 2. I have noticed that when I apply compression to my final mix or loudness maximizing that the mix starts to have some unpleasant distortion to it. I have been looking into ways to eliminate this problem, and suspected it is mostly a result of the way the guitars were processed, so I worked on their tone as much as possible.. and now I'm looking into other avenues, and this is why my attention has come to 32 bit mixing and use of dither.

I originally started my session as 16 bit and I just need to have this confirmed for me.. Is there any real benefit to raising it to 32-bit float, mixing down as a 32 bit file, mastering it then applying dithering as the very last step - as opposed to keeping it 16 bit, mixing down and mastering? Everywhere I read about dither, it says that it eliminates distortion and replaces it with hiss.. But does this actually apply as much to loud recordings as it does to low level passages?


A couple of side questions: A lot of plugins seem to have this "force 32-bit samples" checkbox. When is this best applied? When mixing in 32 bit, to make sure it is really processing at that depth as well?

Also: I use Izotope Ozone 3.. Anyone who is familiar with it, as well as MegabitMax Ultra dithering.. is it the same thing as Ozone's Mbit+ Ultra dithering, or did izotope just try to cash in on the name.. being that megabit max has won a lot of "sound tests" by a huge margin..

I know I'm all over the board here, but anyway just listen to the mix and see if you can spot anything you think might help me achieve a cleaner mix/master without sacrificing the heaviness.. Remember I use Guitar Rig 2 for the guitars, (processed in hi-res mode). But I can change the settings at any time to achieve a different sound. I personally am happy with the tone of the guitars by themselves, but in the mix I think they could stand to be a bit more pleasant if you know what I mean? They also have some compression applied.

Thanks everyone.
 
Mistral said:
Is there any real benefit to raising it to 32-bit float, mixing down as a 32 bit file, mastering it then applying dithering as the very last step - as opposed to keeping it 16 bit, mixing down and mastering?
The 32-bit float format has a couple of benefits that might be significant for you.

First, gain changes are completely non-destructive. With integer samples, you lose resolution if you lower the gain; bits are lost forever. With floating-point samples, gain changes affect only the exponent, and they are completely reversible; no information is lost.

Second, there's practically no such thing as an "over" in floating-point format. Unlike integer samples, where 0dBFS is the absolute ceiling, in floating-point 0dBFS is just another point on the scale. You have over 700dB of headroom! So your processing chain is very unlikely to cause distortion. The only requirement is that the level must be brought below 0dBFS before it hits your D/A converters.

The downside is that 32-bit samples take up twice as much space as 16-bit samples, so your track count might be reduced if disk I/O bandwidth is the main constraint on your system.

HTH,
Don
 
Thank you, that information helps a lot. Disk space is a luxury I have (400+ gb).. I am going back to the theory that it's all in the way I'm mixing my guitars. I think I was simply trying to make the rhythms too similar, and too loud, which ended up overboosting certain frequency ranges.. and as the other guy said, drowning out things like the drums.
 
Mistral said:
Disk space is a luxury I have (400+ gb).
Just to be clear: "bandwidth" refers to how much data must move through the wires in a given time, not the amount of space the data takes on your hard drive.

Don
 
Oh sorry I misread you. I understand what bandwidth is. The data is streaming from 2 fast drives.. the USB2 drive can achieve 41 mb/sec sustained data rates.. I don't think there should be any bottleneck with it. Right now the disk meter in Cubase isn't even registering anything.. I don't know whether that's a good sign or if it's just broken.. lol
 
Last edited:
Back
Top