KVR Audio

Hink · Post by **Hink** » Tue Jul 06, 2004 5:53 am

a few years back I heard my first song on the radio, it actually climbed to number 17 of most requested music. Everytime I heard it I cringed, I could hear all kinds of compressor breathing. It was then that I realized you gotta know your limitations, and when it comes time to polish my work I feel it's worth the investment to send it to a good mastering house...not a place to be mixed, just mastered...some of these guys are artists in their own right.

Crackbaby · Post by **Crackbaby** » Tue Jul 06, 2004 7:47 am

BONES wrote: No, that's when I bypass and hear no difference at all. I never compress any tracks, except with this new bass compressor, and I'm completely happy with my final mixes.

Im happy for you

freeztar · Post by **freeztar** » Tue Jul 06, 2004 8:31 am

Now now...there's more than one way to get loudness. I personally use a dbl mastering limiter. It's easy to use and sounds good (DX though). I like Garret's idea a lot and will try it out. I've just used my ear in the past, but a measurement would be nice.

The most important thing to getting "good" loudness is proper tracking. I strive for a -6 to 0 minimum peak input level for vocals, a -3 to 0 minimum peak input level range for guitar, and 0 for most plugs and syths. That gives me more headroom on the compresser/limiter without hearing the noise floor.

If you plan to have radio play, and can't afford outside mastering (not a good idea to master your own stuff!!!), then keep in mind that radio stations compress everything. That was probably part of the problem with hink's radio track. It's compression probably sounded fine at the mixing desk, yes?

Also, if you plan to have your stuff mastered, don't use compression/limiters and do mix down to a peak of about -3 db.

tweakin · Post by **tweakin** » Sun Dec 12, 2004 3:36 pm

OK, err... Here's a noob question (that I'd be MOST appreciative to find an answer for

)

When I render from Cubase (32bit float WAV), then Import to Wavelab to use mastering tools, the sound is clean and sharp- quite similar to what Cubase plays back (only a few db quieter?!?), but when I use my mastering plugins and tweak the sound, I then "save as" (as a 16b/44k), the sound is heaps lower still when I play it back again...

Conversely, when I run the same mastering plugins in the master channel of Cubase (althought I can't hear the track continuously because of the jumping!!), the end result 16/44 sounds dynamite!!

Is there a standard rule or method that is used to convert from 32bit float format to 16bit?

Does anyone have any links or references to somewhere I can learn how and/or according to what rule all the channels are summed when rendering to WAV?

I'm hoping this info will help me get a louder, cleaner (more like the cubase playback) mixdwown with no dynamic processing to send off for proper mastering. (could always buy a tape machine (

)

Cordelia · Post by **Cordelia** » Sun Dec 12, 2004 4:30 pm

tweakin,
A number of mastering houses these days prefer the 32 bit FP files. Otherwise, they are processing (EQ, compression, etc.) already dithered files. You should always save dither for the final stage, and dither only once.
If you want to find out about the standard methods for conversion from 32 bit FP to 16 bit, Google "dithering".

Glooper · Post by **Glooper** » Sun Dec 12, 2004 4:46 pm

Another thing worth checking that hasn't been mentioned yet is for bass sounds. If your track has a lot of sub bass in, that sub bass can eat up a lot of the volume. However if you don't gave a sub-woofer.. the chances are your not hearing that bass.. and haven't accounted for it. Using spectrum analysers and high pass filters can sort it out easily enough.

Ben

tweakin · Post by **tweakin** » Mon Dec 13, 2004 3:02 pm

Thanks for replying guys! I actually emailed my mastering house and asked if they would prefer 32bit FP files but they said that would only be critical in Classical genres (as opposed to Uplifting Trance which is what I make) and to just send 16bit files.

I'm still curious about rendering though and would like to understand for instance, what happens when you have signal way over 0 db (in your 32bit FP host) and you render to 16 bit, as opposed to having all the signal just below 0db then rendering to 16 bit. The results of the two differ but I can't understand how, ie. what summing or averaging method is used to combine all of the channels...

I don't expect anyone to reel off a verbatime explanation, but if anyone has any links or references where I can read up about it that would be mint!

btw: dithering I do kind of understand and I leave that to the mastering engineers, it's more a rendering/summing query - but cheers!

IIRs · Post by **IIRs** » Mon Dec 13, 2004 3:08 pm

BONES wrote: You know, whenever I apply a compressor to an individual track I never get any sense that it is anything but louder or softer. If I balance the levels I can bypass them and not hear any difference in 99% of cases. I just don't get it.

You need better compressors / monitors / ears

C00kie · Post by **C00kie** » Mon Dec 13, 2004 3:14 pm

tweakin wrote:what happens when you have signal way over 0 db (in your 32bit FP host) and you render to 16 bit, as opposed to having all the signal just below 0db then rendering to 16 bit. The results of the two differ but I can't understand how

I presume anything above 0dB will become "clipped", which is a very bad thing (unless you like the worst possible distortion imaginable)

Do yourself and your public a favour, stay below 0dB when you're going to convert to 16bits. Maybe use a brick-wall limiter (the crushing elephant?) to handle that for you.

Glooper · Post by **Glooper** » Mon Dec 13, 2004 4:30 pm

tweakin wrote: Thanks for replying guys! I actually emailed my mastering house and asked if they would prefer 32bit FP files but they said that would only be critical in Classical genres (as opposed to Uplifting Trance which is what I make) and to just send 16bit files.

Change mastering house! Yes it is more important in classical music due to the dynamic range. However for pop music I would still expect them to ask for 24bit files at least. The ear can hear around to 20bit of definition... If you master a 16bit file then you will create distortion. You should always perform processing above the target medium and render down at the very last stage.

tweakin wrote: I'm still curious about rendering though and would like to understand for instance, what happens when you have signal way over 0 db (in your 32bit FP host) and you render to 16 bit, as opposed to having all the signal just below 0db then rendering to 16 bit. The results of the two differ but I can't understand how, ie. what summing or averaging method is used to combine all of the channels...

It depends on the host.. 0db can be different on different digital applications. Some hosts will leave (say) 6db of space above 0 for headroom, So when mixing down you will be aiming to have all peaks below 6db (This is what cubase does, and is what I am using). And of course if mastering in Cubase (as I do), then the loud sections of track will go into this 6 db. Some digital hosts have 0db as the top of the scale and others will leave say 3db or 8db etc.

tweakin wrote: I don't expect anyone to reel off a verbatime explanation, but if anyone has any links or references where I can read up about it that would be mint!

Read anything by Bob Katz. He is regarded as one of the top mastering engineers. I have one of his books (mastering, the art and the science) and it is superb, easy to understand and very informative. His website is

http://www.digido.com/

tweakin wrote: btw: dithering I do kind of understand and I leave that to the mastering engineers, it's more a rendering/summing query - but cheers!

Well if your mastering engineers are using 16bit files then they won't be dithering! Dithering is used at the last stage of a mastering chain when converting bitage (from 24 -> 16 or 32 -> 16). It is used to make a 16bit file sound more like a 20bit file (for example)...

There is some info on how dithering works found at ....
here!

Ben

IIRs · Post by **IIRs** » Mon Dec 13, 2004 4:48 pm

Glooper wrote:Well if your mastering engineers are using 16bit files then they won't be dithering!

Not true: any processing will increase the wordlength, so if the material is mastered digitally the end result will still need dithering to 16 bit.

Glooper · Post by **Glooper** » Mon Dec 13, 2004 4:52 pm

platinumears wrote:
Glooper wrote:Well if your mastering engineers are using 16bit files then they won't be dithering!
Not true: any processing will increase the wordlength, so if the material is mastered digitally the end result will still need dithering to 16 bit.

true... But it depends on the process and the processes input/output. If the process still chucks out 16bits then no dithering will make a difference...

Still the main point is that you should be keeping your bit rate as high as possible until the very last stage of the mastering chain.

Ben

Cordelia · Post by **Cordelia** » Mon Dec 13, 2004 4:57 pm

So, platinumears, if you're working in a 32 bit FP sequencer and you render a 16 bit file, is the file dithered down? I know in FL there is an option to dither when rendering a 16 bit file, so I assume this means you are either going to dither or (if you uncheck "dithering") you are simply truncating.
If this is true, wouldn't it be foolish to send a 16 bit file to a mastering house?

IIRs · Post by **IIRs** » Mon Dec 13, 2004 5:04 pm

Cordelia wrote: So, platinumears, if you're working in a 32 bit FP sequencer and you render a 16 bit file, is the file dithered down? I know in FL there is an option to dither when rendering a 16 bit file, so I assume this means you are either going to dither or (if you uncheck "dithering") you are simply truncating.

Depends entirely on the host: sounds like FL gives you the option.. other hosts (eg; Tracktion) just bundle dither plugs which you can insert into the chain or not as you please..

Cordelia wrote: wouldn't it be foolish to send a 16 bit file to a mastering house?

Yup!

digitaldoom · Post by **digitaldoom** » Mon Dec 13, 2004 5:07 pm

A few months ago I began playing with Waves L3. That's the best software compressor I've ever used.

If people read up on this thread, and research all of the information posted, I think they will be able to get 'good' masters of their tracks.

Thanks for all the great tips and techniques everyone!

A Simple Mastering Question (regarding low output)