Normalising your BR1600 master track?

[Super 8]

Or perhaps normalising your 9/10 MIX prior to applying the MASTERING TOOL BOX? (Or not at all!) What works best for you? Thoughts please ....

[AndyR]

I copy the final 9/10 mix and normalise the copy.

Then I check the Mastering tools against both. I usually Master the normalised version, I tend to prefer it - seems to make the tools work better?

Then I normalise a copy of the master, just in case I left some headroom. If it's not a lot different (or if normalising it adds something I don't like - has done once), I post the un-normalised master.

[Super 8]

BRILL! Thanks Andy. I'm only just getting around to looking into this side of stuff. For the longest time that Mastering Tool Kit feature on my BR1600 has lain redundant but I've been getting into it of late and getting some quite pleasing results though I say so myself (although tread carefully, some of the Master presets in there are awful in my opinion & can potentially ruin your track!) Very interesting to hear how you 'effectively' (geddit??!) 'Double Normalise' sometimes right at the last stage (ie: on your 9/10 mixdown bounce & then again on your post-mastered finished version), that's something I'm defo gonna explore. I guess the reason why I haven't bothered with this Normalising malarkey to date (hence my question) is because, many moons ago, someone told me (or I read somewhere!) that you shouldn't Normalise stuff at all - bad for your music! I can't rememember their reasoning why though! As such, I haven't entertained the idea ... until now and yes, you're right, I'm beginning to think/hear that normalising the 9/10 bounced mix prior to mastering does indeed make this Mastering Tool Kit feature perform, erm, better! I hadn't thought to Normalise the final mastered mix as well though so I'll go give that a try now too. THank Dude. CHEERS!       

[Super 8]

Then I normalise a copy of the master, just in case I left some headroom.

And hear's where I get confused by 'the rocket science' of it all. Correct me if I'm wrong hear but I've always been of the impression that mixes for mastering should ideally leave a good bit of headroom (say a max peak of about -6dB) in order to give the mastering engineer (or 'Mastering Tool Kit' robot in this case!) some space to work with? Surely if one's mix is normalised to Unity Gain (0dB) or as near as then there's actually nowhere left to go for the mastering bit? If this be the case then why, from my home studio tests this end, do my 'normalised prior to mastering' mixes sound noticeably better than just the original mix put through the Mastering Tool Kit? It sort of flies in the face of convention (especially if you then go on to normalise the mastered mix too!)

PS: While I'm back on, has anyone tried submitting a mix to this newfangled LANDR mastering service? I tried it and it REALLY whacked the levels up on my song but not in a particularly pleasing way. I much prefered the dynamics of my original mix (even though it was noticeably quieter). I guess it would be alright if you're into pumping Dance music and looking for a really squashed, fat, heavily compressed master with a ton of thumping bottom end but it really didn't work for the Folk/Rock song I chose to submit! Again, LANDR stipulate in their mix submission request that submitted mixes should have headroom space of at least 6dB in order to give their mastering software space to work with (which takes us right back to my initial questioning here!)       

[AndyR]

I've heard the "leave a bit of headroom for the mastering engineer" - and if I was supplying a mix to someone else for mastering, that's exactly what I'd do (I think the BR1600's Normalise tool lets you do that. For example, if necessary, you can take a 0db file and normalise to lower).

However - I found that the BR1600's Mastering toolkit seems to perform better (to my ears) if I normalise the mix 100%.

I normalise the final mastered file because I tend to master a bit cautiously so that I don't get any clipping etc. The BR's master fader is the last control in the mastering chain, I sometimes ride this during a mastering run to add further dynamics or apply a fade out, etc. If I do this, I have to make sure I do not go through the ceiling... so sometimes I haven't got the track's max peak to fully 0db. Normalising it ensures that. I have once found that normalising a master sounded awful - not sure what happened... I threw that away and posted the un-normalised one!

====
Now, I've also heard the "don't normalise anything it makes your recordings sh1t" advice.

I followed this for a while, BUT... I now think it's a load of old hoo-ey

Normalising of itself cannot make your recordings sound bad. All it is doing is finding the max peak and raising everything so that peak is where you want it (100% or 0db... wherever you want).

It does mean you can't safely push the fader above 0dB on a fully normalised track without  some form of compression to bring the peaks down. Nor can you safely add a further gain stage like EQ or chorus to the track, without lowering the fader, or the compression. (I guess that's why the mastering engineer likes headroom on the mix - but it's still down to his/her personal favourite tool settings as to how much headroom is needed... SO, perhaps the BR's Mastering Toolkit presets have been designed knowing that most users will send in a 100% normalised mix?)

No, normalising is cool and does not add horridness. It's what you do as well as normalising.

Here's the obvious problems you can introduce if you normalise everything, all tracks, before using them (I'm sure there are others, but here's two):

1. Noise. If you recorded at a low level, boosting the max peak to 0db is going to result in boosting the background noise. If you've got ten guitar or vocal parts, that's quite a lot of hiss you're adding.

2. So-called "brick-walling". This is when you compress or limit the peaks first, and THEN normalise. Means you lose dynamics and can produce things that are fatiguing to listen to.

My understanding now is that normalising isn't the problem at all in either case. Normalising is like burning a perfect "raise the volume control to the max without clipping" onto the recording.

Here's what's actually wrong in the above two cases:

1. The recording levels were too low, you were going to have to turn them up anyway and thereby add the hiss - normalising is not the culprit. You need to sort it out next time - record with a better signal-to-noise ratio.

2. You've compressed or limited the cr@p out of everything, well... what did you expect?

I personally believe there are strong arguments that both, but certainly the second, are ARTISTIC DECISIONS. If someone wants to brickwall their recordings, that's fine. Old dudes on audiophile sites will say your material is unlistenable and it gives them headaches. They talk about DR ranges (dynamic range) and wotnot. And it's all the fault of pro-tools and normalising....

I dunno... I see there point on remasters of The Grateful Dead or wotever... but if [insert name of bright young thing] wants to whack it to their listeners... and their listeners want it... umm....

===========

Finally, and a bit of a "by the way" - I used to fiddle A LOT with the BR1600's Mastering Tools. It is worth doing to find out what's going on. But...

I kinda use the presets now unless I know something more needs doing. Usually the preset sounds wa-ay better than anything I can do!

I'll always have a little experiment with pushing the preset hotter and limiting, to see if a bit of "brickwall" (maybe "wattle-and-daub-wall" in my case) actually works on my track. But I'll try to let my ears decide which I like.

I always go back to comparing to the original mix -

"could I have made it sound like this in the mix?"
(the answer's usually "yes, but I really cannot be @rsed to go back a redo it all" )

"Ok, it's quieter, but does the original mix sound better?"

"what's missing now, or has been added, that I don't like?"

... perhaps a decent mastering engineer with real mastering tools would do a much better job. I've not tried these LANDR etc services. Might do yet!

[Super 8]

Man do you know your sonic onions or what Andy? Awesome reply Dude (I'm gonna print it out for some bedtime reading - LOL!) No, seriously, that's a great reply mate - thanks very much for going to all that effort there, you've really cleared things up for me in this department and I totally agree with pretty much everything you've put down here.

I'm kicking myself for only discovering this Mastering Tool Kit now - so much so that IF I'D KEPT A BACK UP OF MY PREVIOUS SONGS (=NOPE!!!) I would have wanted to go back & re-master them myself. Hey Ho! I'll just make sure everything I try to do from here on in is finished in such a way (I've just started working on a solo album in fact so this dialogue has come at a really good time for me!)

Having tried a number of things, I too agree that it's worth sticking with the on-board BR mastering presets. In fact, having tried 'em all I usually just go with that first one ('P01:Mixdown' off the top of my head - I'm away from my machine at present). I never rely on mastering to fix stuff - I ALWAYS try to get my mixes sounding as good as I possibly can rather than rely on the crutch of mastering to magically 'fix' stuff. Basically all I'm looking for from a master is for it to retain the vibe of my original mix just boost the level to that of perceived industry standard for a commercial release (ie: so it'll stand up against other recordings were one of my songs ever to get played on the radio!) That's what I didn't like about LANDR when I tried it. It seemed to add extra stuff that I didn't want there PLUS on the one submitted track in question it completely mucked up the level of the lead guitar part on the left channel - it was WAY louder than I'd set it in my original mix.

OK, before I skidaddle, can I pick your brains again Andy RE: your advice 'Record with a better signal-to-noise ratio'? To clear this in my mind once and for all would you say what I'm doing is right gain structure wise on input recording? Read on .... I take an audio input (be it Voice; Guitar; Bass; whatever let's stick with bass eh?) So I play my bass at the volume & way in which I intend to record it watching for the Input Trim red light to illuminate. When it does, I roll back the Trim so that it only flickers on the loudest bits. Once I've set this, I then turn my attention to the Input Level control (aka 'recording level setter' right?) and set this so that my signal is hitting the line on the left (which when you look across to the right meter readout is about -4dB I think (I wish the metering was better on the BR1600 ie: it went up in increments of 1dB. Poor meter readout is one of he few things that bugs me about this unit! There should be a behind-the-scenes, sideways meter that let's you set what level you're inputting at a lot clearer - too much guess work as things are I find!) Anyway, I digress, where was I? Erm, yeah, I think that's it actually! That's how I set ALL my input levels to date. That sound about right to you??     

[AndyR]

Yep, that does sound about right to me. I pretty much do the same as you - get it as hot as possible at each stage without clipping. Depending what I'm recording, it might also be coming in through a mixing desk... so I have an output from the instrument, another input gain on the mixer (with nice pre-fade listen metering ), the track fader, then a group master that sends to the BR... and THEN the input trim on the BR, the "record level" - and also, if I have an effect on, there's gain in there too.

But yeah, hot as possible without clipping all the way through. Sometimes I find my signal from the desk is too hot for the BR input - so I have to lower the group master on the mixing desk. Never have figured that one out! I expected that master fader to be on 0db but I can't get ANYWHERE near that... (I probs need to check the back of the desk one day - its line level is possibly mismatched with the BR's? There's probs a switch somewhere ). However it's no problem spotting it - the clip/overload light on the BR shows and I can usually hear the clipping - it usually happens on bass for me. So I just lower the group fader on the desk.

If I'm real worried, I'll unplug everything to take the desk out of the chain. The thing is, the desk and the monitors are the power amp and monitors for three keyboards and three or four amp modellors... If I want to play any of those things, I just turn the desk and monitors on... So it just seemed v clever to take the group send out to the recording machine. Send the right stuff to the group and hit record!

========

Anyway, when I first started (especially on a DAW) I'd be recording a guitar and get hardly any. Yeah, if you normalise it, it loud enough to use - but oooh there's some hiss you find later!

Then someone taught me how easy it is really... they said "it's called gain-staging, or something, sort out each stage through the chain, STARTING AT THE OUTPUT LEVEL FROM THE INSTRUMENT if it's a line-in thing like a keyboard or guitar modellor, then the input trim, then the eq, then the track fader, then the master... etc... Get each stage in turn as loud as you can before nasty distortion... then move on to the next stage... - do that, and you should have decent level, no clipping, minimal noise, at the end of the chain where it's going to tape"

The first time I tried it like that, it worked, and I suddenly stopped worrying about it

Basically, if I'm able to mix to decent levels without having to normalise any individual tracks to get them loud enough to go with the others, then I feel I'm probably doing OK...

[Super 8]

Cool Andy. Thanks again for sharing mate! Glad to know we're singing from the same hymn sheet :-) It's been nice corresponding with ya & I'm looking forward to hearing more tunes from you in the near future. CHEERS!!!

[AndyR]

I've actually demoed a new song on my old MBR, sat on the sofa, over the last few days (no click, just several guitar tracks recorded with Les Paul Special and a Yamaha THR amp, and two vocals... all through the MBR's built in mic).

The demo got done because I needed to tape the idea before we watched something on TV on saturday morning - oh, yeah, the F1 Qualifying... Then she went out, and I kept fiddling with it.

That's the first song written in almost a year. We (me and the missus) aren't convinced the title and some of the chorus lyrics are the best they could be yet... but might be ok.


It's just possible I might take this one up to the BR1600

[Super 8]

YES! Do it. Great to hear the music is beginning to flow again. Very much looking forward to hearing the finished masterpiece. You're back in that musical saddle again amigo - let's hear that Les Paul Spesh wail! Bring it on!