Track looses punch and sounds shit after "mastering". Help?

[Niketin]

Just so everyone knows, I'm still pretty much a newbie when it comes to mixing properly.

So I have this new track almost finished but the "mastering" (making it louder) ruins the track and I can't finish it.
Without any mastering compression/limiting the track sounds they way it should and nice when I crank the volume up from the speakers but very quiet at normal listening volume, having its peaks around -3dB. So when I apply iZotope Ozone "Loudness maximizer" to make it normal volume, the track looses its punchiness on the drop and the volume of the main bass varies a LOT at some points.
How could I fix this? Any advice?


And another smaller question: Whats the common way (if there is one) to have your master channel when mixing? Should I have it empty and just level the channels low enough so that the master won't peak, or should I have a compressor or a limiter on the master channel making sure that it won't peak and just have the other channels hitting almost 0dB? Or what?

Thanks in advance!

[mthrfnk]

Protip: Don't use Ozone's presets.

If you just want the track louder consider using a limiter?

[Samuel_L_Damnson]

Work on getting your mix perfect with nothing on your master channel. Mastering things yourself usually makes things worse anyway (although it will get louder usually). Louder doesnt always mean better. Its best to have someone else master your track anyway because a fresh set of ears will hear the errors and problems easier than yourself.

[hasezwei]

tip: don't use izotope presets without knowing what they really do. i did that a lot myself and yes, it squashes the fuck out of your tune killing all it's punch (most of the time at least)

i normally put a limiter on my master set to -0dB so that my speakers won't blow up when i accidentally do something weird or clip the master, and generally see that it peaks at somewhere between -6 and -3 dB.

now if you wanna 'master' (as in self-master) your tune do the following, it's what i personally like the best:
bounce your tune the way it sounds best, in this case peaking at -3dB. now make a new project file and load in your .wav, normalize it (so it automatically turns it up to the point where the loudest peak in your tune is at -0.00dB, should be an automatic function in every DAW) and then take a look at the waveform. if everythings still quiet with a few peaks going up, put a limiter (i just use the kjaerhus classic master limiter cause its simple as fuck and free) on the channel and carefully adjust the threshhold so that it leaves the actual tune alone and only catches those peaks.
then listen if it still sounds ok at those points or if it's obviously squashed. if the latter is the case, go back to your project file and fix that, if not bounce the result as a .wav.

now your tune should be sufficiently loud while still retaining the punchy dynamics.

[eldoogle]

^This sounds like a really good idea.

[Samuel_L_Damnson]

^This sounds like a really good idea.

This^ Never though about normalising it when really it just makes sense

[hasezwei]

^This sounds like a really good idea.

This^ Never though about normalising it when really it just makes sense

normalize. fucking. everything.

[Samuel_L_Damnson]

Does it change the signal/noise ratio at all or introduce any bad artefacts? I suppose it would be scaling everything up relative but im just wondering. Ill go and try this later though.

[hasezwei]

no its simply turning up the volume til some point reaches 0.0dB. its nothing you couldnt do with a volume fader yourself, just way more precise and fast.

of course if your sample/wav is very quiet and has loads of background noise that will be turned up too, but since we're all working in the box and not recording guitar amps with microphones or stuff like that it really shouldnt matter.
i just prefer normalizing samples cause it gives me greater control over the levels.
if i have for example a kick and a snare and one hits -2 and the other -4,7 dB or so, and i would want both to be exactly the same volume i could fiddle around with the volume faders til it sounds about right (or the numbers on the track meter fit together) or just thwack a limiter on the channels and completely mess up my drum hits. OR i could normalize them and not touch the faders at all.

however thats just my opinion and im sure there will be people with good arguments against this approach, in the end its all about what suits your workflow best

[Vespers]

Couple thoughts here mate. Mastering is soooooo much more than making a track sound loud. It's a whole separate art from mixing. If you really want to learn more about it, start doing some reading, like Bob Katz' book Mastering the Art and the Science.

As for Ozone, as some others have said, you can really destroy your tune with presets and the loudness maximizer. You want to nail your mix as much as possible first. Really watch your sub bass levels (use a spectrum analyzer if you need to) as those frequencies will start robbing you of valuable headroom when mastering.

I don't do the final masters on my own tunes, but when I'm doing quick masters, I always use some gentle bus compression before a limiter. Soft knee, 1.5 or so ratio, opto mode for a bit of a "laggy" quality, and slow release 1 second or more. If you're in Ableton, the built in Mix Gel Compressor preset it good for this.

When you get to your loudness maximizer, or limiter, make sure you're not getting constant gain reduction. You just want it to be shaving off the peaks. If you want your track to slam more, try using a bit more compression and limiting on your main elements in the mix. I usually use some light limiting on my kick bus, snare bus, and bass bus.

Hope that helps. EZ.

[Vespers]

no its simply turning up the volume til some point reaches 0.0dB. its nothing you couldnt do with a volume fader yourself, just way more precise and fast.

of course if your sample/wav is very quiet and has loads of background noise that will be turned up too, but since we're all working in the box and not recording guitar amps with microphones or stuff like that it really shouldnt matter.
i just prefer normalizing samples cause it gives me greater control over the levels.
if i have for example a kick and a snare and one hits -2 and the other -4,7 dB or so, and i would want both to be exactly the same volume i could fiddle around with the volume faders til it sounds about right (or the numbers on the track meter fit together) or just thwack a limiter on the channels and completely mess up my drum hits. OR i could normalize them and not touch the faders at all.

however thats just my opinion and im sure there will be people with good arguments against this approach, in the end its all about what suits your workflow best

Agreed. There's a great batch tool to do audio processes like that. It's called Sample Manager by Audiofile Engineering. You can normalize, change bit and sample rates, rename...all kinds of useful stuff. All in a batch process, in-place. Super useful.

[Smiles]

http://tarekith.com/assets/pdfs/Mastering.pdf

Read this. A lot of good info in this guys stuff.

[Niketin]

Thanks for all the replies!

I tried to normalize it and then use a limiter on it but it still sounds awfully quiet.
I recently started to experiment with the way of my master channel when I start mixing and that's probably the problem here: on this track I first mixed so that my kick and snare hit around 0dB and then later when I wanted to lower the volume of all the individual tracks I had to assign all of the channels to one channel ( because in FL I can't just select all the tracks and lower the volume like in other DAWS).
If I would't lower the volume like this it would be peaking a lot on the master.

Here is a clip from the drop if it helps (no normalizing or anything done to it, peaks around -3dB)
Soundcloud

[Burgeamon]

You don't want your kick / snare hitting at 0db as that'll take up all your head room. As a rule of thumb I get my snare / kick peaking at -10 on their channels and then mix the rest of the track from there. By the time I've mixed all the elements in the track is peaking at between -5 and -3 on the master and then I render and normalise. I have played about with oZone a bit but I honestly feel I get better results from mixing with an empty master channel. Ozone will make your track louder but if you don't know what you're doing then you'll get that volume you want but your track will be squashed as fuck and you'll lose a lot of punch.

What hasezwei said sounds like a good way to do a self master and once you're happy with the track I'd just pay someone else to do it. There are a lot of guys out there who may not be the best mastering engineers in the world but can provide you with competent masters for very little money.

[webstarr]

Get it mastered by a professional engineer who knows what he's doing

[Electric_Head]

Get it mastered by a professional engineer who knows what he's doing

^this

Or don't look at your own mix down as mastering.

[Vespers]

Get it mastered by a professional engineer who knows what he's doing

For sure. And, at the same time, it's also good to understand how mastering will change your tune. I think having the skillset to do a decent quickmaster on your own is a wise investment of time and energy. The more I understood the mastering process, and had an open dialogue with my mastering engineer, the better my mixdowns got.

It's truly amazing how much running into a limiter or compressor will change the balaning of elements in your mix. The whole Mix it Like a Record philosophy of mixing into a bus compressor was something I dismissed earlier in my career, but now realize it has a fair amount of merit when only used for a quick reference.

[knobgoblin]

Never. Ever. Ever. Use presets for mastering. If you dont really know what you're doing, you are probably doing more harm than good. Only use the proccessing that the track needs to reach its desired Sound. Tarekith is a good guy, and he knows his shit. His tutorials on self "mastering" are very good.

I would recommend not normalizing before processing unless you are going to use nothing but a limiter. The problem with normalizing before hand is that you can quickly run out of headroom and be clipping in between plugins. For example, if you normalize and then use and EQ before the limiter, then if you do any boost at all, and often when you use cuts (because of phase issues related to the eq process) you will clip between plugins. Ableton is cool in that it has level meters in between plugins, so you can see when you are doing this.

The basic trade offs in mastering are balancing clarity and punchiness against volume. You have a fixed ceiling for your peaks (really should master to -.1 or even -.15 to accommodate crappy consumer DACs that will often interpret those as clipping). If you are limiting more than 4 dB(in extreme cases), you are limiting too much. It will kill the transients relationship to the rest of the sound. It should be a gentle shaving off of the peaks. The loudness will often come from EQ and compression. Mastering compression is probably the hardest thing to get right other than the EQ. The soft knee/low threshold combo is a great way to boost RMS without destroying the context of the transient. In conjunction with a faster more "limiting" type compression (just shaving peaks by 1-2dB at a 4:1-5:1 ratio) you can get a pretty loud sound before even engaging the limiter. Also, you should use the absolute best quality compression you can get afford of for 2buss/mastering work. The difference in quality of compression is vast from the stock Ableton/logic comps and something like "the glue" or UAD plugins.

Stay away from multiband processing unless the track absolutely needs it (a good example would be your bass volume jumping all over the place or a harsh high hat that only happens in certain sections) and even then, use with extreme caution. This is the place where most people's skills and experience falls apart, and for the most part is mainly used to fix things that should have been fixed in the mix. Same goes for M/S processing (very powerfull in mastering, but a double edge sword...a little goes a long way)

EQ is also somewhere where a little bit goes a long way. It's rare for a mastering engineer to boost or cut more than a dB, possibly 1.5 or 2 in extreme cases, unless the mix has lots of problems and needs to fit with other material better.

Hopefully some of that helps...

[benjam]

Read this its awesome
http://www.izotope.com/products/audio/o ... gGuide.PDF

[didi]

For your second question, as a rule of thumb leave your master channel blank.

If' you want though, you can mix into a compressor on your master channel. I'd only advise this if you had high quality, and appropriate compressors.