I'm at a state with many of my mixes in my current project where I would like learn to do some of the finalizing steps so that I'm better prepared and have some learning under my belt for when the time comes.
I believe what I need to do now is make sure levels are consistent between tracks (songs), using some form of limiting and level checking either using a completely different software other than my DAW (Sonar Platinum) or using some plugins on the bounced tracks.
For the mix, I'm using some hardware (tube preamp for some color, and an old MDC 2001 Dynamics Processor on the 2 bus as an insert along with analog summing through my board, then "live" recording the board output back into the DAW.
I see many posts here already about "mastering", so don't really know if that's what to call this, but what I hope to achieve is finalizing for posting to streaming services, CD stamping, and other distribution medium. Not really concerned with going vinyl as I understand, that's a whole other can of worms.
My questions are:
What workflow do others currently use to do this type of thing.
Which metering plugin would be recommended for ensuring that the bounce is coming in at the appropriate level (not too hot, and not too low for maximum bit depth accuracy), and what level should I shoot for (bounce coming back in)?
Any recommendations you guys have would be helpful, and if I'm not providing enough information for you to help me, feel free to ask for whatever information you need.
I'm using Windows 7 64 bit, Sonar Platinum 64 bit, MOTU HD192 Sound interface, and a Mackie 24/8 board.

 

Hi..

U Write much things..but some of them is different ..

Lets tell you that there is some steps you must Follow. .

Like : create mix master..

U cant master before mix.

Mastering is the final step u do..the last..

Levels are very important.. in a track there is a kick and vocal the dominant in the top. .everything Else should be around the that.

No matter What daw u use ..in a daw u have same thing...like com eq synth etc..

Daw is Just a tool u can use.

Write to me if u need help about everything..

 

YOu can scan for replaygain, and apply offsets to your songs in your DAW, or use loudness measuring plugins like izotope insight, prolly some free ones out there too...

 

I like the old tool called Har-Bal2 for this job. It's very convenient to balance the overall RMS levels of 10+ different tracks on an album without tearing your hair out doing a ton of DAW setup for it everytime.

That's the simple part of what it'll do for you, the actual selling point of the application is that it can also balance the frequency ranges across those tracks to match better and sound better across different speakers, but this use for it is more of an art. There are automatic tools in there for it too, but as usual they don't always hit the spot on first try.

Best part about using a separate application is that it's much less fiddly. Everything geared toward what you're doing and nothing else... This is a stage in the production that we should all really be outsourcing to India, it's pretty much braindead Monday morning work.

 

I'm using similar type of workflow as you do, except the console thing. I think having a one good hardware processor through which you can "analogise" tracks and bounce back is important, but a console on top of a good hardware processor might be too much. Outboard processing adds to tracks something that I haven't been able to achieve with plugins.

Recording levels? Anything goes for as long as it's not clipping the AD converters in your audio card on the way in. Don't be too obsessed with the recording levels as you are recording at 24-bit. The highest optimum level is the one that doesn't clip the peak meter in the DAW, pretty much. However, DA converters can start clipping at levels like -6dB even, so I usually record at up to -6dB peak level. If you're asking about recording the whole mix through outboard, the same rule applies. It's better to leave those 6dB of headroom for digital limiter in your DAW to handle, or a pro mastering engineer.

If you stick to the -6dB rule, you'll get a bit low RMS level recording, but don't worry about it. Apply some good quality EQ for equalising the tracks [songs], to make them sound similar, and then some good limiter and dithering at the end. Good metering plugin is also quite essential, especially if you're not making square waves, but quality stuff. Keeping things simple is the key to good sound.

If you're making square waves, then use a multiband compressor before the limiter for cleaner limiting. If you're not making square waves but leave about 8-10dB of headroom in your songs, then you don't need it and your mixes will sound wonderful and more natural. Cheers!

 

Last edited: Mar 14, 2017

Thank you, I will look around for this. Exactly the type of tip I'm looking for in terms of getting volume between tracks for finalizing. I was looking at this as you said, "brain dead work" thinking how tedious this was going to be getting levels right using the DAW.

I went looking for the HD192 specs to see at what level clipping occurs, and in the process read that on top of the level meters the top most LED is a clip indicator which if triggered, will stay on until it's reset. So I assume all I need to do is adjust the Master Fader on the board so that the clip indicator LED never lights and I'll be in good shape as far as introducing clipping distortion. I need to look on the front and see if -6dB is marked or not, can't recall off the top of my head. I'm assuming you say -6dB (if 0 dB is where clipping actually occurs on the HD192) so that you have some extra headroom for further processing. Thanks for that tip.
As for using the console, it seemed to add more dimension than just the outboard processors did. Something about running it through the console strip and master bus on the board gave it a nice cohesive sound I had been struggling to find. Not saying you couldn't do the same with a plugin chain, this is just what worked, and was dead easy to implement with equipment I already have. I call it "poor man's summing box".

So I guess the final question I need to get answered is what level to shoot for as a "finalized track", ready for distribution. And how to measure that (plugin, software, etc.) I see talk about different LUFS levels, RMS, etc. What's the proper way?

Also you mentioned dithering at the end. Can you elaborate on this or point me to some further info that you found useful.

 

The proper way? In my opinion the proper way is to aim for K-14 or K-12 levels. http://www.digido.com/how-to-make-better-recordings-part-2.html
But most of the people will tell you to just use your ears to get the proper level, or they'll tell you that having 3dB of dynamics in a song is great. I guess what I want to say is - it's your choice. Loudness wars are coming to an end with most of the audio streaming services serving equal loudness songs, meaning no matter how loud or not loud your songs are they will sound just as loud as other songs after they process it.

Some more very important reading on loudness: http://dynamicrangeday.co.uk/about/

To become aware of what loudness wars do to a track, try limiting one song as loud as possible or as loud as you think it can go [possibly K-8] without sounding like Metallica album, then limit the same song to a K-12 standard, and then put both at the same level and listen. The "louder" one will sound like shit, and not louder at all. It'll sound limp, powerless, and lifeless, and more distorted, too. Everything about loudness is just a psychological thing exploiting the traps of human perception of sound.

For metering you can use free Voxengo SPAN or even better ToneBoosters EBULoudness plugin [more visual loudness metering], or his Barricade limiter which is also a good limiter when you're not trying to push things into the square wave territory, but the aforementioned K-14 or K-12.

K-8 [doesn't exist] would be average RMS level at -8dB with occasional RMS peaks to -4dB. Leaves 4dB of headroom for peaks.
K-12 is average level of the song at -12dB with occasional RMS peaks to -8dB. Leaves 8dB headroom for peaks.
K-14 is average level of the song at -14dB with occasional RMS peaks to -10dB. Leaves 10dB headroom for peaks. Most albums from the early 90s are kinda K-14 as I found out even though it hadn't been invented yet. Depeche Mode album "Violator" for [a good] example is within K-14 limits.

I haven't even touched LUFS because if you're new to metering just try to understand one thing and then it will be easier to understand the other. LUFS metering is all the shizz these days and is replacing K-system which actually never really took off for one reason or another. But I can say it anyway that if you make a track at K-14 or K-12 level it will sound great even 20 years from now, which I cannot say for all those square waves with only 4dB [if that] of dynamics. It is also recommendable to keep a non-mastered or non-limited version archived for - you never know.

 

For a freeware combination for metering and limiting you can use Voxengo SPAN and LoudMax combination, but LoudMax doesn't include dithering, so it would be better to use ToneBoosters Barricade as a limiter, or you can use MDA dither after LoudMax to get a properly dithered track. Mr. Airwindows has a nice assortment of dithering plugins here. I suggest you to use a common TPDF dither for a start.

What's dithering? Here and here. Loads and heaps of great info there! Enjoy mate!

Regarding recording into DAW, I've read at multiple places that -6dB peak at AD is the safe value for every audio interface. My TC SK48 actually starts to clip/distort at 0 dB they [TC] say because it's got bigger headroom, but I still use -6dB for just in case. You never know when you could get a clip anyway so it's better to be safe than sorry.

 

Last edited: Mar 14, 2017

SineWave, I've pretty much given up on dither since most of the music I work on these days is aiming for iTunes/Soundcloud/beatportapotty. What do you think about dither in context of inevitable mp3 encoding and these brave new frontiers?

 

If you're encoding into a lossy codec like MP3, AAC, or AC3 it is actually better not to dither but encode the song directly from your DAWs master output. Again, this is only relevant if you're encoding a song directly into a lossy codec purely because of the psychoacoustic encoding that lossy codecs use to encode the songs. Have you noticed that songs encoded in MP3, AAC etc have no bit depth? Just the sampling frequency. You could say that bits change dynamically in a lossy codec, and dithering is unnecessary in this case. Cheers!

Edit: Personally I always dither [simple TPDF] and archive songs in 48/24 or 96/24, unlimited. I use these archived tracks as a "glass master" [oldskool term...lol] to make limited-EQued-mastered copies from in whatever standard is needed later on. It is very convenient to have a non-destroyed copy in any way. For you never know, eh?

Edit2: An easy way to master a song without dithering directly into a MP3 or AAC is then to use [on the master output] an EQ [EQuality? Pro-Q?, SlickEQ?], a good compressor [Cytomic Glue? Pro-C? Kotelnikov?] into LoudMax [Pro-L? Elephant? Limiter nr6?] and encode it that way. EQ could be used to make the song sound better but generally to make the compressor work less... and both are used to make the limiter work less so you get a louder master with less distortion which is the worst enemy and hardest to defeat when it comes to making "square waves".

 

Last edited: Mar 15, 2017

@SineWave , thank you very much for your time. I'll be studying what you've posted. You've clarified a great deal simply in the posts.

 

Wow, HarBal2. I forgot about that one. The first eq ripper I used. It could do batch processing too if I remember correctly. I used to use it to rip an eq curve from a commercially released track and apply it to my tracks. You can definitely use this to make all the tracks sound like theyre all eqed to sound like theyre on the same album.
Now I use Pro Q2 or the eq in ozone do this when needed. To rip eqs curves and apply them to my tracks.

 

I am mainly working with house music since 2010. What i learned all these years is to refrain from using dynamic processors on individual channels as much as possible. Just watch your levels.

 

Don't mention it.

I'm sure others have loads of stuff to add to it, too. There are many ways of doing it, and their ways might differ. Like this Harbal... I wouldn't touch it with a shitty 10-foot stick. You can get similar results with Waves Q-Clone and Voxengo CurveQ. I've been playing with it long time ago, it's of no use when you make a good mix and mastering - it's a tool for repairing badly mixed songs *IMO*, so I'm not gonna say that it is totally useless. It could be useful, just not every time.

It is up to you to find your own way. I think of music production as a science, because it is not exact and it constantly evolves.

 

Last edited: Mar 17, 2017

Agreed in so far as the EQ matching functionality. I'd much rather get that fixed at the track/stem level vs. overall mix. The different songs are beginning to sound very similar already, likely because of some of the mix choices I'm making, and the mic usage was pretty similar when tracking.

I would only be looking at such a tool for level matching between songs, if it is useful for that (haven't tested yet). Hence the questions about metering methods and levels to shoot for, so each song is similar in terms of volume/level. Kind of an eyeball sanity check for my ears, and to make sure I'm within specs for whichever delivery method.

Really interested in reading about dithering. I've seen it referred to in some of my research before, so was pretty sure it would come up at some point, but anything I've read already was very vague.

 

The Har-Bal is a funny thing, my version has got a clunky Windows 3.11 interface straight from 1990. So no wonder it cracks Sinewavey up thinking about someone using it on albums. Here's the thing though, it's fundamentally not the same as CurveEQ or Q-Clone. Har-Bal can do what these do, but I never used it for that myself because I don't like the sound of "matched" frequency responses, I can never get it right and it pisses me off. So to save some sanity, I leave it the hell alone.

But the Har-bal is basically a broadband linear phase EQ, plus a special RMS-like accumulation tracker for frequency ranges (not amplitude). So it doesn't function by matching your track's response to a different track (though you can do that if desired), rather its main power is in giving your track "harmonic balancing" using that broad EQ... It's a little like EQing the entire track's 1/3 octave response averaged over time. It shows you obvious dips where the accumulation tracker determined your track's lacking - but then it's up do you whether you want to boost/cut, and where exactly to do it. The different approach takes some helpfile reading to get used to, but surprisingly good results! I got put on to it by Craig Anderton ages ago and been using it ever since.

In the meantime, it's also perfect for the simpler task of matching the volume of different tracks on an album. Can do it in almost braindead batch mode.

 

Last edited: Mar 18, 2017

@Satai, Har-Bal2 is definitely on my list to check into as the volume matching seems right up the alley of something I'm looking for. I post my findings here. Might be a week or so until I can get into it in more detail.

 

how about that one?

 

Haha, no, nothing automatic! I can't give up the control!

 

Volume matching? Use SPAN with K-14 metering, but best of all is still using ears. VU-meter is not bad, but K-standard works more like human ears so results are better, and LUFS metering even more so. However, there's no LUFS metering in SPAN, yet. I know all these standards by heart so you can be sure I'm not talking from the place where the sun doesn't shine.