A Mastering Tip

[Anderton]

As many of you know, I do a lot of mastering (mostly digital...so there). As many of you also probably know, I like dynamics and try to give a "competitive" sound without smashing dynamics...and for that, I've developed various techniques.

 

But lately, on several cuts I found a way to make them sound less brittle and more organic: I put a steep shelf on the highs. Yup, I restrict frequency response. Basically, I do a sharp drop around 15kHz or so, and fool around with cutting above that until it sounds "right." In conjunction, I also add a teeny tiny bit of "exciter"-type processing on the high end, which adds a sheen without sounding brittle, and sounds smoother than the highs I reduce.

 

The difference is very, very subtle, but what's interesting is that after listening to it for a while, the cuts sound "wrong" when you switch out the very highest frequencies.

 

I don't do this with every track by any means, just the ones that need it (of course), but so far every client has said their track sounds much "sweeter" than they thought it would, and I think it relates to the highs.

[Ernest Buckley]

 

As many of you know, I do a lot of mastering (mostly digital...so there). As many of you also probably know, I like dynamics and try to give a "competitive" sound without smashing dynamics...and for that, I've developed various techniques. But lately, on several cuts I found a way to make them sound less brittle and more organic: I put a steep shelf on the highs. Yup, I restrict frequency response. Basically, I do a sharp drop around 15kHz or so, and fool around with cutting above that until it sounds "right." In conjunction, I also add a teeny tiny bit of "exciter"-type processing on the high end, which adds a sheen without sounding brittle, and sounds smoother than the highs I reduce. The difference is very, very subtle, but what's interesting is that after listening to it for a while, the cuts sound "wrong" when you switch out the very highest frequencies. I don't do this with every track by any means, just the ones that need it (of course), but so far every client has said their track sounds much "sweeter" than they thought it would, and I think it relates to the highs.

 

 

Mike Rivers mentioned this years ago before anyone else I can remember and it has always stayed with me... something along the lines that "what people love about analog is that its not as bright as digital". I`m paraphrasing. He went on to say that if we take out the top end it`ll sound better.

 

Simple Advice and very Effective.

 

I felt suddenly enlightened because I realized how many great sounding records don`t have much going on in the high end. Even mics and pres... my favorite ones lean on the dark side = more bottom end/less top. I usually run my mixes through some Avalon EQ and take out freqs in the upper range.

 

Its a good habit to get into... for me EQ is about taking things out rather than putting them in.

[roomjello]

 

As many of you know, I do a lot of mastering (mostly digital...so there). As many of you also probably know, I like dynamics and try to give a "competitive" sound without smashing dynamics...and for that, I've developed various techniques. But lately, on several cuts I found a way to make them sound less brittle and more organic: I put a steep shelf on the highs. Yup, I restrict frequency response. Basically, I do a sharp drop around 15kHz or so, and fool around with cutting above that until it sounds "right." In conjunction, I also add a teeny tiny bit of "exciter"-type processing on the high end, which adds a sheen without sounding brittle, and sounds smoother than the highs I reduce. The difference is very, very subtle, but what's interesting is that after listening to it for a while, the cuts sound "wrong" when you switch out the very highest frequencies. I don't do this with every track by any means, just the ones that need it (of course), but so far every client has said their track sounds much "sweeter" than they thought it would, and I think it relates to the highs.

 

 

Cool, i'm going to try this on some cymbal smashing tracks.

[Kotch]

I suggested cutting the highs on some forum once and got my arse handed to me.

 

"But duuuude, you can't cut 20K, it'll sound completely lifeless."

 

Yeah, right.

[roomjello]

and u47s' sound like {censored} because they only go to 15k

[BushmasterM4]

I always cut everything above 15k to 17k and below 40. The slope varies from time to time. Hell I HPF and LPF all individual tracks too. I too was told thats a NO-NO. There isnt any NO-NO's when it comes to music. Most consumer speakers cant reproduce stuff below or above those freq. anyway. Another nice trick is to duplicate the track your mastering and compress the second one alot !!! And set its level to taiste (blend the two). Like a cheap side-chain compression.

[Ernest Buckley]

I suggested cutting the highs on some forum once and got my arse handed to me. "But duuuude, you can't cut 20K, it'll sound completely lifeless." Yeah, right.

 

It must have been on Gearslutz.

 

Whatever you do, never admit you have any popular piece of gear on that site. You`ll get excommunicated.

[Lee Knight]

 

...I do a sharp drop around 15kHz or so, and fool around with cutting above that until it sounds "right."

 

 

Me too. Craig, I know you've used HarBal a bit. There are those genre curves in the Reference section. I typically use the "Rock" one for high end guidance. And it rolls off around 15k. As far as "exciter", I use the HarBal stereo widener a little. Not really an exciter though.

 

What type of exciter are you using?

[Lee Flier]

Yeah, I tend to cut the extreme highs and lows on individual tracks anyway. As has been pointed out, most people's listening equipment can't reproduce those frequencies anyway, so they're just going to take up headroom and produce distortion.

 

There are, in fact, analog recordings that are very bright (depending on the tape formulation and speed). Tape can certainly reproduce high frequencies. But it tends to do a gentle rolloff so you're still hearing the highs, but they aren't so strident that they distort and/or just annoy the crap out of you.

[roomjello]

They are generally saturated a bit too which helps.

 

I wonder, speaking of plugs and software....Does anyone have that UAD studer thing?

I would like to hear how it does this high end saturation job.

Can someone try putting an EQ pre the tape, say on a vocal or acoustic gtr and boost the highs with a shelf pretty hard, then hit the tape pretty hard and then report the results?

Does it work? Does it allow you more highs that otherwise without making you feel like there is a sword being pushed into your brain?

[WRGKMC]

I dont touch my highs much. They arent there when I get done mixing so I must be doing something during my mixing stage right.

 

Just remembered, I been using Waves multiband limiter in my mains lately befor mixdown. Its likely compressing things down above 10K with that. My finished mixes roll off above 15 to 17 K so I dont have to touch them.

 

I can see using exciters if you have really good monitors to dial them in correctly. Its basically re-synthasizing the highs without the associated noise trebble frequencies can carry with them. Any transistorized units used in the recording chain can contain ambiant transistor biasing noise background boosting the highs can produce. One track of that may not be audiable, but you get a few and they're cummulative on a mixdown. Chopping those down then resynthasizing them will work. You just have to be careful dialing it back in. You get too much coming through your tweeters and it can sound awful.

 

The only time I may have to selectively cut the highs in my mixes when I have that hiss happening. Some plugins can produce some high frequency hiss if theres too much gain dialed in. I usually try to nail it down there at track level whenever possible. Creating silence between vocal parts can help alot to cut down on that white noise too.

 

Dont rule out the use of a good noise reduction plugin either. Used mildly it can be as effective as any other technique for those high end frequencies. I sometimes use it before brickwalling cause I know the Waves L2 limiter I use will add some brightness to the high end. I can use the Waves noise reduction and and set it for listening to the difference (what the reduction is removing) and then dial out the exact amount of high frequency cut I may want to remove. When set that way you also hear all the digital background hiss and wash that will be removed.

 

When done right removing all that crappy digital wash and really make the sound much more transparent. You also have the benifit of the noise reduction responding dynamically so you retain the transient spikes of high frequency instruments, (mostly cymbals) where they should be vs just cutting everything down and loosing those dynamics from coming through.

 

A De-esser is another item that can be used is its the type where you can dial up those high frequencies. Its not a go to tool for me but in some cases it can work.

[temnov]

I use couple of UAD-2 Quad cards with all plugins. Studer A800 is fantastic. I worked with 24 track recorders in 80s and I can tell you UAD nailed it.

 

here is the link with some examples:

 

 

http://www.soundonsound.com/sos/mar11/articles/uadaudio.htm

 

 

Repeating my points and posts regarding modern plugins - it's all very different and very good today, technology is here.

 

As about highs in mastering - watch this, guy actually saying he adds some 27k (6 min 37 sec point) while mastering.

 

It's not about LPF of course, but interesting nevertheless.

 

[video=youtube;ttr2RcUc93E]

[flatfinger]

Something I read over at SOS (and haven't tried yet) is that in the mixing stage , you strap you best eq across the 2buss and put a bit of air to mix into ; This keeps you from using as much track or group buss eq ..( which , with a heavy track count , may be choosen for low cpu footprint and might have some grunge that will accumulate in the total mix)

[WRGKMC]

 

 

 

Mics cant pick up those frequencies, Most instruments wont produce them, playback equipment wont reproduce them, and human ears wouldnt hear them if they did. I wonder what benifit there might be by is by doing that other than entertaining dogs hearing abilities. The only reason I can think of is he's looking ahead and compensating some for the high frequency slope loss when the song gets compressed to MP3's or something, but even there i've never had any kind of benifits to boosting frequencies that have no musical sound there.

[Anderton]

i've never had any kind of benifits to boosting frequencies that have no musical sound there.

 

Maybe it's a really gentle filter, and by the time the skirts hit 15kHz (only an octave down, really), they're just the right amount of boost...and if he put the filter at 15kHz, there would be too much boost at lower frequencies due to the gentleness of the slope.

[Ernest Buckley]

 

Mics cant pick up those frequencies, Most instruments wont produce them, playback equipment wont reproduce them, and human ears wouldnt hear them if they did. I wonder what benifit there might be by is by doing that other than entertaining dogs hearing abilities. The only reason I can think of is he's looking ahead and compensating some for the high frequency slope loss when the song gets compressed to MP3's or something, but even there i've never had any kind of benifits to boosting frequencies that have no musical sound there.

 

 

Its an interesting topic for sure. The Avalon 737s I use for mixing and mastering have a 20k and 32k selection and on rare occasions I will tap them on. Now we may not be able to hear that high up but I do know when I`m boasting at 32k and I go too far, things get airy.

 

Its hard to describe, its almost as if 32k lightens or softens the mix. I don`t recall the last time I actually used that setting but I will experiment at times so whatever the case, yes, we cannot hear that high but we do sense something when we are boasting at those freqs.

[Anderton]

 

 

 

What you're hearing is the boosting of lower frequencies as well, as filters cover a range of frequencies. There are also phase issues at the lower ends of the skirts that may have an influence on the sound.

[Kotch]

 

 

 

Heh. It wasn't Gearslutz, but Gearslutz could easily have been the ultimate source of that sentiment.

 

I read Gearslutz, but I don't post. I wouldn't trust myself to be civil.

[Extreme Mixing]

Yes Craig. You can never trust the paint. I remember being at Bernie Grundman's and he boosted 2 db at 17K. Holy Jesus that made a tremendous difference--because it was so wide. It must have gone down to 3 or 4K. Sometimes you are better not to even know the center of the frequency that you are boosting. Just listen...

 

Steve

[WRGKMC]

Maybe it's a really gentle filter, and by the time the skirts hit 15kHz (only an octave down, really), they're just the right amount of boost...and if he put the filter at 15kHz, there would be too much boost at lower frequencies due to the gentleness of the slope.

 

Yea thats what I was thinking. It either makes the high end rolloff slope more gradual or it influences the harmonics. If the higher harmonics are lifted and not flatened, it may allow the lower harmonics to take full shape. I'm sure we've all seen the series of harmonic bumps that ride on root notes of a sine wave. If thie higher bumps are being removed through a filter, that filter isnt exactly a surgical cut. The lower order harmonics we can hear, probibly dont have as good a shape as they could. Its sort of like taking the upper layers off a pyramid and the layers left get blurred or rounded off in the process.

 

In your method, you remove the upper layers, then replace them with reconstructed harmonics developed from the root notes. This insures the high frequencies are uniform. In the trimming process, you remove all the uneven and distorted harmonics, as well as digital noise prevelant up there.

 

My sound cards arent super. I'm using M-Audio 1010LT cards. Frequencies below say 40hz die out and get fuzzy, and get about 18K on the high end recording at 24/48. I do know the sample rate does influence the abilities of the cards though. If I record at 16/44 I may only get 16K super clean on the high end. I suppose if I record at the max of 24/96 it may very well be clean up to 22K where the boards are rated. May tighten up the low end too.

 

When I get done with this batch or recordings I have on my drive I may revisit recording at the maximum sample rate again. The first generation of the recording at the present rate is good for what I do and its why I use it. But manipulating the wave files with plugins and processing does take its toll. I think I've gone as far as I can on the front end tracking and would need some serious upgrading to make some improvements there. from what I can tell A 24/48 mix can wind up sounding more like a 16/44 file before I even mix down to 16/44 when alot of processing with plugins is being used. The mastering process is needed to put the polish back on whats left. My instinct tells me if I do record at the cards maximum sample rate I may wind up with better results in the end. I'll only know by trying though.

[WRGKMC]

 

 

 

I'm sure its the quality of those frequencies that are being captured is the real key there. If the mics responce doesnt taper off till 32K you arent dealing with a weak and fuzzy waveform 20K. If I record a guitar track with an SM57, theres just nothing musically usable above its top frequencies. Using Craigs method can synthasize what that mic would have reproduced if it had a greater frequency responce.

 

There also may be something pleasing about tape emulation too. We all know tape adds smooth high frequency background noise from the tape itself. Even with Dolby you could hear that HF sound through your tweeters when theres dead air on a tape being played. Its the contrast between that high frequence and the other musical frequencies occuring in that same region that give them air. An analogy would be, seeing an image on stage with a black background. You have no depth perception because you cant see the back wall. you pretty much see the image as being two dimensional, height and width. Put a lit background scene in back of the person and now you have thre dimensionality. Height width and depth.

 

I realise sound and picture arent the same thing but that background tape hiss tricks the ears into hearing depth because thay now have contrast. This is also why the frequences dont have to be super high to give you that air. The thing that was hard to get adjusted to coming from analog was not having that tape hiss his to judge my high frequences mixing. I'd hear it and and know where my upper frequency limit was. I havent gone back to tape emulators to recreate that same hiss which they can do. But I have found some interesting uses with them none the less. Last mix I did I put a JS Ferox in the mains and tweaked my mix so I was edging the saturation like I would tracking analog tape pushing the tape limits for that warm drive. Then I'd remove the effect.

 

The results were very cool because I had some air I was missing and things were pumping where they should be. Later when I masteres the results were even sharper than I had been getting. I suppose I could do the same If I had top of the line gear, state of the art monitors and my ears were 40 years younger, I wouldnt have to rely on crutches like that to get a great mix. But I embrace anything that saves time and gets me there.

 

While running that plug, I could hear how much the low end would distort when I pushed the level up. A littel mix tweaking to the tracks and I eventually got the saturation even across the entire frequency spectrum. I was of course trusting the plugin specifications the makers decided to use to emulate tape when doint this. If the plugin was poor the results would have been poor. The ears of course are the final determining factor, but I have used several different plugins this way in the past as test tools vs using them as actual effects and had some interesting results.

 

I do have a habit of mixing to commercial quality, and often get it wrong. It may be hearing fatigue or just focusing on a particular part too much and it winds up dominating the mix a bit too much. I usually catch it in the next polishing session or after the first mastering attempt. So the tape emulation trick is simular to pushing individual EQ sliders into the red so each frequency is on the edge of distortion. A bass track kind of acts as a low frequency slider on an EQ, a guitar mids, cymbal highs etc. What I wind up with is a flater responce mix, something simular to what a multiband limiter will do without the artifacts. It surely wont work with all mixes, but its something you can experiment with.

 

Simply put the tape emulator in the mains, push the level up and listen for the first instrument that starts cracking distortion on the way up. Its likely its transients are a bit stronger than it should be. If you end up having the snare and vocals just punching the ceiling first followed quickly after by guitars kick and bass, you can be pretty sure you got a good flat balance to the instruments.

[roomjello]

 

I'm sure its the quality of those frequencies that are being captured is the real key there. If the mics responce doesnt taper off till 32K you arent dealing with a weak and fuzzy waveform 20K. If I record a guitar track with an SM57, theres just nothing musically usable above its top frequencies. Using Craigs method can synthasize what that mic would have reproduced if it had a greater frequency responce. There also may be something pleasing about tape emulation too. We all know tape adds smooth high frequency background noise from the tape itself. Even with Dolby you could hear that HF sound through your tweeters when theres dead air on a tape being played. Its the contrast between that high frequence and the other musical frequencies occuring in that same region that give them air. An analogy would be, seeing an image on stage with a black background. You have no depth perception because you cant see the back wall. you pretty much see the image as being two dimensional, height and width. Put a lit background scene in back of the person and now you have thre dimensionality. Height width and depth. I realise sound and picture arent the same thing but that background tape hiss tricks the ears into hearing depth because thay now have contrast. This is also why the frequences dont have to be super high to give you that air. The thing that was hard to get adjusted to coming from analog was not having that tape hiss his to judge my high frequences mixing. I'd hear it and and know where my upper frequency limit was. I havent gone back to tape emulators to recreate that same hiss which they can do. But I have found some interesting uses with them none the less. Last mix I did I put a JS Ferox in the mains and tweaked my mix so I was edging the saturation like I would tracking analog tape pushing the tape limits for that warm drive. Then I'd remove the effect. The results were very cool because I had some air I was missing and things were pumping where they should be. Later when I masteres the results were even sharper than I had been getting. I suppose I could do the same If I had top of the line gear, state of the art monitors and my ears were 40 years younger, I wouldnt have to rely on crutches like that to get a great mix. But I embrace anything that saves time and gets me there. While running that plug, I could hear how much the low end would distort when I pushed the level up. A littel mix tweaking to the tracks and I eventually got the saturation even across the entire frequency spectrum. I was of course trusting the plugin specifications the makers decided to use to emulate tape when doint this. If the plugin was poor the results would have been poor. The ears of course are the final determining factor, but I have used several different plugins this way in the past as test tools vs using them as actual effects and had some interesting results. I do have a habit of mixing to commercial quality, and often get it wrong. It may be hearing fatigue or just focusing on a particular part too much and it winds up dominating the mix a bit too much. I usually catch it in the next polishing session or after the first mastering attempt. So the tape emulation trick is simular to pushing individual EQ sliders into the red so each frequency is on the edge of distortion. A bass track kind of acts as a low frequency slider on an EQ, a guitar mids, cymbal highs etc. What I wind up with is a flater responce mix, something simular to what a multiband limiter will do without the artifacts. It surely wont work with all mixes, but its something you can experiment with. Simply put the tape emulator in the mains, push the level up and listen for the first instrument that starts cracking distortion on the way up. Its likely its transients are a bit stronger than it should be. If you end up having the snare and vocals just punching the ceiling first followed quickly after by guitars kick and bass, you can be pretty sure you got a good flat balance to the instruments.

 

 

A very interesting approach.

And so did you take the tape off in the end because the distortion wasn't pleasing at all? What if you just touched it, no benefits?

[WRGKMC]

 

A very interesting approach. And so did you take the tape off in the end because the distortion wasn't pleasing at all? What if you just touched it, no benefits?

 

 

I could have backed it down and just gone with it. I may still try it that way.

I used to have to do battel that edge in my analog days to get maximum gain vs clean headroom.

Adding it in intentionally kind of goes against my instinct.

 

In tape I'd push the levels on the tape right to the edge while monitoring the playback head output.

You'd get the high frequency harmonics saturating in a musical way that just sounded cool.

You would only do it on specific instruments though. It would put a smoother edge on driven guitars, round off snare transients, give a nice edge to vocals etc.

Other items like bass and kick you would often leave below the tape saturation level so that remain crystal clear in comparison.

Its a method of tweaking gain levels tracking.

 

Adding that saturation to the mains after track effects can cause some horrible effects. Reverbs, echos, Chorus, anything adding time based ambiance or air will create all kinds of beating an modulations that are dissonant to the ears. Its like using reverb in a guitar chain before overdrive. Distorting the trails is very unnatureal. Reminds me of hearing a song on the radio that is crushed badly. Maybe thats a sound someone likes, I dont know. If you do thats one way of getting it.

 

My whole thought process is to preserve clarity so by the time it gets beyond mastering, and gets hit by radio compression, MP3 compression or a 3" computer speaker theres still something left thats pleasing to the ear. Degrading artistically in a passage to give it an AM radio quality would be a whole different thing of course.

 

This is why I said I used it as a tool to get a balanced mix, using the saturation levels as a practical red level meter for the ears to detect overs, then remove it before mixdown.

 

If the mixdown was very dry you could use it on the mains. Or on Individual tracks before any other effects. That would provide more realistic track saturation.

I'll try one of my mixdowns both ways tonight and see if the results are helpful or harmful then maybe post the results.

[roomjello]

Yes i agree with all you said on the individual instrument benefits, and Tape can be a threat to mixdown if you hit it too hard. And yet there are all these guys with Ampex atr's out there using them specifically for that.

[WRGKMC]

True, true. It is a high fidelity analog method that preserves and even enhances the harmonics though. Using digital to get that just doesnt cut it the same. I heard there was a good tape emulator called Magneto that was part of s Steinberg DAW version that was discontinued. It was supposed to be a very good plugin for tape emulation. Maybe it would produce better results then the latest one I been messing with. Theres also some new plugins that are supposed to be good but they arent cheap.

 

I could pull out my decks and pass the mixdown through them but thet take up alot of valuable space and require alot of effort to use them. Sonar does have external effects routing so that would make it easier to use. I'm not using the best interfaces though so I dont think the extra D/A to A/D conversion would be super benificial. I still want to test it out though. I do have a bunch of decent hardware effects that I could also test out that way. Maybe when I take my next weeks vacation from my day job where I have the time to screw around with it. i also want to try using a mixer. The two 1010LT cards have 8 analog ins and 8 outs each so thats one way you can use them. Its another thing I havent done because My largest board is only 12 channels. I really need to sell off some of my old gear and upgrade.