Posted on 20. Mar, 2016 by Daniel Duskin (Record Producer, Mixing & Mastering Engineer) in Mastering, Saturation, Ssh!
Disclaimer: In revealing this secret, I’m refraining from judgment of the Award Winning Mastering Engineers that service using this technique. The use of this technique may upset some when they learn of it—don’t blame the messenger.
Okay, here we go. I’m playing with fire by even discussing this outside of Professional Mastering circles. I’m about to blow the lid off this secret… Are you ready???
Mastering Engineers are digitally hard clipping our mixes!
You heard that right—after all the work we’ve done making headroom to avoid clipping before sending it off to mastering, Mastering Engineers are often clipping our music after you’ve worked so hard to avoid it. It can be certain that at least some of my readers are already mad after hearing this.
Of those who are, some of you are upset at the possibility of that happening to your mixes, and some of you may even feel the need to deny that it’s even happening.
I never thought that [engineers] would need to, or wish to clip [the converters].
How wrong I was…
— Dan Lavry (Lavry Engineering) [source]
For those who are upset at the notion that any professional Mastering Engineer might do such a thing, I can at least ease some—but not all—of your worries, by letting you know that they are only clipping the transients.
Clipping is still a very popular method of many pro mastering engineers and many rock records that you’ve listened to over the years have achieved their loudness this way. [...] It’s still about a thousand times better than any brickwall peak limiter if we’re talking about most rock and pop music.
— Steven Slate (Slate Digital) [source]
If you’re in the camp of people who feel that any digital clipping is a horrendously bad move—regardless of how many samples are clipped—nothing will make it acceptable to you that Mastering Engineers are doing this. I’ve seen many so shocked at this notion that they flat out refused to believe it.
This might seem an odd thing to say to those who see the loudness war as causing an obvious increase in clipping, but the notion of professionals using pure digital clipping (not limiting) regularly outside of a few horrid mistakes—such as Metallica’s Death Magnetic—seems unprofessional at most, or rare at best.
However, I’m not speaking of a few rare events, or simply unprofessional decisions. I’m speaking of a shift in what is the professional industry standard regarding mastering over the last decade or more, without exaggeration.
For those who can handle it—as right or wrong as it may be—not only is it happening, I can help explain the rational that’s being used to do it.
Note: Just because I can explain a rational, doesn’t mean that I subscribe to this rational. I’m refraining from making any judgement one way or the other.
If you need a mix to be extremely loud—either to compete with current music of that genre, or because the client asked for it—there are many ways to try and obtain that volume, and they give far from equal results.
There was a blind test on the Pro Sound Web forum comparing L2 vs clipping, and the majority of Mastering Engineers responding preferred the clipped sample.
If you take a heavily dynamic mix with lots of transient peaks, and try to use a limiter to attain a very competitive level with modern releases, the end result will often have a lot of pumping.
I usually just clip the input of my converters because I’ve been doing that for a long time and know how to do it properly. I find brickwall limiters really just squash up the sound and I try not to use them at all.
Remember, pumping caused by compression or limiting is reducing the volume. This statement is obvious, and technically correct. It still needs to be noted—sometimes we can miss the implications of the obvious.
To put it another way, when a limiter reduces the volume, the music is getting softer, and softer isn’t the goal when one is engaged in a loudness war.
Disclaimer: I’m not arguing in favor of the loudness war. I’m simply explaining what has been happening, and what the justifications for these actions have been.
Let’s not overlook some important details
I better mention a few things before I get into too much trouble here. If you talk to Mastering Engineers that won’t deny the use of clipping, they have some interesting points to add.
First of all, some Mastering Engineers claim to rarely use this technique. We’d have to pry into their rooms to really find out, but we’ll take them at their word anyhow. Second, many of the Mastering Engineers who use this technique claim to use it so lightly that it’s virtually harmless.
Either this is true, or it’s exaggerated minimization, but we’ll take them at their word anyhow.
The clipping i’m usually doing is generally just grabbing the half dozen or so highest kick or snare hits, and even then the parts that are clipping are a fraction of a millisecond long. I think there’s a big difference between that and flattening every beat.
—Scott Craggs (Old Colony Mastering)
Scott makes a great point. Clipping only some transients that last a 1-3 samples or so in length isn’t very destructive, and can reduce post-clip limiter pumping, resulting is less ear fatigue. You head that right, sometimes—if done correctly—clipping can result in less ear fatigue.
The importance of 16x oversampling is often discussed. However, I’ve noticed—as have many other Mastering Engineers—that when using plugin based clippers, results vary. Some sound better with oversampling on, while others sound better with oversampling off.
Because 16x over sampling is incredibly rare outside of hardware converters (you might notice a trend beginning here), the justification amongst many Mastering Engineers is that this clipping technique is only attainable with expensive outboard gear.
There are two possibilities here.
- Phenomenal (and phenomenally expensive) outboard mastering gear allows for clipping without sounding bad.
- Mastering Engineers need it to be true that expensive outboard gear is required for this technique, to protect their careers.
I personally feel that both of these hold some truth. While it’s understandably true that any professional will defend their job, it’s also true that oversampling on converters has exceeded that of oversampling in DSP (or native) plugin algorithms. A good question would be to ask how much of a difference does this make?
It should be mentioned, that while oversampling can reduce general aliasing, oversampling for setting a clip peak will force the clip threshold lower—further into the audible range. This can be very useful for preventing additional unaccepted clipping down the road—such as during playback through sub-par D/A converters, or during MP3/AAC conversion.
Argument: It’s Not True Digital Clipping..?
This goes deeper. Mastering Engineers have made yet another argument in favor of clipping expensive outboard mastering A/D equipment over using plugins for the same task. It’s been said that clipping converters clips the input circuit rather than the converter itself. One can’t help but wonder if that’s over-stated or over-reaching a bit. It seems to me that a well designed A/D converter would have a very low THD (total harmonic distortion), and therefore be less likely of clipping the analog circuit before the A/D process.
With that said, it just so happens that some A/D converters in the pro audio world have been designed to saturate the analog circuit before the converter.
Mastering A/D Converters
Universal Audio had a (now discontinued) A/D-D/A converter (the 2192) with analog circuitry that had a slow onset of analog harmonic distortion before the A/D converter was hard clipped. While I have not seen this unit used that much for mastering, UA published a blog article here that specifically noted this design.
We calibrate the 2192 to provide 18dB of headroom in the transition region from nominal to clipping, and gradually increase the harmonic bloom and class-A compression as the signal approaches digital clipping.
Universal Audio Webzine Article sep-2003
A more commonly used mastering A/D converter that contains processing that can produce a slower onset of harmonic distortion is the HEDD, manufactured by Crane Song. This hardware unit includes a digital process that emulates tape saturation.
The gold standard of A/D converters for mastering engineers is (arguably) the Lavry Gold AD122. Like the Crane Song, a digital process is utilized, called ‘Soft Saturation’—”a digital emulation of analog tape saturation which leaves the signal completely linear below the threshold.”
Challenging Expensive Gear — Can you do it ITB?
I have noticed—as have some other mastering engineers—that the slower onset via soft THD to clipping in some of these more expensive boxes doesn’t give a better result than pure digital clipping for some material. The question therefor becomes, why is it assumed that clipping via expensive outboard gear must be superior?
I’ve come to prefer a little bit of oversampled clipping in the box before the limiter. It just seems more transparent. A while back, I sent an ‘Eagle Eared’ client two versions of a master. One clipped at the A/D converter (Lavry Blue), and one clipped ITB. He greatly preferred the ITB version, and I agree.
—Scott Craggs (Old Colony Mastering)
Mastering Engineers Aren’t Stupid
These guys really know their stuff! I’m continuously amazed at the things they hear, and how true their opinions are. Sometimes I realize the factual nature of one Mastering Engineer’s opinion months later. Any honest and experienced engineer knows what I’m talking about. This has to be mentioned, because while I believe a medium dose of skepticism is healthy, an overdose of skepticism can hold you back. Don’t take anything in this article as supreme truth. I’m documenting what Mastering Engineers are doing, what opinions they are giving, what doubts some people have, and even documenting some refutations to those doubts.
I feel that I need to add the perspective that different music benefits more from different techniques, and different music is more or less tolerant to different techniques. Because of these variables, it becomes extremely difficult to make any broad-sweeping generalizations. That’s good news for you, and good news for Mastering Engineers, because it demonstrates that you may have the tools at your disposal to do it yourself, and it also demonstrates that seasoned Mastering Engineers have more real-world experience to better make those decisions based on the material.
What Should You Make Of All This?
Once again, it appears to remain true that it’s the talent & experience of the engineer that makes the difference, not the gear.
It’s The Engineer…
… Not The Gear!
It should be obvious, but sadly many wish not to believe it. For some reason, it’s easier to accept the notion that there must be some magical equipment being used that makes professional mixes and masters sound better than amateur mixes and masters.
While it’s true that some of that expensive gear sounds great, and some of the cheapo gear can give poor results, it’s rarely the case that this has very much to do with the reason professional mixes and masters sound better than amateur ones. If you’ve noticed that those big time engineers have a lot of expensive gear, and you’ve assumed this must be what’s making their work sound so good, you have it backwards. In most cases, they gained the opportunity to get that great gear because they do great work!
When you hire a professional engineer, you are paying for their skills, not their gear. There’s something attractive as an artist to think that it’s the gear—and not the engineer—to allow you to claim more of the personal talent that contributed to your total art. While that may feel good, the honesty that it takes to accept that a professional might be able to mix or master your music and make it sound even better can be the edge that gives you more of the sound that the greats have achieved, and your friends and colleagues haven’t.
How You Can Use This Information To Get Great Masters ITB!
The most important lesson here is that what sounds best wins—the end listener can’t tell what you used, or what settings you used, to get there.
It just so happens that digital clipping ITB sometimes / often sounds better than clipping expensive converters.
While digital clipping is generally considered a bad thing, digital clipping can be the most transparent (sonically invisible) type of clipping before it dips below the audible threshold. Because of this, digital clipping can be really useful as a tool to obtain extra level as the final stage before dither, or to reduce transient peaks pre-limiter for more level without—or with less—audible pumping.
You can experiment with clipping plugins that have oversampling options, but pick the setting that sounds best—don’t over-sample just to over-sample. While oversampling can reduce general aliasing, oversampling for setting a clip peak will force the clip threshold lower—further into the audible range. Because of this, it may be safer to set the clip threshold to -0.6dB without oversampling, which will eliminate any unwanted surprise clipping during MP3/AAC conversion, or during playback on sub-par D/A converters.
Whichever you choose, let your ears be your guide, and be sure to listen to the end result on as many mediums as you can. Gain Stage while clipping. This will allow you to find the audible threshold. If you can’t hear the clipping, don’t worry about it. It’s the point at which it becomes audible that it becomes questionable.
If volume is more important than quality, you could push it further into the audible range, but it’s not recommended. Be sure to check to see how audible your clipping is on small one-way speakers, as well as on your main monitors. Sometimes the clipping can be masked by high frequency information heard on your main monitors, but becomes apparent later (surprise!) when listening to your mix on smaller—and cheaper—speakers.
Recommended Plugins !!!
Has a 2X oversampling option, and a softclip option which can drastically reduce the harshness of the clipping but lower the clip threshold / onset. For a free plugin, you’d be a fool not to have this in your toolbox!
This plugin very accurately reproduces the sound of clipping a very high-end A/D converter. You can choose between limiting or clipping modes, and it also has a softclip option which can drastically reduce the harshness of the clipping but lower the clip threshold / onset.
This is a phenomenally advanced plugin! It’s a limiter, a clipper, DC offset fixer, dither, and much more. Plus, it sounds great! Never underestimate Voxengo plugins—they are often better than anything else out there, and this may be no exception!
Combines a 1-3 band Maximizer / Saturator with a clipper (which is labeled a limiter). The added 1-3 band saturation functionality can be very useful, has the ability to adjust shape and single or multi-band, and can be completely bypassed to utilize the clipper (which the plugin calls a limiter). The sound is very rich and analog.
Summary – Best Practices
- Check to see if the clipping is audible in small one-way speakers, and on your mains.
- Compare limiting to clipping, and choose which gives the best sonic results.
- If both limiting and clipping have sonic benefits, consider placing your limiter after your clipper (keep limiting to -1dB in most cases).
- Experiment with different clipping plugins, and oversampling settings, and always pick which one sounds the best.
- Consider keeping the max output to no higher than -0.3dB, and consider -0.6dB if you plan on converting to MP3 or AAC.
- If it’s still not working, the problem is likely somewhere in the mix. If you’re unsure, you can consult a professional mixing or mastering engineer.
- If you’re sending your music out to be mastered by a professional, please avoid all digital clipping and leave it to the professional—the more times your music is clipped, the worse it will sound.
Be sure to check out Secret Mastering Techniques They Don’t Want You To Know Part 2! — Coming Soon.