clipping the master

[dan_s]

heey Funky, harrison has my favorite path station. hah.

I'm by no stretch a professional, so I know jack, but I try a great deal to make my shitty sounding music be of the professional volume and overall sound, lately I refuse to even use limiters at all, and i've been quite happy with my sound.

i guess theres a lot more to loudness then just raising the volume though. To that end, i put more effort into watching my rms levels, and bringing them up to get a better overall average volume, i believe that has helped mine music anyway.

[kilroy]

In general...clipping is a more transparent method of dealing with the transient peaks that eat up available audio headroom. Although the process is brutal, it is also mercifully quick...you just cut and run.


Since most emperical dynamics processors rely on some sort of time specific envelopes to control/moderate the leading edge of the waveform, one normally perceives greater audible deformation/shaping of the audio than what is noticed by simply "topping" the few samples in the fast rising amplitude of the transient portion. However, there are limits to what degree this is practical. As the clipping removes more and more samples, this relative transparency is lost. This becomes most noticeable when the clipping removes too much material in the ramping amplitude of the waveform immediately preceding the comparatively narrow, and somewhat "expendable", fast rising transient spikes.

Clipping is often viewed as distortion, and so it is. But in mathematical terms any audio deformation that significantly alters the dynamic or spectral makeup of the waveform, for good or ill, might also be termed similarly. So in essence, when we debate the artifact residue of clipping versus other dynamic control processes we are merely comparing one type of distortion with another.

A good strategy when one is looking for more RMS is to use dynamics processing to work on the body of the sound for some more "floor fill", and use clipping to "throw away" some peaks. This plays to the strongest virtues of either method. The trick, as always, is to find that elusive magic balance which renders the least apparent "damage" to the audio for the greatest return in loudness and perceived dynamics preservation.

Or, as a peer of mine so eloquently put it, "It always sounds better if you can find a good way to spread the junk around as much as possible."

[Kim Lajoie]

DVDs are all mastered so that the audio is almost constantly below 0dbfs

Music is almost always below 0dBFS too.

-Kim.

[Kim Lajoie]

I've written about clipping the master here (with audio examples), and more generally about mixing for loud masters here.

Man, I love not having to re-write the same thing every time.

-Kim.

[grillo]

Sorry Kim, i was just a bit surprised to hear that given as advice so i made this post. Shoulda known it was already an old trick in the book.

And, just for the record i went back to check and it is the producer masterclass with Dj Tayo on CM mag 121, had a little mixup in my head with the Layo and Bushwacka feature on FM and took a long time to sort that out

Also thanks for the Gclip suggestion, i'll go ahead and try it to see how it works.

[Kim Lajoie]

Don't be sorry!

-Kim.

[Shy]

kilroy: yeah, yeah, but the biggest question is will you ever reconsider offering your soft clipping algorithm/s in a software product? (that is, and reach a conclusion that you should )

[martian]

He must be talking about maximizing loudness with a limiter or perhaps putting some saturation (plugin) on the track as well. You should not digitally clip the master, I can see no advantage in that.

That being said, there have been many discussions here about clipping A/D converters (not your host's master bus), which is more of an advanced mastering engineer-type idea; but I doubt he's referring to that.

digital clipping for loudness on drum transient heavy material has been done for ages. i find it funny though how it remains a mainstream secret, not because its not shared but becuase no-one will believe you, and they wont try it themselves.

digital clipping a dance track 4db is almost always more transparent than any limiter. limiters smooth and pull down drum transients relative to the rest of the mix, the faster settings you use the less they do this and the more relative transient level is preserved, but at the expense of more distortion. digital clipping is like a limiter with instant attack/release, where you get maximum posible transient definition preservation at the expense of the most distortion, but if the distortion occours on a drum transient then this is masked by the drum sound (which is already comprised of unrelated harmonics, ie. noise). that said its only ok up to a point and having to do anything for loudness isnt ideal. clipping will give you the largest intersample peaks on playback, cured by 3-4x oversampling though.

also most people dont realise that if they dont leave headroom on their tracks before mp3 conversion then there will likely be digital clipping, so chances are you've already (not)heard it on your own music.

[m-ac]

Is there anything similar to Gclip for the intel mac?

[tetsuneko]

digital clipping, Huh? no wonder all new music sounds like shit

[tonAP]

.. so instead of 'hip-hop', a better name would be 'clip-hot'

don't you think so?

[Kim Lajoie]

.. so instead of 'hip-hop', a better name would be 'clip-hot'

Am I the only one who first misread that as something much ruder than it is?

-Kim.

[bduffy]

He must be talking about maximizing loudness with a limiter or perhaps putting some saturation (plugin) on the track as well. You should not digitally clip the master, I can see no advantage in that.

That being said, there have been many discussions here about clipping A/D converters (not your host's master bus), which is more of an advanced mastering engineer-type idea; but I doubt he's referring to that.

digital clipping for loudness on drum transient heavy material has been done for ages. i find it funny though how it remains a mainstream secret, not because its not shared but becuase no-one will believe you, and they wont try it themselves.

digital clipping a dance track 4db is almost always more transparent than any limiter. limiters smooth and pull down drum transients relative to the rest of the mix, the faster settings you use the less they do this and the more relative transient level is preserved, but at the expense of more distortion. digital clipping is like a limiter with instant attack/release, where you get maximum posible transient definition preservation at the expense of the most distortion, but if the distortion occours on a drum transient then this is masked by the drum sound (which is already comprised of unrelated harmonics, ie. noise). that said its only ok up to a point and having to do anything for loudness isnt ideal. clipping will give you the largest intersample peaks on playback, cured by 3-4x oversampling though.

also most people dont realise that if they dont leave headroom on their tracks before mp3 conversion then there will likely be digital clipping, so chances are you've already (not)heard it on your own music.

Sorry: not sure what your point is here. Are you talking about the A/D converter thing?

My first sentence was referring to pushing your fader up until it (digitally) distorts, which just doesn't sound like a good idea to me.

[Kim Lajoie]

My first sentence was referring to pushing your fader up until it (digitally) distorts, which just doesn't sound like a good idea to me.

That's pretty much what we're talking about here.

There's this false dichotomy[1] between digital saturation and analogue saturation. Somehow analogue saturation is beautiful everywhere and we should push everything into the red whenever we can, but digital saturation is untenable and we need to stay below 0dBFS as all costs.

You see that red light above your meter? When it lights up, it means "keep going, you're almost there".

Use your ears. Digital clipping doesn't sound nearly as bad as everyone claims.

-Kim.

[1] I just wanted to use the term "false dichotomy".

[tonAP]

My first sentence was referring to pushing your fader up until it (digitally) distorts, which just doesn't sound like a good idea to me.

..And also the OP @grillo talking:

he really meant hard clipping after everything

So what are we talking about?

@Kim:

Am I the only one who first misread that as something much ruder than it is? Laughing

I'm absolutely not rude in my heart but, since there is *NO LIMIT*
meaning no limit over 0dbfs?!? Then 'clip-hot' is more than adequate