I'm ..confused. Clipping digital

  • Thread starter Thread starter mixsit
  • Start date Start date
M

mixsit

Well-known member
I don't know where to put this, it's 'mastering, or mixing.. or..
Anywho, I'm a bit confused now. I get gain staging at tracking and mixing. I've heard nasty distortion, pretty much know where it's coming from when it happens. But pretty much took it on faith for the most part that going over 0dBFS was nasty. I didn't bump into the question much since I tended to work staged 'correctly'.
Lately I've noticed happening to push past zero, limiter on, limiter off, don't seem to be all that. A little different, sure.
Tonight I was working on a track that even though it had been compressed and limited, in the master project following a sparser less dense track, just seemed like it need to be pushed.
Non of these are real loud by today's norms, but with added gain this thing is now hitting (snare most likely) +2 (+4 on the intro..) and sounds.. fine.

Questions - re-education check please.
If it's +X' over 0 in the app, 32bit float, perhaps that is not a source of clipping right?
But, I'm listening through my D/A so it has to be clipping there right?
Now it occurred, well that's fine, maybe it's sounding ok on these D/A's, but what about other players (systems)?

But if I bounce it down (export it) now it's rendered max 0dBFS. And still sounds the same (I think.

I tried after adding about 3-4db to this mix that was already hitting -.5', the same TimeWorks limiter used on the mix previously. With or with out it, works fine. If I had to choose I'd actually go w/o, sounds as it it's a faster release and lets it breathe a bit more.
:facepalm: IDK, just looking for help to get straightened back out here.

Should I post a clip maybe?
 
The notion that digital clipping is horrible and must be avoided at all costs is a myth. Hard clipping in digital is no different than the hard clipping you get from overdriving most solid state audio devices. Personally I try to avoid that while recording, since I want everything as clean as possible. If I decide a track (or full mix) will benefit from subtle distortion, I'd rather add that while mixing when I can hear everything in context.

--Ethan
 
Yep. The hitting the 0dbfs ceiling isn't much different from a brickwall limiter. In fact, some ITB limiters really do just that - if it's louder than X, make it X.

And this kind of thing really isn't terribly noticeable unless it lasts for a while. Just clipping off the first sample or two from a snare hit, which already has a lot of high frequency energy, isn't going to be particularly noticeable. Frankly, most of the harmonics generated will be dumped at the anti-aliasing/reconstruction filter anyway.
 
mix at 0 db? no that how its done. I wish they never derived that vu meter off the the bit scale. Without having to write a paper about this here. that 0db is 100% as all of the bits. digital clipping is 0db on this scale wich transcribes to 100% DC to the DAC. If someone converts back to standard dbmV scale (VU) it would be +20 vu or so. Dac usually clip at +18m vu.

normal mixing should happen at 0db on the VU (or "T" on the Nordic Scale) or -10db Digital scale in DAWs. This will preserve all transients and high frequency details.

now you have the mix standardized and be able to do the next step: mastering

Vynal : no amplitude adjustment needed. -3 db slope eq at 25hz and 18K is desired but the cutter
engineer can determine if the eq needs to be applied.

CD: now cds dont have a fixed standard like they should. this is where the loudness war with record producers happen and alot of good music was sacrificed in this game.
commercial commercial pop has 6 db or less of dynamic range peaking at -4 db digital.
the reason why they slowly have decresed the dynamic range is that it is louder on cheap equipment, but blairy like on adiophile equipment. You'll also notice that the mixing style has changed to adopt this high output as the channels are mixed closer together (in the extreme, they are the same and advanced processing like ducking is used to justapose channels in the mix) unfortunatley this kills the creative process of the production because things are now focus around 'numbers' instead of 'feel'.

so if you master a good mix and it seems too soft, go back and mix the tracks closer together in amplitude if you need to.
 
Ok, but don't know if you read the post, but this was/is talking about intentional clipping and on the final mix.
Not really addressing here how loud the mixes are, the levels, comp, limiting (or clipping), that all pretty much followed where the tracks led.
 
mixsit said:
Ok, but don't know if you read the post, but this was/is talking about intentional clipping and on the final mix.
Not really addressing here how loud the mixes are, the levels, comp, limiting (or clipping), that all pretty much followed where the tracks led.
Click to expand...

that works a little different depending on the DAW and the pan law it uses
and depends on the mix engine. (pt10 and above and cubase6.5 and up limits instead of compresses overs for example)
 
Couple of questions here then- How would pan law figure into this?
And, some apps 'limt (w/o putting a limiter on).. as opposed to just 'clipping?
This is Sonar, but I've not heard of this.
 
And frankly, I think if the DAW is doing anything other than clipping off at 0db (when rendering to fixed decimal bit rate), it absolutely must be an optional setting somewhere in the menus or else be labeled in big letters somewhere that it becomes non-linear at some point on the meter.
 
ashcat_lt said:
And frankly, I think if the DAW is doing anything other than clipping off at 0db (when rendering to fixed decimal bit rate), it absolutely must be an optional setting somewhere in the menus or else be labeled in big letters somewhere that it becomes non-linear at some point on the meter.
Click to expand...

nope, sorry, there is 8-16 bits above the 32bits in the mixer (depending on DAW).
is there wasn't, people would crunch all over the place and send DC out of thier DAC.


btw it works that way with consoles too. And even the extreme with Neve. Thier's is hard set 10 db of headroom past the highest adjustment of the clip level light.

Even the cheap mixers have 3db above the clip light.

but truely clipping digital is nasty and destructive. and should be advoided.

true VU is logarithmic. That digital scale isn't but only will show you 32bits out of the 40 it has (for example). if you can change the meter to a true Vu meter type, you'll see it is not linear (logarithmic).

when I switch my scale in cuebase7 to nordic, the big "T" is there. lol
 
Dafuck you talkin about boy?

If I'm rendering to 24 bit, and what I'm rendering wants more than 24 bits on then I want my daw to clip it at 24 bit, not try to get cute and compress the thing. If I want to use compression to avoid clipping, that should be my deal. But it should not "save me from myself" without asking me politely, or at least telling me that it had this "feature" so I know WTF it's doing.
 
its how modern daws are programmed..... go back to protools 6 (for example) if you want to know how nasty true digital clipping is.
Now these days, they have a limiter of sorts at 0db threshold. and it release at 0db but that function is based on bit values instead of a calculation of signal interpolation.
true clip does happen at the msb (most significant bit) value equals the value of the bus but daws are now set up so you don't have that at rendering.
Mixing enviroments are significanly higher than that (40-64 bits) This is why your DAC hasn't blown up already bringing your signal over 0db.

in your practices, 0db digital is 23 bits when you mixdown 24bit format. if the msb is at 24bit of 24 bits it is registereed as an over wich equals logic 1 out of a dac which is DC voltage . you'll never get 24 as a value in your 24 bit file because the program won't let you.
 
Naw dude, my DAW (Reaper) just clips. Smatterfact, since I'm not the kind of guy who just takes these things on faith, I went and tested it. It's pretty easy to do, really. So I did. Ran a triangle wave through the thing at +6dbfs and rendered. The resulting file ramps nice and straight right up to the 0dbfs mark and then flattens out with a nice sharp corner and sits there till it comes back down the other side. There is absolutely no rounding off at the top to indicate any sort of compression. It just clips.

We are all well aware that the internal processing in the DAW can get way louder than 0dbfs without clipping. That is not at all what I'm talking about. When it finally comes time to render, 0dbfs is the ceiling, and I am glad to know that my DAW will let me go right up to it without messing with my audio unless I ask it to.

So what about your speaker killing DC voltage? It doesn't happen. All of our devices are AC coupled, meaning they can't actually pass DC current. But, you say, the squared off tops of the waves aren't flat long enough for the big caps to catch them. Okay, but in order for the DAC to reproduce those squared off tops it essentially has to create an infinite series of harmonics, and it can't. The anti-alias and reconstruction filters make it impossible. What comes out of the DAC and goes to the speakers cannot be actually flat on the top, it's wiggly as though built from a finite number of harmonics. Bottom line, there isn't really any DC hitting the speakers.

BTW, I've been doing this for a while now. You really want me to look "all the way back" to PT6? :thumbs up:
 
Well if nothing else this is some hopefully nice additional insight into the subject.
Going back around to the begining, what's been continuing to be interesting is playing with and w/o the normal '-0.5dBFS limiter.
 
yea...... rendering .... cuts it like what you said. but the outcome is really only compatible to digital media.
 
drtechno said:
yea...... rendering .... cuts it like what you said. but the outcome is really only compatible to digital media.
Click to expand...

Man, you have a really odd take on things. I will leave it to others to comment. You don't need to hear my opinion again....
 
jimmys69 said:
Man, you have a really odd take on things. I will leave it to others to comment. You don't need to hear my opinion again....
Click to expand...
I stopped leaving mine on his posts a while ago.

Ed...? :laughings:
 
trust you ears

if it sounds distorted, it is distorted.


don't trust your ears?

then stay from peaks on everything, you'll be fine.

why do you want to fly so close to the sun anyway?
you can stay closer to the ground, and view is so much better.
 
Guys (and ladies!) I need to hit this again. To clarify I'm really still grappling with this one part of the question-

I am clipping my mix +X' over 0dBFS -just snares are hitting this for the most part, and it's good -intentional.
Now if that was a limiter there doing this clipping, I presume that sound would be exactly what I capture on export. End of story
The heck of it is, and why I'm back with this, again tonight on another mix I'm liking it better w/o the limiter.

But, I presume now I am hearing the effect of my D/A's contribution to being hit with this track with 'overs.
So the question is - would this be a case where the result is going to be unique to a) the actual clipping of a D/A(?) and b) the effect also unique (to some degree) to this particular playback chain?
Last, maybe the only way to capture it would be D/A > A/D back in again?

I hope that's clear :o :)
 
You must log in or register to reply here.
Back
Top