So, I started a project and didn't get anything on my master channel (fl studio), then I get some melody drums etc and it sounds PERFECT even with nothing on the master and going way over 0db, but if I just stick a limiter, a clipper or anyshit there to control the peaks it distorts the sound... and it makes sense... BUUUUUUUUUUUUUUUUUT, as long as I know there's no way to a digital signal get *really* over 0db (in real world) so how could I imitate the way my daw deals with the sound going over 0db (I presume it does something) to keep the song sounding the same when I export (because when i export it, it goes to really nothing holding the sound on so I really get clipping sound, but in fl studio, even if its actually over 0db it doesnt sound as clipping)

I uploaded it so you guys can see that it really doesnt go above 0db:

 

Last edited: Mar 27, 2021

You are clipping your converters. There's a big rabbit-hole of clipping converters, if you want to go down there.

 

what do you mean?

Thanks I got the problem solved

 

i dont have a clue what your saying either but you should always Always have nothing on your master and turn stuff down

 

Last edited: Mar 28, 2021

Converters are 24bit at most. Meaning 0dBFS is the maximum level. You are clipping your converter. There's tons of info on clipping different converters, if you want to tumble down that hole.
Since you seem to be new I would advise to learn gain-staging first (especially before exporting mix to be mastered). Then tumble down the hole.
Edit: Maybe I read your question wrong. Are you asking about 32bit FP internal processing (which has 1528dB of dynamic range and thus cannot "clip"). Or are you asking about clipping converters?

 

Last edited: Mar 27, 2021

If you have a way to see the output that is fed into your audio interface you might notice that there is a healthy pad between your DAC and Your DAW sometimes as high as 20dB.

If you want to hear what clipping your DAC sounds like you are going to have to be sure what signal level it is actually getting.

Different brands, different amount of headroom.

The solution to your problem is to turn down your master-fader and turn up your speakers, it is literally what is happening now anyway.

 

I think I wasnt clear enough
So, what I mean is: My project is going over 0db, but it doesnt sounds bad, sounds good actually. So I want to keep it sounding that way, if I export the project it sounds distorted, but in the daw it doesnt sound distorted at all, so, Im asking if someone knows how does FL Studio deals when the audio goes over 0db (because tecnically thats not even possible) so I guess its doing something to the audio that I cant see

 

It's because you're working at 24 or 32 bits but you are rendering to 16 bits.

 

Forget it, Baxter. People just don't get it that the 32-bit floating point sound that doesn't clip at the *output of your DAW*, whilst still in a digital form, goes out to your audio card and into a Digital to Analogue converter that converts this pristine, unclippable 32-bit FP sound into a very much clipped 24-bit analogue sound.

I keep telling that to people but they simply don't get it. *You can't possibly hear a 32-bit FP sample from your computer*. There's no such thing. Actually there are some 32-bit DA converters in some audio cards, but they're very rare and they are 32-bit integral, not FP, and they will still clip everything over digital zero because all DA converters have analogue limits [read the specs of your audio card, those are it].

So in conclusion: to get the same sound that everyone can hear in the same way you do by clipping it in an analogue way at your DA [which is kinda cool for some things], you have to put a clipper that sounds similar to your card's DA on the master output of your DAW, set at 24-bit output, no dithering, True Peak at about -1dB at least [because clipping the clipped sound will change it too much at the DA stage], and export it that way if you want your listeners to hear what you hear when you let your DA converter play a 32-bit FP non-clipped output from your DAW which then converts to 24-bit, and clips that audio in your DA converter in your audio card - that's the sound you *hear*.

Cheers!

p.s. you can also record the output of your audio card into another one to get the authentic recording. But record it at about -1dBFS because when you listen to it, and MP3 it, you will still get analogue peaks to 0 or even overs at the DA stage.

 

Last edited: Mar 29, 2021

As others have said, whatever happens, happens after it leaves FL.

It's entirely possible that it sounds better because it's a bit louder, 'cause that's a real thing, volume evens our ears' frequency response (provided you don't boost it to painful levels, that is). You might add slight saturation on your master (very slight) and lower the master so that it doesn't go over -0.5dB or thereabouts.

 

oh, thats too much information, I couldnt understand it yet, but youre saying that whats clipping is hardware and no software?

 

I mean, it sounds better because it doesnt sounds squished, it sounds like a soft clipper is there but in reallity it isnt

 

Nothing is clipping.

Your output has built in headroom, nobody is going to make an output with zero headroom because the engineers that design even the shittiest DACs are still engineers (even the ones with neckbeards.)

The only thing the FP of 32b-FP is doing is keeping your signal from squaring (beyond maximum value NAN.)

When you render as 24bit/16bit that is clipping digital clipping [squaring oscillation].

Reamp your signal and get it steadily above +20dBdBFS at least in your daw if you want to hear your DAC clipping.

If you want to clip your signal use a clipper on the bus or loopback and clip the input stage of a second interface.

All you are doing using the built in headroom on your interface & would get the same effect by turning down your masterfader by however many dB it is over and turning up your speakers by the same amount.

The rest is in your head.

I don't know how this can be explained any clearer. This is one of those times where the answer is so simple the only thing keeping you from seeing it is you.

 

(Note I don't know FL, so there are some assumptions here).

Modern DAWs represent the numbers in floating point numbers (for the mathematically inclined: https://en.wikipedia.org/wiki/Floating-point_arithmetic ) , for 64bit floating points this roughly translates being able to count each grain of sand in Sahara (slight, but not too much of an exaggerated number, you get the idea). So yeah, it's not the software that's to blame. What happens though is that at some point those numbers that can represent ridiculous amounts of accuracy have to be translated to real world. So your DAW processes those numbers, yes, even those that go beyond -0dBFS, but then it hands them to the audio driver which might not (or doesn't want to) handle numbers that big, so it has to be translated to 24bit audio. That cannot represent those things so everything above -0dB is clipped to -0dB. And even if the audio driver can handle it, the hardware can't.

It's kind of difficult to explain this without going into mathematics and how your computer and your AD/DA handles it, but you get the idea. Whatever happens inside the software can be "ideal" and it can handle going over, but it has to be translated to the real world at some point to be heard.

 

errrr.... but in reality it is. You're *hardware clipping*.

@JMOUTTON "Your output has built in headroom, nobody is going to make an output with zero headroom because the engineers that design even the shittiest DACs are still engineers (even the ones with neckbeards.)"

This is absolutely true and great to know, but as Kuuhaku said he hears soft clipping, he really does hear soft clipping. That can sound good, that's why the headroom is important: the more headroom - the more room for some nice soft clipping, but when you eventually fill that headroom with really loud output that won't sound nice at all because you'll get real hard digital, very harsh sounding clipping. The quality of this hardware clipping depends on the quality of your DAC.

 

Last edited: Mar 29, 2021

Not that I disagree but that's a bit harsh. It can be hard to understand the mathematical IEEE representation at the best of times for the best of us.

For instance, Python (the programming language, but most languages that deal with FP do this, hence you need epsilons to deal with things) does this:

Code:

1.2 - 1.0 = 0.199999999999999996

Makes sense, no?

Always record at less than 0dBFs. This form of clipping is called intersample peaking caused by the reconstruction of the digital signal at the DAC. I've heard that the magical number is -0.3dB but I keep it safe (like you) and keep it at -1 at most. Not really a problem if your signal is limited anyway.

 

@Kuuhaku

Sorry this got a bit too technical, but keep your channels and especially your master from going over (-1dB max).

If you like (and a lot of us do!) that slightly saturated sound do it in your DAW, or process it after exporting.