Does it matter when mastering if your tune peaks above 0db

[wormcode]

Just 24 would work, most would probably convert down anyway

yeah it'll always end up at 16bit for cd anyhow. i just was referring to keeping everything at 32 during production to prevent clipping. i was watching a steinberg video and they recommended this due to:

cubase processing everything internally at 32bit, so less cpu load as no conversion is required

more headroom, as its practically impossible to clip 32bit

but for the master output, i've noticed that it will clip if i put the volumes up too loud, so i guess he was referring to internal clipping.

Yup truth. For what it's worth, I'm pretty sure every major DAW besides maybe Pro Tools (I've read conflicting reports on PT & 32, but don't use it...) is operating at 32 bit float internally now by default. Probably not dubturbo though.

[Shum]

Pro Tools does use 32 bit internal processing, indeed it can be as much as 64 bit depending on your setup.

[skimpi]

Just 24 would work, most would probably convert down anyway

yeah it'll always end up at 16bit for cd anyhow. i just was referring to keeping everything at 32 during production to prevent clipping. i was watching a steinberg video and they recommended this due to:

cubase processing everything internally at 32bit, so less cpu load as no conversion is required

more headroom, as its practically impossible to clip 32bit

but for the master output, i've noticed that it will clip if i put the volumes up too loud, so i guess he was referring to internal clipping.

im pretty sure it can still clip, having more bits doesnt mean you can have it louder, as in you can go over 0db or 0db is louder than with less bits, it just means you can have more dynamic range, due to the sound being represented with more bits. if you are working with 32bits and you turn it louder than 0db it will clip, it just may not be as noticeable until you turn it up much louder as when the top of the waves get clipped off, its not that much of a volume difference to the next step down in volume. really you should be able to mix perfectly fine with 16 bit, definately with 24 bit, if you have to use 32bit in order to not clip, then you are going somewhere wrong when mixing. having more bits means you can have everything lower in volume and sound better aswell with less noise/bit reduction.

[VirtualMark]

Hey, just to clear this up. I've just run some tests in cubase to check this. I got an 808 kick, increased the gain by 24db so it was clipping on the master. Exported it in 3 formats, 16, 24 and 32bit float. Both the 16 and 24 were clipped and i had therefore lost data from the original sample. The 32 bit was not clipped, and i was simply able to reduce the gain to get my original sound back. So the Steinberg advice is sound.

I agree that its not ideal to not be able to mix your tunes without it clipping. However, by exporting everything in 32bit float, you are preventing ANY clipping during the production process. So i can confidently answer that in Cubase, peaking above 0db does not matter, as long as you export in 32bit float. i can't say if this is true for other daws as i haven't tried this test on them.

[macc]

^ That's largely correct, but it's not a great habit to get into. Much better to simply keep your goalposts in the same place by avoiding clipping in the first place - in which case you may as well only export at 24-bit. 32-bit float and 24-bit are practically identical when you aren't clipping.

Another factor to consider is that it will sound different. If you are clipping the master output throughout the mix process, then exporting at 32-bit float and turning it down, it won't sound the same as it did before. You have been monitoring a constantly clipping signal, which is now not clipping. Lots of mix decisions will have been made based on how it sounded when clipping. Those decisions could now be mistaken...

There's also the question of plugins that operate with fixed point calculations - they will clip if you run them over 0dB.

To summarise simply, it won't clip in the DAW's mixer, but there's a 'bit bottleneck' at the master output (ie, the output DA convertor). Keeping it under that is very important for a host of reasons.


One other thing worth reiterating yet again here:

The benefit of 24-bit is that it allows you to turn stuff DOWN further without worrying about quality loss. So don't be scared to turn stuff DOWN, just be sure to export at 24-bit.

You just simply don't need to have anything anywhere near 0dB in a digital environment. It's completely unnecessary and usually detrimental to the signal.

[atticuh]

One other thing worth reiterating yet again here:

The benefit of 24-bit is that it allows you to turn stuff DOWN further without worrying about quality loss. So don't be scared to turn stuff DOWN, just be sure to export at 24-bit.

You just simply don't need to have anything anywhere near 0dB in a digital environment. It's completely unnecessary and usually detrimental to the signal.

Is this because in the analog realm the electrical strength of the signal plays a part in a good mix?

[VirtualMark]

^ That's largely correct, but it's not a great habit to get into. Much better to simply keep your goalposts in the same place by avoiding clipping in the first place - in which case you may as well only export at 24-bit. 32-bit float and 24-bit are practically identical when you aren't clipping.

Another factor to consider is that it will sound different. If you are clipping the master output throughout the mix process, then exporting at 32-bit float and turning it down, it won't sound the same as it did before. You have been monitoring a constantly clipping signal, which is now not clipping. Lots of mix decisions will have been made based on how it sounded when clipping. Those decisions could now be mistaken...

There's also the question of plugins that operate with fixed point calculations - they will clip if you run them over 0dB.

To summarise simply, it won't clip in the DAW's mixer, but there's a 'bit bottleneck' at the master output (ie, the output DA convertor). Keeping it under that is very important for a host of reasons.


One other thing worth reiterating yet again here:

The benefit of 24-bit is that it allows you to turn stuff DOWN further without worrying about quality loss. So don't be scared to turn stuff DOWN, just be sure to export at 24-bit.

You just simply don't need to have anything anywhere near 0dB in a digital environment. It's completely unnecessary and usually detrimental to the signal.

some interesting info there, thanks. for the record i do actually do as you said above, the only difference is that i keep it at 32 until the end. when i do a final export i do it in 24, but for bouncing parts i tend to keep it at 32 just to make sure nothing is lost, and also as it was recommended by steinberg as it apparently reduces cpu load slightly. i don't know how much but anything is welcome.

[macc]

Is this because in the analog realm the electrical strength of the signal plays a part in a good mix?

Eh? Perhaps you've been reading the internet too much...

In the analogue realm the main reason for keeping things relatively loud is to avoid noise floor concerns. This has carried over into the 'digital age' (vomit) with everyone trying to get as close to 0dB as possible to 'max out all the bits' (bollocks, pointless) in order to avoid the digital 'crunch' (audible, fact) you get with low resolution signals. The advent of 24-bit makes all that a pointless load of cack, as you can go extremely quiet and still be better than 16-bit.

If you mean about louder analogue signals causing 'pleasant' distortion, well, then maybe, but you can't judge it by the same standards. It doesn't work the same way.

[Divane]

Just 24 would work, most would probably convert down anyway

yeah it'll always end up at 16bit for cd anyhow. i just was referring to keeping everything at 32 during production to prevent clipping. i was watching a steinberg video and they recommended this due to:

cubase processing everything internally at 32bit, so less cpu load as no conversion is required

more headroom, as its practically impossible to clip 32bit

but for the master output, i've noticed that it will clip if i put the volumes up too loud, so i guess he was referring to internal clipping.

Alright so this means that if I select 32bit float in the project setup it will reduce my cpu load to some extend?

[VirtualMark]

according to the steinberg video i saw, this only applies to cubase tho. its because of the 32bit internal processing. i don't even know if its true for 64bit cubase, i'll try to find out how old the vid was i saw.

also, from steinberg on this site: https://www.steinberg.net/forum/viewtopic.php?p=88934

"We would recommend using 32 bit floating point for the file types in the project so that you take maximum advantage of Cubase's processing and mixing capabilities. This allows the any changes in the gain, plugins, mixing, panning, etc to be processed in their highest bit depth fidelity. I hope this information helps."

[oWLinDaylight]

. The advent of 24-bit makes all that a pointless load of cack, as you can go extremely quiet and still be better than 16-bit.

^^^This.
24 bits = 2^24 possible values = 16,777,216
16 bits = 2^16 possible values = 65,536
So you can be a 1/256 of the full volume in 24 bit audio and still be at cd quality.

[Sharmaji]

the simple, 1-word answer to OP is "yes."

re: 32-bit bounces... 32bit floating point is a really efficient operating environment for plugs, but at a bounce level it's a total waste of HDD space. if your mixes need to be bounced at 32bit to avoid audible distortion, you need to create better mixes.

[J-ohyeah-M]

And btw don't know if you heard but the moneyshot thread is shit.

Please tell me that was a joke...

I go through about 3 of the 50 odd pages and usually find fuck all apart from people arguing and giving advice about stuff they know nothing about. The stickies are the only things that put me of dsf

Come on man...I've learnt sooo much from the stickies (esp The MoneyShot Thread). And I'm always going back to them when I need to.
If you can't learn new things from them, or get a little help with problems that you're encountering, then I really don't know what to suggest.

Maybe stumble forward on your own and hope a production miracle happens???

[Earjax]

J-oh yeah, i think by stumble forward you mean experimentation which is pretty much how every producer learns the trade. Usually I'm looking for specific answers to a problem I encounter when I go on dsf, so maybe thats why I find the stickies not so good, I just don't find stuff that isn't relevant to me at that point in time interesting or useful :/

[oWLinDaylight]

i think by stumble forward you mean experimentation which is pretty much how every producer learns the trade.

But they also learn it by, you know... learning
And based on the fact that you were about to have a track mastered while it was peaking over 0 db makes it sound like you're stumbling.

[Earjax]

You can learn while experimenting no? anyways I don't want to argue, your probs right, but I'd rather learn at my own pace and in my own way to get the results I want, since everyone is different I've noticed dramatic improvements in my production and the sole things it has come from is practice and time

[Littlefoot]

One other thing worth reiterating yet again here:

The benefit of 24-bit is that it allows you to turn stuff DOWN further without worrying about quality loss. So don't be scared to turn stuff DOWN, just be sure to export at 24-bit.

Well said mr Macc . Sometimes the problem is with digital faders is sometimes people are afraid to turn em down, the advent of 24 bit recording should have wiped that out I guess..

[Earjax]

oh god...most embarrassing question of all time :/

[drift]

oh god...most embarrassing question of all time :/

nothing wrong with the original question, that's what the forums for.



this is where you went wrong

And btw don't know if you heard but the moneyshot thread is shit.

[bassinine]

i mean, if he's exporting stems the stems probably won't clip by themselves. then the ME can just lower the volume of each stem before they master.