Pro Tools 10 - 32 Bit Floating Point - How It Helps?

[WestSideSpitterz]

can you record your voice and mix a beat with your voice to make a song?

[jasoncomeau92145]

major issue here. i tried your way of doing it. it works great! the problem with this vid is...you raised a clip to... above the clipping point from a clip that was already below clipping. again, your vid is correct...to a point. my problem is, ive imported other bands multi-tracks, where some of the clips were already well beyond past the clipping point. this is after importing the .wav audio. when i followed your steps to the imported (clipped) audio...nothing happened. correct me if im wrong.

[jasoncomeau92145]

if you want, i have no problem of making a video to show you what i mean. thanks in advance for your response. i would love a workaround to my problem, i have tons and tons of other bands multi-tracks that already have audio that was recorded incorrectly. (with clipping) and i cant have them re-record. im stuck. lol. so yeah, let me know, ill make a video to show you what i mean.

[fredlillo]

Merci pour ces informations !

Good job

Fred L (France)

[2khockeytv]

Cubase have this for years, just saying :)

[brownpride315]

@2khockeytv cubase sux just saying lol

[OpenSourceArtists]

If u are recording vocals would u want to record @ lower bit rate and then increase it during mixing/mastering stage? Or would u record @ 32bit to begin with?

[after11tv]

@OpenSourceArtists no interface that i know of expect maybe the hd stuff and other systems like that, will record in 32 bit float. at least not yet :p

[PovertyakaGargamel]

@after11tv I have the Focusrite Scarlett 8i6 and it gives me the option to record 32 bit float in Pro Tools 10

[ThaRealBummyDavis]

Besides all the arguing great video...

[90041hood]

"What is the colour of puberty?"

[Paul7Diesel]

Nice information in the there, thanks man.

[matthewtryba]

This is telling me that 32 bit floating is totally unnecessary if you know how to gainstage properly.

[stevenkurtscholz]

@matthewtryba That is completely correct. Its all about headroom we as humans cant hear in the end. And keep in mind when the digital audio get converted to be able to play out our speakers is will be 24 bit. Humans cant hear a difference after 20bits. No point in 32bit if you know how to make headroom.

[harvestsoulfly]

@matthewtryba Exactly what I was thinking.

[JSprayaEntertainment]

lol FL Studio and some others have had 32 bit floating point for a while

[993kRecords]

@JSprayaEntertainment along with terrible algorithms

[JSprayaEntertainment]

@993kRecords should learn more

[993kRecords]

@JSprayaEntertainment I know quite a bit sir.

[JSprayaEntertainment]

@993kRecords u talk about imageLine algorithms like you really have compared the pro tools ones to the FL studio ones , not to mention the 100 other things a PC can do to your audio , i own pro tools by the way , but most important its about how it sounds at max/clipping levels and there the same after you bounce down / render and in the programs ... both program wont make you famous if you suck , but One can if your good enough

[Starlite4321]

Yes, Yellosoul is correct. To say all this differently, when you record a 24 bit signal into a 32 bit PT track, or increase the bit depth of an existing 24 bit track to 32 bits, you get 8 unused bits to use from that point on within PT. Anything that happens within PT (summing/mixing, gain changes, processing...) gets to use the additional headroom provided by those bits (which is a LOT). There's almost nothing you can do that will result in clipping while in the box-which is a very big deal.

[Yellosoul]

Yuvalishes, no you wouldn't be able to bring back something that was clipped in the initial recording

[Benprogfuse]

is this because at 24 bit, there is no digital code for any gain that reaches above 144 DB, but at 32 bit float, though it can redline, it is able to generate a code so that when you normalize the signal, the peaks are not cut off and the peak portions of the audio are still there?

[bs1313a]

Nice!! Thanks.

[gives9]

This is really cool. Thanks for the nice example. G

[portessa]

Thanks for this great video - really puts 32-bit Floating Point into perspective.

[fridrikur]

Great tutorial thank you

[Starlite4321]

What can be done is, the magnitude of a signal recorded with 24 bit words can be increased until it saturates (signal is lost - sound breaks up), then record the SAME signal (the same way) with 32 bit words and increase ITS magnitude the same amount, and it WON'T saturate - because the 8 extra bits give the increased signal room to live in - its not literally lost. So anything that happens in PT set to 32bits that would have saturated at 24 bits, won't. That's what's new and valuable in PT10.

[Starlite4321]

The reason is that you can't ever recover lost information. When the waveform is clipped by excessive volume, that "top part" is literally lost and can't be recovered. What actually happened was, you saved off the original file in 32 bit format (which was never actually saturated, it was only being PLAYED too loud), brought it back in with the volume still set too high, then reduced the volume back down so it sounded correct again. The 32 bit element didn't play any role at all.

[Starlite4321]

Thanks for the video! Although your ultimate point is correct, the example you presented is a little misleading.  Viewers might think what you did was, increase the gain on a properly recorded track and cause it to saturate (increased its volume beyond what the original the track could handle), converted it to 32 bit (which CAN handle the larger volume) and saved it, brought it back into PT, and reconstructed the original signal because of the 32 bit word size. This is actually not possible.

[Chamilamusic]

hi thanks for the video tutorial..very help full.. but my equation is do i need my sound card 32bit support to work 32bit projects ? if any one know would be very help full.. (im using saffire pro 40)

[JoeOnKeys]

Pro Tools 5.3 LE and onwards also had floating point processing. What is the point of this discussion???

[ten21recordingstudio]

Well it's true .... you never stop learning. Great real world demo. Thanks Russ

[pestpest1]

very good

[Shushsound]

Great vid Russ

[yuvalishes]

Thanks for this video, It's about time we get some real world application to this 32 bit float feature. A quick question. If the original recording was clipping hard at the source you wouldn't be able to bring it "back from the dead" would you?

[musicman8942]

@yuvalishes For something like you're talking about I'd use Izotope rX2 and it's Declip function to bring it back to useable.

[leosaramago00]

@yuvalishes No, you couldn't. If hard clipping occurs before or at AD convertion stage, there's nothing you can do, you're dead for good. Keep in mind AD/DA converters are 24 bits fixed, not 32 bits Floating Point. Aim for something like -12dBfs peaking at input stage in the digital world. Output gain, compression and limiting towards loudness can(and should) be done at mastering stage.

[yuvalishes]

@leosaramago00 I understand and agree, thanks for the info.