64-bit floating precision - pros/cons?

Post general topics related to Cubase Pro 9, Cubase Artist 9 and Cubase Elements 9 here.
alexis
Grand Senior Member
Posts: 4248
Joined: Tue Dec 21, 2010 7:55 pm
Contact:

64-bit floating precision - pros/cons?

Post by alexis » Wed Nov 15, 2017 6:06 pm

I always heard that 32-bit was way more than we would ever need. Why 64-bit? And does it take up more CPU or storage?

Thanks-
Alexis

-Cubase "Safe Start Mode" (CTRL-ALT-SHIFT)
-Get variable-tempo audio to follow a grid here,
-Replacing freely-timed section into a variable tempo project

Cubase 9.0.20; i5-4570 3.2GHz, 16GB RAM; W10 Pro 64-bit on Samsung SSD 840 Pro 256GB; Seagate 1TB SATA 600 Audio; UR28M; Motif8; UAD-2 Solo; Jamstix 3.6; RevoicePro3.3; EZDrummer 2

GregOndo
New Member
Posts: 40
Joined: Thu Dec 16, 2010 12:03 am
Contact:

Re: 64-bit floating precision - pros/cons?

Post by GregOndo » Wed Nov 15, 2017 6:10 pm

It offers more precision in the audio engine. It will add a little more CPU usage but does not affect the file size.

All the best,

Greg
Cubase 8, Nuendo 6.5, WaveLab 8.5
Hexcore I7, 12 gigs RAM, Windows 7 64bit
MR816CSX (2), CC121, Houston
Lots of VSTis and External MIDI

alexis
Grand Senior Member
Posts: 4248
Joined: Tue Dec 21, 2010 7:55 pm
Contact:

Re: 64-bit floating precision - pros/cons?

Post by alexis » Wed Nov 15, 2017 6:14 pm

Thank you, Greg!

More precision ... what are the advantages of that (greater dynamic range? ... or is it something completely different than that)?

(Hobbyist more than engineer here, sorry!)
Alexis

-Cubase "Safe Start Mode" (CTRL-ALT-SHIFT)
-Get variable-tempo audio to follow a grid here,
-Replacing freely-timed section into a variable tempo project

Cubase 9.0.20; i5-4570 3.2GHz, 16GB RAM; W10 Pro 64-bit on Samsung SSD 840 Pro 256GB; Seagate 1TB SATA 600 Audio; UR28M; Motif8; UAD-2 Solo; Jamstix 3.6; RevoicePro3.3; EZDrummer 2

JimH
New Member
Posts: 13
Joined: Thu Dec 16, 2010 6:25 am
Contact:

Re: 64-bit floating precision - pros/cons?

Post by JimH » Wed Nov 15, 2017 6:25 pm

I always thought summing was done at 80-bit floating point anyway. So I don't know what 64-bit is about. Tell me why this isn't just a marketing number.
Cubase 9.0, Windows 10 64bit, Intel Core i7 920 / ASUS P6T, John Bowen Solaris, Korg Kronos, Yamaha S90ES, Hammond XK-3c, Minimoog, VAX77, Kurz K2000, etc.

studiodaz
Junior Member
Posts: 85
Joined: Fri Jan 07, 2011 9:56 pm
Contact:

Re: 64-bit floating precision - pros/cons?

Post by studiodaz » Wed Nov 15, 2017 6:25 pm

less rounding numbers so less errors when manipulating the audio signal. Every change to the signal will be closer to the original quality.
+++++++++++++++++++++++++++++++++++++++++++++

User avatar
djw
Senior Member
Posts: 1592
Joined: Sun Mar 15, 2015 2:49 am
Contact:

Re: 64-bit floating precision - pros/cons?

Post by djw » Wed Nov 15, 2017 8:07 pm

Do a blind test and see if you can hear a difference. ;p

It's cool to have though.
🥕
Specs: Latest versions of Cubase Pro, Dorico, WaveLab Pro, Windows 10, AMD Phenom II 1100T, 8GB RAM, AMD R9 390, YS CI1

roel
Member
Posts: 762
Joined: Wed Jun 27, 2012 2:20 pm
Contact:

Re: 64-bit floating precision - pros/cons?

Post by roel » Wed Nov 15, 2017 8:22 pm

on older machines (2010 and before) it's a no go terrain.
Way more resources are needed.
But ok, for me it's a testing environment so no prob.
Cubase 9.x, vienna ensemble pro 5,
Master:laptop Clevo, P771zm, i7 4790S, 16 gig, 2xSSD samsung 840pro/evo, 2x Samsung SSD850 +touch screen
Slave: none, will probably never be replaced because latest processors are powerfull enough
Sound: Motu Ultralite MKIII (usb), Focusrite 18i20, behringer 16 ch mixer
control: Synergy, novation zero SL
vsti: Steinberg, East West, Vienna, Native Instruments, Spectrasonics and a lot of others
hardware: S760, CD3000xl, SY77, studiologic 990, alesis Q88,

User avatar
zibelnik
New Member
Posts: 45
Joined: Mon Mar 31, 2014 11:17 pm
Contact:

Re: 64-bit floating precision - pros/cons?

Post by zibelnik » Wed Nov 15, 2017 8:29 pm

It sounds better, no question, clearer, deaper, more natural now, I am sure
Ljubljana, Slovenia
Cubase Pro 9.5|Win10 Pro on NMVE Samsung 950 Pro, 980ti graphics,
ASUS X99-DELUXE II | Intel Core i7-6850K | 64GB RAM
UAD Apollo 8p on thunderbolt + Symphony on SPDIF

alexis
Grand Senior Member
Posts: 4248
Joined: Tue Dec 21, 2010 7:55 pm
Contact:

Re: 64-bit floating precision - pros/cons?

Post by alexis » Wed Nov 15, 2017 8:38 pm

djw wrote:
Wed Nov 15, 2017 8:07 pm
Do a blind test and see if you can hear a difference. ;p

I haven't bought it, that's why I'm asking ;-)

djw wrote:
Wed Nov 15, 2017 8:07 pm

It's cool to have though.
I'm happy you think so! Why do you think that? Do you hear a difference? Are there other practical advantages?
Alexis

-Cubase "Safe Start Mode" (CTRL-ALT-SHIFT)
-Get variable-tempo audio to follow a grid here,
-Replacing freely-timed section into a variable tempo project

Cubase 9.0.20; i5-4570 3.2GHz, 16GB RAM; W10 Pro 64-bit on Samsung SSD 840 Pro 256GB; Seagate 1TB SATA 600 Audio; UR28M; Motif8; UAD-2 Solo; Jamstix 3.6; RevoicePro3.3; EZDrummer 2

User avatar
djw
Senior Member
Posts: 1592
Joined: Sun Mar 15, 2015 2:49 am
Contact:

Re: 64-bit floating precision - pros/cons?

Post by djw » Wed Nov 15, 2017 8:43 pm

alexis wrote:
Wed Nov 15, 2017 8:38 pm
djw wrote:
Wed Nov 15, 2017 8:07 pm
Do a blind test and see if you can hear a difference. ;p

I haven't bought it, that's why I'm asking ;-)

08)7'*
djw wrote:
Wed Nov 15, 2017 8:07 pm

It's cool to have though.
I'm happy you think so! Why do you think that? Do you hear a difference? Are there other practical advantages?
Plugins that calculate in 64-bit fp don't have to convert to and from 32-bit.

I can't see where you would hear a difference, but others would disagree with me.
🥕
Specs: Latest versions of Cubase Pro, Dorico, WaveLab Pro, Windows 10, AMD Phenom II 1100T, 8GB RAM, AMD R9 390, YS CI1

ultradust
Junior Member
Posts: 150
Joined: Tue Jan 08, 2013 12:37 am
Contact:

Re: 64-bit floating precision - pros/cons?

Post by ultradust » Wed Nov 22, 2017 3:56 am

From Philippe, developer of Steinberg WaveLab:
The interest of 64 bit float is not about "headroom" / "dynamic range"...

pro:

no need to convert between 32 bit and 64 bit float: 64 bit is needed by some plugins for their internal computations. 64 bit then means: small performance gain and no precision lost between succeeding 64 bit plugins.
Better audio precision when mixing audio signals. I explain this at the end of this message.
If audio devices ever go beyond 24 bit precision, 64 bit float will be needed (because 32 bit float means, in fact, 24 bit precision)

con:

requires more memory, which can mean a performance lost (more memory to move). But as soon as a sophisticated plugin is used, this one is likely to become the bottleneck, compared to the memory overhead. Therefore, this is a "relative con".
64 bit CPU instructions are as fast as 32 bit instructions, because the CPUs are 64 bit today. But certain rare instructions are faster with 32 bit float, because the CPU can conjugate 2 of them while in the same time, only one 64 bit instruction is performed (SIMD).

Now, an explanation about 32 bit float vs 64 bit float, for mixing.
While 32 bit float means in fact 24 bit precision, 64 bit float means in fact 48 bit precision. This means, far more precision.
I can illustrate this difference with elementary school maths (this is an analogy of what happens in reality).

Let's say samples can have only values 0, 1, 2, 3, 4, 5,...
Let's start with a sample that has value "3"
An audio gain of "divide by 2", is applied. We get the value "1.5", but this value is not allowed hence must be rounded, eg. the new value becomes 1.
Later another gain "multiply by 2" is applied. The new sample becomes "2".

Consequence: we started from value "3" and ended up with value "2", while the two gains should have cancelled each other.

When this kind of loss is performed multiple of times (complex mixing), then errors stack up.
The consequence is not dramatic, because some errors are (randomly) compensated by others (round-down / round-up), but this compensation actually means "digital fog" aka noise.

64 bit float processing pushes the digital fog far from the 24 bit domain. Hence a cleaner result at the end of the audio chain.

The difference 32/64 is therefore about "audio definition", if your ears can sensible enough. But that's another topic!
Philippe
viewtopic.php?f=244&t=125236

Magnus_N
Junior Member
Posts: 104
Joined: Fri Mar 04, 2016 11:26 am
Contact:

Re: 64-bit floating precision - pros/cons?

Post by Magnus_N » Wed Nov 22, 2017 8:49 am

Thanks ultradust, for sharing Philippe's explanation. Very informative.

/Magnus
Cubase Pro 9.030, Windows 10, RME HDSP 9632, i7-6700K CPU @ 4.00 GHz, 32 GB RAM

Carvin Man
Junior Member
Posts: 55
Joined: Thu Oct 12, 2017 10:47 am
Contact:

Re: 64-bit floating precision - pros/cons?

Post by Carvin Man » Wed Nov 22, 2017 11:11 am

Magnus_N wrote:
Wed Nov 22, 2017 8:49 am
Thanks ultradust, for sharing Philippe's explanation. Very informative.

/Magnus
Even though no one can hear a difference.... Have a read here:
viewtopic.php?p=687057#p687057

Then have a read on the whole thread, and the links provided:
viewtopic.php?f=240&t=88517

User avatar
Denis van der Velde
Junior Member
Posts: 155
Joined: Fri Jun 23, 2017 10:41 am
Location: Netherlands
Contact:

Re: 64-bit floating precision - pros/cons?

Post by Denis van der Velde » Wed Nov 22, 2017 12:37 pm

The main thing is ? It it affecting processor speed ? It could be 64 Bit is better on a 64 bit machine. And 32bit is more processor heavy because of conversions to and from 64bit processing. So anyone did a speed test ?

To me 32bit processing is ok, i leave it for now. Because no one can hear any difference between 32bit and 64bit..
Denis van der Velde
http://www.curioza.com
AAMS Auto Audio Mastering System for Windows

User avatar
MixterRader
New Member
Posts: 24
Joined: Mon Dec 12, 2016 10:54 pm
Contact:

Re: 64-bit floating precision - pros/cons?

Post by MixterRader » Wed Nov 22, 2017 1:34 pm

Unfortunately, none of my Waves plugins seem to work at 64 bit.
Cubase Elements 9.5
Windows 10 64-bit
Intel i7-7700K 4.20 GHZ
16 GB ram
RME Hammerfall HDSP 9632 + Totalmix

My Website | My Facebook Page | My Soundcloud Page

User avatar
Denis van der Velde
Junior Member
Posts: 155
Joined: Fri Jun 23, 2017 10:41 am
Location: Netherlands
Contact:

Re: 64-bit floating precision - pros/cons?

Post by Denis van der Velde » Wed Nov 22, 2017 1:56 pm

MixterRader wrote:
Wed Nov 22, 2017 1:34 pm
Unfortunately, none of my Waves plugins seem to work at 64 bit.
That is why i leave it on 32bit, because if the plugins are actually 32bit processing, they must have 64bit processing. If not i did not know what happens. Maybe more can test this. I assume you have waves x64 plugins ? (not using Jbridge?).
Denis van der Velde
http://www.curioza.com
AAMS Auto Audio Mastering System for Windows

Dallon426
Member
Posts: 229
Joined: Sat May 04, 2013 8:00 pm

Re: 64-bit floating precision - pros/cons?

Post by Dallon426 » Thu Nov 23, 2017 7:14 am

I don't think anyone can tell the difference in audio quality. You'd have to be super human.
I've got an album here if you like Folk Rock
https://dallonghan.bandcamp.com/album/shrink
also more music on spotify, itunes, etc


I7 7700K
AsRock z270 itx fatal1ty
16GB DDR4 Corsair 3466mhz
1 512 GB M.2, Nvme
1 SSD 240GB (OS and apps drive)
Noctua nhl-9i heatsink
Inwin Chopin
Windows 10 64bit Pro v 1703
Cubase Artist 9 (updated)
Motu 828x (Thunderbolt)

vinark
Member
Posts: 376
Joined: Wed Dec 15, 2010 5:24 pm
Contact:

Re: 64-bit floating precision - pros/cons?

Post by vinark » Thu Nov 23, 2017 10:13 am

Dallon426 wrote:
Thu Nov 23, 2017 7:14 am
I don't think anyone can tell the difference in audio quality. You'd have to be super human.
And your audio interface, amps and speakers need to have a dynamic range and noisefloor beyond (-)144db.

User avatar
Raphie
Senior Member
Posts: 1318
Joined: Thu Dec 16, 2010 5:20 pm
Location: Netherlands
Contact:

Re: 64-bit floating precision - pros/cons?

Post by Raphie » Thu Nov 23, 2017 11:13 am

64bit mixing IS NOT about dynamic range, it’s about roundig accuracy and adapting word length.
Analogue Mastering
MSI raider X99 - Intel i7 6800k - MSI Gaming X 1070GTX 8GB - OCZ RD400 nvme SSD - 16GB DDR4-3000
Windows 10 x64 up to date - Cubase Pro 9.5x - Wavelab Pro 9.5x
RME MadiFX and racks full of outboard

Everything you need to know about remote control editors

Buyaicia
New Member
Posts: 35
Joined: Mon Aug 31, 2015 5:10 am
Location: Catalunya
Contact:

Re: 64-bit floating precision - pros/cons?

Post by Buyaicia » Thu Nov 23, 2017 11:14 am

Ultradust,
I think that using integers to demonstrate the error in floating point is not a good way.
I will use another example
I start with value 1
Divide by 3 equals 0.33333333333333333333333333333333333333333333333333333333333333333333333333333
32bit floating point precision will cut at 8th decimal value
You will get 0.3333333
Multiply by 3
you get 0.9999999
The error is 0.0000001 of expected value.
This error can be measured, but, can it be heard?
With 64bit precision it will cut at 17th decimal value and the error will be:
0.00000000000000001 of the expected value.

Regards,
Lluís

ultradust
Junior Member
Posts: 150
Joined: Tue Jan 08, 2013 12:37 am
Contact:

Re: 64-bit floating precision - pros/cons?

Post by ultradust » Thu Nov 23, 2017 11:35 am

Lluís et al,

In practice, DAW audio engines and plugins process audio in widely varying - often extreme - ways. Yes, in abstract, a single truncation might not affect much. But multiply the lack of precision across a hundred tracks in an electroacoustic sound design project, each with multiple plugins (perhaps being converted between 32-bit and 64-bit processing). Multiply the lack of precision in and out of multiple components of each processing chain. You can imagine that in certain situations, the sheer number of calculations and the way the processors are used could amplify rounding errors significantly and cause deviations further and further from "accurate."

Whether you would like the resultant sounds better or not is a matter of preference. You could arguably prefer the 32-bit-processed results. But the 64-bit processing capabilities built into Cubase and WaveLab allow our work to be more precise. Plus, it just makes sense to keep 64-bit plugin processing in 64-bit rather than convert back and forth needlessly.

Dallon426 and vinark, I agree with Raphie. This isn't about hearing an increased dynamic range above a noise floor due to 64-bit processing. It's about maintaining higher precision at more points along the processing path.

Dallon426
Member
Posts: 229
Joined: Sat May 04, 2013 8:00 pm

Re: 64-bit floating precision - pros/cons?

Post by Dallon426 » Thu Nov 23, 2017 12:11 pm

What does is matter if it all sounds the same? We're talking about sound.
I've got an album here if you like Folk Rock
https://dallonghan.bandcamp.com/album/shrink
also more music on spotify, itunes, etc


I7 7700K
AsRock z270 itx fatal1ty
16GB DDR4 Corsair 3466mhz
1 512 GB M.2, Nvme
1 SSD 240GB (OS and apps drive)
Noctua nhl-9i heatsink
Inwin Chopin
Windows 10 64bit Pro v 1703
Cubase Artist 9 (updated)
Motu 828x (Thunderbolt)

vinark
Member
Posts: 376
Joined: Wed Dec 15, 2010 5:24 pm
Contact:

Re: 64-bit floating precision - pros/cons?

Post by vinark » Thu Nov 23, 2017 12:16 pm

Raphie wrote:
Thu Nov 23, 2017 11:13 am
64bit mixing IS NOT about dynamic range, it’s about roundig accuracy and adapting word length.
And what do you think rounding errors cause? (Distortion) Noise! And I meant dynamic range in the sense of usable range above the noise floor.
Still this update is meant for streamlining the audio processing not improved audio quality as stated by Fredo.

User avatar
Raphie
Senior Member
Posts: 1318
Joined: Thu Dec 16, 2010 5:20 pm
Location: Netherlands
Contact:

Re: 64-bit floating precision - pros/cons?

Post by Raphie » Thu Nov 23, 2017 12:47 pm

ultradust wrote:
Thu Nov 23, 2017 11:35 am
Lluís et al,

In practice, DAW audio engines and plugins process audio in widely varying - often extreme - ways. Yes, in abstract, a single truncation might not affect much. But multiply the lack of precision across a hundred tracks in an electroacoustic sound design project, each with multiple plugins (perhaps being converted between 32-bit and 64-bit processing). Multiply the lack of precision in and out of multiple components of each processing chain. You can imagine that in certain situations, the sheer number of calculations and the way the processors are used could amplify rounding errors significantly and cause deviations further and further from "accurate."

Whether you would like the resultant sounds better or not is a matter of preference. You could arguably prefer the 32-bit-processed results. But the 64-bit processing capabilities built into Cubase and WaveLab allow our work to be more precise. Plus, it just makes sense to keep 64-bit plugin processing in 64-bit rather than convert back and forth needlessly.

Dallon426 and vinark, I agree with Raphie. This isn't about hearing an increased dynamic range above a noise floor due to 64-bit processing. It's about maintaining higher precision at more points along the processing path.
Yep, this is a good summary, PG and Fredo summed it up nicely from a accuracy and streaming processes perspective.
people confuse wordlenght, bitdept, samperate
Analogue Mastering
MSI raider X99 - Intel i7 6800k - MSI Gaming X 1070GTX 8GB - OCZ RD400 nvme SSD - 16GB DDR4-3000
Windows 10 x64 up to date - Cubase Pro 9.5x - Wavelab Pro 9.5x
RME MadiFX and racks full of outboard

Everything you need to know about remote control editors

User avatar
peakae
Senior Member
Posts: 2139
Joined: Sat Jan 05, 2013 8:15 pm
Location: Bedroom
Contact:

Re: 64-bit floating precision - pros/cons?

Post by peakae » Thu Nov 23, 2017 12:55 pm

True, on the audible side of things I would not claim that there is no difference.
Just that I in my room with my speakers dont hear anything conclusive, it feels better but that could easily be something I want to hear.
Cubase Pro 9.5, Wavelab Elements 9, I7 3770K , win10x64, 16Gb Ram, RME Raydat, Steinberg MR816x, Motu 828mkII, Behringer ADA8200, Yamaha moXF6, Steinberg UR242, Yamaha THR 10, Cambridge Audio DacMagic Plus, CMC TP, CMC CH.

Post Reply

Return to “General”

Who is online

Users browsing this forum: beatpete, JMCecil, Vendimion and 9 guests