24 v. 32 bit files from clients?

[Darius van H]

Someone wants to send me 32 bit files to master from.

Somewhere in the back of my drug-addled brain i seem to remember their was no benefit in 32 bit mixes to mastering, but perhaps i'm hopelessly behind the times.

Anyone got the technical hoo-haa on this??

Thanks in advance!

[bendermastering]

Originally Posted by Darius van H

Someone wants to send me 32 bit files to master from.

Somewhere in the back of my drug-addled brain i seem to remember their was no benefit in 32 bit mixes to mastering, but perhaps i'm hopelessly behind the times.

Anyone got the technical hoo-haa on this??

Thanks in advance!

There is really no benefit in sending a 32bit file but there's also no reason for you not to accept it that way... (Unless you have problems with bandwidth and the overall size of files)

[tonemaker]

I'm no mastering engineer, but I thought one of the major benefits of 32-bit float was that you could do tons of additive processing and not have to worry about clipping or degradation due to DSP. You can always normalise 32-bit float DOWN to just below 0 dBfs, right? This assumes that the software supports 32-bit float of course, and is less relevant for hardware of course. But it's nice to know it's there, right?

[Virtalahde]

24 or 32 makes no difference to me.

Sometimes I've had odd problems with 32bit files, and haven't been able to import them correctly. They might have been floating point files.

[Red Mastering]

Originally Posted by Virtalahde

24 or 32 makes no difference to me.

Sometimes I've had odd problems with 32bit files, and haven't been able to import them correctly. They might have been floating point files.

I thought it was only me
for some reason which I don't understand, had to convert them to 24 bit, as 32 wasn't working...

[wado1942]

There's different formats of 32-bit float so sometimes you have to alter the way the software interprets the files. Aside from that, I'm fine with 24-bit, I'm fine with 32-bit linear or float.... While I CAN do 64-bit float, I prefer not to deal with the huge files.

[wcb123]

I thought there were some benefits to this when it comes to dithering, I'm almost certain I've heard ME's mention this before.

Sent from my HTC Desire using Gearslutz.com

[JR465]

Originally Posted by wado1942

There's different formats of 32-bit float so sometimes you have to alter the way the software interprets the files. Aside from that, I'm fine with 24-bit, I'm fine with 32-bit linear or float.... While I CAN do 64-bit float, I prefer not to deal with the huge files.

What kind of software does one need to handle 32 bit files/ or convert to 24 bit?

[Cellotron]

32bit files offer no real advantage to me over 24bit. Any difference between the two is way below the noise floor, and if you are outputting to an analog process chain - pretty much every DAC converts this to 24bit fixed point in order to be able to output it - plus real world performance of even the best DAC out there is really only 22bits.

So 32bit float files generally represent just bigger file sizes for no actual sonic benefit. However I accept these without a problem as I just want clients to be able to send in whatever format is convenient for them.

Best regards,
Steve Berson

[wado1942]

Originally Posted by JR465

What kind of software does one need to handle 32 bit files/ or convert to 24 bit?

It's one of the few reasons I still keep a copy of Cool Edit Pro 2.0. It does 32-bit linear and three types of 32-bit float. I think most editors will do the standard 32-bit normalized float though.

[Darius van H]

Thanks!

I've been told that in the newest version of PT, you can export the mix to 32 bit for the first time. Could be quite a pain if it becomes fashionable and everybody starts sending me 96/32.

[Red Mastering]

Originally Posted by wcb123

I thought there were some benefits to this when it comes to dithering, I'm almost certain I've heard ME's mention this before.

Sent from my HTC Desire using Gearslutz.com

yep.
I think it was wavelab mastering presentation - Tischmeyer, if I am correct mentioned it

[huejahfink]

Not an especially technical answer here, but I see 32bit files submitted about once every 6 months! Never perceived any real benefit from that but it's a perfectly good format and I've got no worries receiving these from clients.
I'd not really considered it before reading this thread but for people whose last port of call in the production is 32 bit plugs on a DAW buss I don't see why it would be more beneficial to add a dither to it and/or change it to a 24bit with the exception of drivespace economy.
I tend to use them a lot more as a temporary file during the working process.
eg. on longer projects I often like to use them as a handy intermidiate format if last stop is with floating point plugs. Out > first pass and plugs> 32bitFP export / RX2 if necessary for any tidying > and then finally pull the whole lot of floating-point files back into cubase for the final level, sequencing and limit.

[da goose]

I get quite a lot of 32bit fp files and no problems here with Wavelab. Wavelab supports 32bit since IIRC v5 and i'm using v7 now.
If a client wants to send me 32bit, fine with me! I really don't see the problem here.

[Darius van H]

My problem with it is late delivery of files for mastering, due to huge data size, failed uploads, retrieval from FTP servers etc.!

[da goose]

To be honest, the differences in file size between 24 and 32 are not that big actually. And with todays internetspeeds and uploadservices like wetransfer i don't see the problem with file sizes at all.

[wado1942]

Originally Posted by huejahfink

on longer projects I often like to use them as a handy intermidiate format if last stop is with floating point plugs. Out > first pass and plugs> 32bitFP export / RX2 if necessary for any tidying > and then finally pull the whole lot of floating-point files back into cubase for the final level, sequencing and limit.

I do that too. I always send out flat dithered 24-bit files though if a client wants high-res media. I can download 32-bit fine but uploading something 75% the size that sounds essentially the same is fine by me.

[Jerry Tubb]

No problem accepting 32-bit files here, we get 'em quite often.

On the pitching side, Peak Pro, RX2, WL7 will play those.

I generally try to avoid converting to 24-bit unless needed.

As I could hear a slight difference in the sound, on my system.

A client Did stump us a couple years ago, when they showed up with W64 files,

Nothing we had would read 'em, but these days ...we ready.

JT

Originally Posted by wcb123

I thought there were some benefits to this when it comes to dithering, I'm almost certain I've heard ME's mention this before.

There are - I can almost always get a cleaner result dithering from 32-bit than 24-bit (MBIT+).

[Cellotron]

Hearing the difference on your system between two files when the only differences between them should be at least 16dB below the noise floor?? Methinks either something is broken in your software or converters or that you are victims of the placebo effect.

Again - for internal processing math if you are doing any digital processing on these files - then yes, the bigger the bit depth of the accumulator and arguably the bigger the bit depth of the resulting figure prior to being passed on to the next process in the chain the better. But for a static storage format? - especially for one that the first thing you do is play it out to a DAC?? Then you're really not getting an effective difference above 122dB worth of dynamic range - let alone above 144dB.

I'm also very curious also as to how many revision requests folks have ever had that required changing dithering options, or were done because people complained how the tails or very ends of fades sounded on their refs. For myself across my career is zero.

Again - I have no problem accepting 32bit files and find that usually the best policy is being able to take whatever is convenient for your client to provide without requiring them to do any conversions on their end(which have a potential for not being done correctly or optimally). But I question the reasons folks are providing as to why the believe there is an advantage to 32bit files over 24bit - particularly if playing them out to an analog process chain is the first thing done.

Best regards,
Steve Berson

Originally Posted by Cellotron

Hearing the difference on your system between two files when the only differences between them should be at least 16dB below the noise floor?? Methinks either something is broken in your software or converters or that you are victims of the placebo effect.

Not interested buddy.

[mirochandler]

Working with Samplitude here. You can set it up, that it only sends 24bit to the DA´s and working internal totally float.

I love the float files, because if you tell your clients that they should not use any effect on the Mix-Master-Channel, they sometimes forgot to set the output right and then you receive files with overs. No problem at all with 32bit float, overs does not matter.

When you use WinRAR for the Mix File it isn´t bigger than an 24 bit Files, it compresses all unused bits to zero disc space.

The best for me is 32bit float compressed as RAR File.

[dcollins]

Originally Posted by Cellotron

Hearing the difference on your system between two files when the only differences between them should be at least 16dB below the noise floor?? Methinks either something is broken in your software or converters or that you are victims of the placebo effect.

I thought mastering engineers had built-in immunity to the placebo effect?


DC

[Cellotron]

Originally Posted by soulstudios

Not interested buddy.

Not interested in what? Discussing how PCM Audio, DAW's and DAC's actually work?

Best regards,
Steve Berson

[Herbeck]

When questions like this pops up I recommend this little test.

60-second test

In my experience you can develop your ability to hear small differences in sound.
That's what perfect pitch ear training is all about.

[Bob Olhsson]

I've experimented with this quite a bit.

Unshaped dithering mixes to 24 sounds better to me than 32 bit float in the final 16 bit product. I really like the PSP Xenon because you can bypass the limiter while leaving the dither on.

[Cellotron]

Originally Posted by Herbeck

When questions like this pops up I recommend this little test.

60-second test

In my experience you can develop your ability to hear small differences in sound.
That's what perfect pitch ear training is all about.

Relative pitch, absolute pitch, or even the ability to quickly equate pitches to their frequencies in Hertz (a definite skill that would be a boon to a mastering engineer), have nothing to do with claims to being able to actually hear differences that should only exist below the noise floor of your electronics, monitoring system and/or room.

Best regards,
Steve Berson

[Herbeck]

I know that what I claim is a bit controversial but don't take my word for it, try it for yourself.

This little program is really useful too.
Pitch Train 2011 - Absolute Pitch music ear trainer


Cheers,

Herbeck

[Darius van H]

I love the smell of placebo in the morning......

[oldeanalogueguy]

Originally Posted by Darius van H

Someone wants to send me 32 bit files to master from.

Somewhere in the back of my drug-addled brain i seem to remember their was no benefit in 32 bit mixes to mastering, but perhaps i'm hopelessly behind the times.

Anyone got the technical hoo-haa on this??

Thanks in advance!

there are many benefits to 32 bit files
but if you have to truncate them to 24 for your software
then you lose most of the benefit