Actually all 32bit means is that it is working with 24bit signals at 32bit floating point.

Vegas already does this. It's nothing but a cunning marketing ploy by other software companies who's products do not process in floating point unless you select that option.

Not working at floating point means that it's possible to digitally clip signals internally.

There is no point.

Recording is done at the bit depth specified by the hardware, which is fixed point with all but on specific card that I am aware of that supports 32 bit floating point ASIO drivers.

The format we write to is always the defined open format - 8, 16, or 24 bit. This is only a file format and since the hardware is streaming fix bit data in the desired format, there is no overhead.

Internally all data is converted to 32 bit floating point when processing, so there is no "24 bit" math going on.

Peter

Maybe we should all buy 32 bit soundcards ! I saw a SOundblaster 32 once, but that was years ago. Duh.


geoff

PS It will be many years before we even get true 24 bit converters. And then we will still have nothing to record that takes advantage of that 144dB dynamic range, unless you want to hold a mic 1m fom the space shuttle on take-off....

you won't catch my valve mic under some space rocket thanks!

i can think of absolutely no reason to have 32 bits of dynamic range. unless it's 5 years from now and your a marketing dude trying to get people to replace their grungy 192/24 hardware and "upgrade" to 384/32 ;)

Look.

I can understand the overzealous audiophile back at my overclocking / casemodding forum going nuts about 32 bits ("because my mp3:s will sound better!!!") but why here? We're supposed to be pros, and pros know their shit, right? So why even MENTION 32-bit? All this accomplishes is scaring people who just got into the recording business into thinking "OMG, I just bought a 24-bit, 192 kHz, 8-in, 16-out megawoopidoocard to record my brother's two-bit trashmetal band, and now I've gotta go 32-BIT?!") which, needless to say, might be what crap companies like Creative wants, but you can't seriously believe that it's a widely-used, all-improving sound format because it isn't.

I still wanna meet the engineer who can tell the difference between 16-bit and 24-bit under normal circumstances - since we all know 24-bit is for the benefit of dynamic range during RECORDING, not playback. Hell, you could play me any normally mastered pop song in 8-bit and I wouldn't raise an eyebrow.

32-bit is totally overkill. Deal with it.

There's no point to 32 bit unless you're making recordings for dogs.

>>There's no point to 32 bit unless you're making recordings for dogs.
Well that is not true and I think you are confusing bit depth with sample rate.

some of you are missing something here: the 32bits that is mentioned actually has nothing to do with the quality the soundcard records at: as perter mentioned, there are no 32bit A/D converters available at the moment.

32bit refers to the maths which 24bit files or 16bit files are manipulated with.

If you do maths with a fixed point, then you get rounding errors. Small they may seem, but this means distortion... nasty distortion.

now, modern processors can operate using a floating point to prevent these mathmatical errors. This means, each time we add a plugin, we do it without quality loss if it operates at floating point.

applications that currently refer to 32bit, refer only to 32bit maths.

also, don't get confused with bit depth and sample rate.

bit depth affects dynamic range. 16bit has a dynamic range of 96db, 24bit has a dynamic range of 120db.

A/D converters often don't get close to this.

Sample Rate affects the frequency response. More importantly, processing at higher sample rates mean that changes in the higher and lower end will be clearer. Not because it adds new frequencies that we can here, but because processing distortions etc will be pushed into the frequencies we can't hear.

hope this is all clearer now.

In fact, why Soundforge have 32bits recording and not Vegas ??

I think that if Vegas would convert the bit depth of your recording card to 32bits directly when it is recording (like soudforge does), it would avoid to have to do it when it plays and let entire CPU ressource for FXs.

"Recording is done at the bit depth specified by the hardware, which is fixed point with all but on specific card that I am aware of that supports 32 bit floating point ASIO drivers."

Are you speaking of the out of production Yamaha DSPFactory (DS2416)?

>> since we all know 24-bit is for the benefit of dynamic range during RECORDING, not playback. Hell, you could play me any normally mastered pop song in 8-bit and I wouldn't raise an eyebrow.<<

Although *you* may not be able to hear the difference many of us can. I agree that a lot of pop, rock grunge, and hip-hop recordings are so over compressed that one can barely tell the difference in a 24/96 master and an mp3 file. But obviously this is not the case for other genres (or the some of above tracks mixed and mastered with some *musical* dynamic range left in it). I've personally done A/B testing of 16/24 bit mixes (24 bit dithered to 16 when burned to CD). With tracks that have a lot of acoustic instruments, including drums and percussion, clear vocals, ambient, new age, or classical types of music, the difference was at times dramatic. I wouldn't even think of using 16 bit at this point.

drbam

>>16bit has a dynamic range of 96db, 24bit has a dynamic range of 120db

I believe the theoretical dynamic range for 24 bit is 144 dB. To my knowledge, there are no converters in existence that get near that number. I'm pretty sure the more recent high-end 24 bit A/D's can give you something in the ballpark of 120 dB of *real world* useable resolution.

I can hear quite a difference between my old 16 bit A/D's and the new 24 bit ones I have. I also did some experiments converting 24 bit files to 16 bit with and without dithering and was fairly impressed with the benefits. I am fairly convinced that 16 bit / 44.1 CD's (if done right) are quite adequate for 99% of the public. I do feel that recording/mastering etc. done with 24 bit files is well worth it and will lead to better sounding 16/44.1 CD's in the end.

As far as higher sample rates go (96/192 KHz), I have yet to be convinced the audible differences are worth the file size and processor use penalties. Many do not realize that the newer over-sampling converters all but eliminate the artifacts caused by anti-alias filtering. Remember that doubling the sample rate will use twice as much CPU power to do the same thing. I would rather have more of my CPU available than use a high sample rate that provides very little benefit IMHO.

There are many out there who swear they can hear a huge difference between 48 KHz and 96 KHz sampling. There are also those who spend over $1000 on six foot silver conductor speaker cables for their home stereo and are convinced they hear a difference too.

yeah, my bad fishtank... it is 144db.

it good to see you actually *read* posts here, unlike some others!

I still record at 24bit 44.1khz ... simply because if I pushed higher at the moment I couldn't have enough tracks. I'm looking forward to adding some more DSP cards and trying 88.2Khz however.

Jason

The drawmer unit claims 130..
still a long way from 144

Once we have true 24 bit the issue of recording at 32 bit then becomes an issue.

As long as everyone is clear of the difference between recording bit depth and datapath and processing depth within the DAW

Forge has 32 bit recording as a file format, nothing more. There is NO realistic 32 bit converter, let alone a 32 bit floating point converter.

The overhead of converting from a fixed bit file to the internal floating point is trivial and will have little if any impact on CPU usage. Sample rate conversion is the expensive part, so you are better served to convert all your media to the project sample rate if you want to eek out additional CPU.

Peter