From digido.com:
http://www.digido.com/portal/pmodule_id=11/pmdmode=fullscreen/pageadder_page_id=54/
Dear Bob,
Would you be so kind to give me info (or links) about 32-bit audioformat.
I'm interesting in what's the difference between 32-bit audio format and 32-bit floating point audio format, used in most audio processing software for PC.
Why are the mix engines of soft audio sequencers like Cubase VST and Cakewalk based on 32-bit floating format?
I've heard this format operates audio with level exceeding 0 dBFS. Is it possible?
What is dynamic range of such system?
I've spent a lot of time finding answers in WWW, but still unsuccessful.
Thank you in advance,
Fyodorov Alexander, sound engineer, Russia
Dear Fyodorov:
I am not a mathematician, but I will explain in the simplest words whatI know to be true. If you need a more mathematical explanation, you'll haveto crack a textbook!
Fixed point format is the language of the "outside world". That's becausein the real world, full scale is full scale---it represents the highestanalog value that can be encoded. 24 bit fixed point is the language ofthe outside world, and its encodable dynamic range is 144 dB. This is thehighest resolution allowed in the current AES/EBU transmission standard.
Some designers choose to use floating point chips in their internal calculations.This is very popular with native applications like Cubase because thecomputer CPUs that they are using like to talk floating point. I hear thatthe Power PC chip can work in either fixed or floating point, but for someesoteric reason, designers like to use its floating point capabilities.Probably because you can take an existing library of floating point code,and compile it for the Power PC very easily if you stay in floating point.
Once a number has been encoded into floating point, yes, it is true thatthe numbers can now represent overflow (above full scale) without overload,as well as smaller values than the 24th (LSB) of a fixed point number. Soyou end up with more internal dynamic range than 24 bit fixed point. Thisallows easy calculations for the "mathematically impaired". You don't haveto worry about overload when increasing gain, boosting a filter, summingchannels, etc. Many authorities also claim that this improves the internaldynamic range of the calculations (particularly filtering and compressionalgorithms) inside the processor. My distortion measurements comparing somedevices using floating point calculations against others doing fixed pointshow that with some kinds of filtering work, the floating point processorsshow less distortion. However, other designers working in fixed point producejust as low or lower distortion. Depends on the designer and his/her DSPtalent.
For example, in the most expensive and advanced processors, modern designersusing fixed point processing have progressed to internal calculationsusing "double precision" (48 bit in most cases), which doubles the internaldynamic range, and many authorities feel this performance produces bettersonic results than 32 bit floating point. This is at the cost of cyclesand power, but with more chips and more speed, it's not a big cost dealat this time.
But the whole "race" changes again when the floating point designers startusing 40 bit floating point. At that point, using equal types of algorithms,the two types of calculations likely produce equal sonic results, withoutquibbling. When working with double precision, it is very easy for a designerto design 24 dB (or more) internal headroom without losing meaningful dynamicrange, so when working with double precision, fixed point becomes as powerful(some say more powerful) than floating point.
However, a certain really talented designer working "only" in 32-bit floatingpoint produces excellent, low distortion results. My take on the matteris that designers concede that it takes a lot more effort (design talent)to prevent floating point work from giving you trouble than fixed point,perhaps because of rounding errors from calculation to calculation. Butone programming mistake, or a few cost-saving shortcuts, can ruin etherfixed or floating point work, especially if shortcuts are taken at themost critical time, when the final output number is converted to fixedpoint 24 bit at the end. If those numbers are not converted (and properlydithered to 24 bits) at that time, then the sound of the entire systemcan be compromised.
Bottom line: Don't be confused by the specs or the numbers or the claims. Distortion measurements may give us some clue as to why some systemssound better than others, but even distortion measurements don't tell thewhole thing, because they don't always reveal all the shortcuts that adesigner is taking under all circumstances. All other things being equal(and they never are!), it doesn't matter whether you're working in fixedor floating point. Only the results count. And there are real sonic differencesbetween platforms. "Cheap digital" still costs--it does not sound as goodas cheap analog.
Best wishes,
Bob Katz
