are new interfaces really any better?

[PorterhouseMusic]

And often its about noise and gain

Good point.

 

[TalismanRich]

yea Dave , your comment made me think from my work experience in TV's and the HD thing and flat panels happened, the low end does come up because the factory's stop making the old stuff and parts made obsolete. The newer HD parts are all that's available so the cheaper end products have to upgrade and use the new available chips.

Like the CRT TV's., Flat Panels took over so the assembly lines and marketing dumped CRT and the CRT TV's were just dead technology, not because they didn't work but because of the "chain" of parts were dumped and so it was over. I can still recall the last dinosaur CRT on the shelf of WalMarts, and stores ...watching the CRT die after a 50 yr rein or whatever it was. There's no more 480 TV now it was 720 then 1080 then more and more beyond a human ability really. Our group made alternative technology to compete against LED ...so we died off, like Plasma and the other losers....LED was just too good. Samsung was the largest customer playing ALL the technologys but eventually when LED showed to be the best, Samsung and the big customers dumped their Plasma, DLP, and other tech lines...LED HDTV...60inch for $300!! lol....used to be $12000 in the beginning.

My son-in-law took my VERY old Sony 27" Trinitron CRT tv to use with his old game systems. It still worked after sitting in the basement for almost 20 years. I paid almost $800 for it, probably 40 years ago, replaced it with a 42" plasma that lasted 18 years. It had stereo audio output, which fed my old Sony receiver and Marantz 5G speakers. Those old VHS movies were great!!

I paid about the same for my new 55" OLED 4K TV, but it just doesn't have the "warmth" of my old CRT and plasma. Those things would keep the den warm on a cold winters day.

 

[ecc83]

"32bit float would be worthwhile since it means you don't need to worry about overloading, and your inherent noise floor will be essentially zero. Plus audio processing is moving towards 32 bit processing."

I freely admit I don't quite get that ^? I do understand floating bit arithmetic, logs and ****! Did it at school but how can 32 BF change the overload point of say a mic pre amp? That is determined by the supply rails and the gain you set. I understand the REALLY top end stuff uses two preamps in parallel and two A/Ds and some very clever circuitry switches (seamlessly?) between high and low gain. But the technique has drifted down to cheaper stuff...not so sophisticated surely?

24 bits has a DIGITAL dynamic range of ~144dB and nobody, even the best of them can get their analogue noise floor within 20dB of that. Those that do best are a hugely impressive design as not only is analogue noise very low but enormous care has to be taken to keep digital and analogue earth paths apart. PSU design is equally important to get a noise floor better than -120dBFS.

I would be most grateful if some technical bod could explain 32 bit float and how it "appears" to beat physics?


Dave.

 

[rob aylestone]

I don't think it impacts the preamp, Dave - the rules for that remain the same - but it assists the mix bus, to the output. You can have a very low output from the preamp that can stay low in almost any stage - then boosted later on without noise creeping in. I've noticed that the old requirement for input level setting is not as critical or vital to get right, and for me, it's almost laziness. I might see the input level just tickling the meters and know that it's still perfectly fine, and when a big transient comes along, it copes. Having channel faders that need more, because they're at the top or are so low because the signal is potent doesn't matter - I can adjust everything so that my fader travel is optimal. If a mix gets a bit potent with far too many faders high, I can jus knock the masters down a bit - and with 32 bit operation, all that just works. I tried to use an old computer a few weeks back, in the office with an old version of cubase on it and I kept hearing distortion, because my lazy level setting just wasn't transparent at all. I don't think I can hear better quality from 32 bit maths, but it just doesn't behave like old fashioned mixing.

 

[Slouching Raymond]

I would be most grateful if some technical bod could explain 32 bit float and how it "appears" to beat physics?

It doesn't.
PCs do all floating point calculations in 64 bit.
Most of the bits are for the main number, and some are for the exponent.
Very big values and very small values can be represented, BUT there are very many values in between that cannot be represented.

 

[ecc83]

It doesn't.
PCs do all floating point calculations in 64 bit.
Most of the bits are for the main number, and some are for the exponent.
Very big values and very small values can be represented, BUT there are very many values in between that cannot be represented.

Yes, thanks Ray, I get how floating point arithmetic works in the digital realm. What bugs me is the implied claim by some manufacturers that it endows the whole analogue/digital audio chain with a near infinite dynamic range.

The idea that you can record at very low levels, below -30dBFS e.g. and still bring the track up DIGITALLY with no noise penalty (save the inherent analogue noise. DR stays the same ) has been around for as long as we have had 24 bit operation. 32 bits seems overkill to me?

Been on audio forums for over 10yrs and we have only fairly recently weaned most bods off recording to a gnat's wedding tackle below 0dBFS!

Dave.