Re: Latency vs. Sample Rate


Arny Krueger wrote:

> Which console?

I was trying to avoid mentioning that to protect the potentially
guilty. And I won't tell you if you guess right, but you probalby will.


> I note that the spec'd delay of my 02R96 is given as < 2.0 ms
> at 48 KHz and less than 1.1 ms at 96 Khz.

That's more like it, and how I expected the relationship between delay
and sample rate to be. But I still wonder if when all is said and done
and heard, it's still too long for comfortable tracking. Since the
usual multitrack monitoring path is from the returns from the recorder,
whether it's an analog multitrack recorder, a hard disk recorder with
dedicated "track" outputs, or individual outputs, submixes, or a total
mix in a DAW, you have the mic going in to the console and its signal
arriving at the recorder 2 ms later. An analog recorder can turn that
around instantly but a digital recorder does the "input monitoring"
switching at the digital end of the output stage. If it's analog,
there's about 1.5 ms delay going from analog to digital and back to
analog, and then another 2 ms or so (optimistically) to get from the
monitor return out to the headphones on the talent's ears. In the case
of the console I asked about, this could be more 10 ms worst case all
around.

They always show you how cool it is to mix on these things, but they
never show you what it's like to track on them. <g>

> IOW, its like being 5 feet from the voice coil of a typical floor monitor
> which is true for everybody but really short people.

They always say that, and it always works out in practice. But how
about if rather than standing 5 feet from the monitor, you were wearing
headphones, as is the case most of the time in the studio? And besides,
you have only one delay path in the mix going to a stage monitor
speaker, you aren't dealing with a round trip to a recorder and back to
the monitor mix path.

> In the case of the 02R96, the extra 2 ms effectively move the stage monitor
> back another couple of feet. I've never seen a muso run from the room
> screaming when such a thing happened onstage.

You're using it on stage. Studio is different. There may be some
psychology here, and the comb-filtering is never an issue since the
monitor volume is always louder than the acoustic volume. Maybe this
is part of the reason why digital consoles are becoming so well
accepted for live sound (because they don't make things significantly
worse than real life) but are slow to be accepted in the studio, at
least for tracking.

> My point was that the stylish business of building the filter so that it had
> a roll-off starting at far lower than Nyquist could affect things in
> unexpected ways.

If you're sampling at 48 kHz, you want the input filter to axe anything
above 24 kHz. You can do that with a really DSP intensive filter that
probably takes a lot of itme to run, or you can oversample so that the
effective Nyquist point is out at around 3 MHz (64x). That way, you can
build a gentler low-pass filter that takes less time to run, and still
have it be down far enough at 24 kHz so as not to let the nasties in.
But if the oversampling drops to 32x when you double the system sample
rate, in order to build a filter that will pass another octave above 24
kHz while operating at the same "oversample" effective rate, it has to
be steeper, thus requiring more horsepower, or for the same amount of
horsepower, taking more time to run.

So that might be the explanation, sort of, in layman's terms. But I
still think it's starting to smell a little bad.

.

Relevant Pages

  • Re: Latency vs. Sample Rate
    ... and how I expected the relationship between delay ... > usual multitrack monitoring path is from the returns from the recorder, ... you have the mic going in to the console and its signal ... > monitor return out to the headphones on the talent's ears. ...
    (rec.audio.pro)
  • RE: MONITOR command
    ... "Activate MONITOR independent of a console" was an enhancement to z/OS ... R7 as part of the console restructure. ... specify the SETCON MONITOR command in COMMNDxx ...
    (bit.listserv.ibm-main)
  • Re: Think I want a console (dorsey, rivers, et al, please respond!)
    ... Here's a link to the manual from Soundcraft: ... is normally what you monitor when you're tracking. ... pressing the TAPE buttons on the main channels routes the ... An in-line console has the monitor level ...
    (rec.audio.pro)
  • Re: MS Fax basic send failure
    ... There's no tools>options on the fax console. ... entry pending in the fax console's outbox folder. ... Select tools, and choose fax monitor. ...
    (microsoft.public.windowsxp.print_fax)
  • Re: Double hires mode color artifacts
    ... So we need to know how the filters worked in an actual monitor. ... an emulator and I've got my plate pretty full with other things. ... should take 15 minutes or so (once the filter coefficients are ... And this should stay in the cache. ...
    (comp.sys.apple2)