From: Dominique Bassal (dominique.bassal@videotron.ca)
Date: Mon Nov 22 2010 - 17:14:50 EST
There may be a middle ground between falling into the "vintage" or the hi-fi "witch-doctoring" on one side, and hearing no difference between softwares that the whole industry uses or rejects unanimously on the other side?
If I don't hear a difference, or if I hear a marginal difference - as you say, the kind of difference that won't survive exportation on different playback situations -, then the testing session will be very short. Likewise, if I hear a significant difference between tools, it is in my interest to note those differences for future use. In my sustained bass example, if I had followed the principle that "most implementations are pretty much the same", and not bought the three limiters after having observed their different behavior and usability, I would have not solved the crackling problem...
Best
Le 2010-11-22 à 14:24, Jörn Nettingsmeier a écrit :
> On 11/17/2010 08:28 PM, Dominique Bassal wrote:
>>
>> That's absolutely true - in theory -.
>>
>> But try the following test. On a reasonably flat, preferably
>> calibrated system, open and listen to the same sound file in:
>> -Quicktime -Logic -DP -Peak -Studio One -Max/MSP -Nuendo -soundBlade
>>
>> I am just using my own list here, but it can probably be extended to
>> any DAW. The differences in "sound quality" are evident. Which is
>> right, which is buggy?
>
> unless there is significant conversion going on, any system that doesn't
> yield the original output, bit by bit, is doing something weird that
> should probably be fixed.
>
> not that i'm advocating "bit transparency" as an absolute measure of
> quality - there are valid reasons to sacrifice it in specific cases. but
> not in summing, which is all a DAW should do, the rest is plugins.
>
> stuff that could happen is format conversion from float to int or
> between bit depths, at which point your dithering approach can
> make a difference. (but there is an optimal way to dither, it's
> all in the literature, no place for magic.)
> all that might cause very subtle differences, but it's not really
> relevant, in my book.
>
> most of us are probably obsessed with quality, which is a safe and
> reasonable approach, but ultimately, if your music depends on such
> subtleties as the alleged difference between different summing busses or
> whatnot, it's going to be steamrollered by real life once it leaves the
> sanctuary of your studio.
>
> i don't have access to all these systems, but it would be enlightening
> to do a null test between them: start your test file with a pulse that
> allows for manual resyncing later, then play it through various systems
> and record the output with a program that you trust. invert the phase
> and add it to the original, match the gains and sync to the pulse.
> i'd expect the only difference you get is some dithering in the least
> significant bits, and even that only if there is bit depth conversion.
> if you do find significant residue (i'd be surprised), i'm sure many
> major studio magazines would like your write-up as a front cover item.
>
> if your null test yields all zeros, and you still hear a difference when
> you play it, you have just proved the most important selling point of
> the plugin industry: the placebo effect.
>
>>> even with eqs, compressors, panners i'd venture that it's less the
>>> absolute sound that matters (most implementations are pretty much
>>> the same and only reveal their peculiarities in corner cases), but
>>> how the controls react and whether you are comfortable with that.
>>
>> I am actually mastering a piece with long sustains in the low bass.
>> The first end chain limiter that I try (Slate Digital FG-X / very
>> pleasant sounding) generates crackling at all release settings. I
>> switch to tc PowerCore Brickwall limiter (the most transparent I
>> have / known to be capricious): crackling. I switch to Flux:: Pure
>> Limiter II (never failed me, but slightly colored): no crackling,
>> under all settings in the "Dynamic profile".
>>
>> Even if I was completely hallucinating about the sonic
>> characteristics of these 3 tools, there is still no crackling with
>> the 3rd one.
>
> sure. but "crackling" is sort of natural when limiting... you are
> playing with the envelope. that's inherently non-linear, hence
> distortion artefacts will occur, always.
> a good dynamics processor will let you find settings where these
> artefacts are inobtrusive (usually that means they are being masked by
> the program material) or use tailored slopes that will
> only result in very low distortion levels. but yeah, sustained bass
> notes have little chance of masking that. heck, even many cheap analogue
> units fail in that case.
>
>> I could go on with EQ, passive and active, compressors, etc. Huge
>> differences in basic sound between them.
>
> hmm. i'd not call them "huge". certainly gear (or plugins) can be a
> source of inspiration, in which case subtle (or even "hallucinated")
> differences will be of musical importance - fine with me. but an eq has
> an amplitude and phase response (plus you can fake some analog
> distortion). in any case, there is not much room for the metaphysical there.
>
> my point was: if you need a bit of corrective eq to deal with some
> acoustic problem, any correctly implemented filter that can be set to
> the required response will do.
>
> naturally, a passive mastering eq will yield different results than an
> active one, easily measured and heard. but my gut feeling is that it's
> the difference in "action" when you play it that will cause most of the
> sonic difference in the final product, rather than the actual circuit
> characteristics.
> much like a strat tends to make you play other stuff than a les paul.
>
> computers are not sexy, and musicians want sexy, and mysterious. fair
> enough. but it also means that plugin manufacturers have to resort to a
> certain level of bullshitting to meet demands (and some arguably get a
> bit carried away with that), because there is nothing sexy about adding
> and multiplying. which is why most of the audiophile witch-doctoring
> crowd has flocked to analog circuitry, lately, in search of the "little
> something".
>
> sadly, computers are perfect machines, and correct implementations for
> all linear and simple non-linear maniplations exist. there is no
> reasonable next year's killer application for basic signal processing on
> the horizon. still, new products are needed to keep the business going,
> hence brushed aluminum widgets with "TubeGlo" fake vu meters and vintage
> dials you have to turn(!) with your mouse(!). go figure.
>
> much as i hate it, it's melodyne and friends that define the technical
> apex of digital audio manipulation - there is real progress there, and
> still much room for improvement...
>
>
> best,
>
> jörn
>
>
>
>
> --
> Jörn Nettingsmeier
> Lortzingstr. 11, 45128 Essen, Tel. +49 177 7937487
>
> Meister für Veranstaltungstechnik (Bühne/Studio), Elektrofachkraft
> Audio and event engineer - Ambisonic surround recordings
>
> http://stackingdwarves.net
>
This archive was generated by hypermail 2.1.5 : Thu Dec 23 2010 - 18:30:23 EST