Best sound quality - run 96khz samplerate, do not oversample.

[Fritze]

But then again - what do I know, I'm just a normal "audio guy".

Not directed at you specifically, but I'm more interested in hearing than knowing. That brings me to a cool idea. Anybody interested in some ABX testing?

[Compyfox]

Maybe it's the ultimate task for the platinum ears "I hear bits flop" people over at Hydrogen.

[Fritze]

Maybe it's the ultimate task for the platinum ears "I hear bits flop" people over at Hydrogen.

You mean it's better to keep this on a theoretical blah level than to actually listen to how this might effect some audios?

[Fritze]

Thanx for that. Just what I thought.

[AudioGuy720]

Another thing you need to look out for which the charts don't have is the amount of pre-echo a steep filter has.

It is shown in the "Impulse" chart.

Haha, now I feel dumb. I totally forgot about the impulse chart.

Despite what you guys are saying/what the charts say the Weiss SFC2 hardware is not shabby at all. But more affordable software has overtaken it in recent years (the SFC came out in 2000 when most SRC's were uglyful) so I guess it's a moot point.

This will be my last post to this thread. I stand by my recommendation of recording at 24-bit/48 kHz if you have an older system or 24-bit/96 kHz if you have a bad ass machine. Export 32-bit or 64-bit floating point files for further processing.

I'd even recommend recording at 24-bit/96 kHz, downsampling to 44.1 kHz for mixing and then when you're finished and ready for bounce replacing your files with the higher quality files. This is what they do a lot in video editing where they'll take a low res "offline file" and then as the last step use the high-res HD (or nowdays 4k/6k) "online files". That's IF your DAW supports changing project sample rates!

[Fritze]

This will be my last post to this thread.

Strange statement before you come up with the good stuff. Thanx for your recommendations.

[Debutante]

chiming in on this old flamewar topic, I've heard this since I joined KVR. SO a long time ago I bounced snippets at 44 96 and 192.


To sound the same, they should be silenced when inverted. they are not. IIRC there was always a high freq coming through with different sample rated pieces playing against each other. 44 sounds fine to me... but why is it the lower ends cancelled out and the very highs bled?

[SuperFly76]

I had a feeling this would happen. We need Ethan Winer now. Now we will have the back and forth game of "Nyquist proved 44.1 is always adequate" vs. "I don't care, I can hear a difference". I've seen this play itself out to very caustic, argumentative results on other forums (but I'm assuming it's happened here before as well).

I've seen knock down drag out fights between professional audio engineers vs. people with a background in digital processing who have viciously disagreed about whether a basic Sound Blaster card from 2003 is adequate for D/A conversion (i.e. there is "no discernible difference") vs. high end converters.

The bottom line is this: no one will be able to resolve this without extensive ABX testing, and even then it may completely depend on who is tested.

Otherwise we are doomed to just go around and around again, and believe me this has happened many times before with no appreciable constructive result.

To quote Battlestar Galactica: "All this has happened before, and all this will happen again."

[SuperFly76]

I'd be more willing to use higher sample rates if all plugins worked properly at higher-than-44.1k rates.
Sadly, they don't. It's something that I keep an eye on now, when trying out new plugins.

I agree with you 100%. I worked at 88.2 for a while but discovered that quite a few of my plugins didn't work well, especially plugins with fixed filters that haven't been coded properly to readjust for higher sample-rates. That's the main reason I still work in 44.1 or 48 respectively. That, and honestly, I heard no perceivable difference. Where I have heard noticeable differences is with soft synths, and most of them now have oversampling as an option anyway.

[pwal]

hm, i've been running @ 88.2 for the last year or so, haven't come across any plugs failing/going weird there

[aciddose]

there are definitely some cases where constants are hard-coded and not adapted correctly across rates. the error will sometimes be more, and sometimes less than what you'd consider "acceptable". what you consider "acceptable" depends upon the context in which it occurs as well.

that said, you may just not notice the difference if you're not comparing, and as long as you do all your work and renders at the same rate you'll be getting exactly what you expect and so all is fine.

normally you don't care what the dc-blocking filter frequency is when you plug in an instrument and start using it. it's just that if you were using something like a bass-boost eq, and suddenly the frequency changed from 55hz to 70hz, then you'd be scratching your head and or throwing an eq out the window.

[Meffy]

normally you don't care what the dc-blocking filter frequency is when you plug in an instrument and start using it. it's just that if you were using something like a bass-boost eq, and suddenly the frequency changed from 55hz to 70hz, then you'd be scratching your head and or throwing an eq out the window.

*wince* You strike mighty close to home there. Dumb dumb dumb me... so many hours wasted chasing down wrong paths. Ah well, live and learn.

[HanafiH]

To sound the same, they should be silenced when inverted. they are not. IIRC there was always a high freq coming through with different sample rated pieces playing against each other. 44 sounds fine to me... but why is it the lower ends cancelled out and the very highs bled?

Now that's a good question. It's because many (though not all) plugs fail in some way, the closer you approach nyquist. Bandlimiting filtration, skewed and cramped EQ curves, aliasing and anti-aliasing algorithms can all cause issues the closer you get to nyquist. By moving nyquist up, you have room to avoid these problems in audible frequencies.

The refusal to ackowledge that there is any difference between pure sampling and pure processing allows the argument to fester endlessly. The best advice I ever saw is to do what you and I have done. Test it. Testing, testing testing: separates the engineers from the primadonnas.

[kelldammit]

chiming in on this old flamewar topic, I've heard this since I joined KVR. SO a long time ago I bounced snippets at 44 96 and 192.


To sound the same, they should be silenced when inverted. they are not. IIRC there was always a high freq coming through with different sample rated pieces playing against each other. 44 sounds fine to me... but why is it the lower ends cancelled out and the very highs bled?

that's precisely the behavior i'd expect.
at the higher sample rates, freqs aren't getting filtered or aliased until they hit 44.1 or 48k, instead of 22 or 24k. so there's some difference there already, should you happen to have dog ears

the lows would probably cancel out because the higher sample rates won't have such a pronounced impact on how those frequencies are reproduced.

additionally, since when you change from 44.1 (or 48 )k to 88.2 or 96k, you're moving the nyquist threshold, so you're changing what gets aliased, and where that aliasing will show up (frequency-wise).
again, this aliasing simply won't be as pronounced at the bottom end. it might help to think of aliasing as visually looking like a reversed reverb-tail type waveform, with the highest amplitude at nyquist (or thereabouts). by the time it gets to the low freq's, the "reflections" have largely petered out. Does that make sense?


k

[pwal]

that said, you may just not notice the difference if you're not comparing, and as long as you do all your work and renders at the same rate you'll be getting exactly what you expect and so all is fine.

that'll be me then