Hi Mani,
Well, I have a tentative theory that's actually been brewing for quite a while now (>1 year) and that has only been supported by having the NOS1 to play with. My theory is that all oversampling/upsampling (i.e. filtering) performed in the digital domain is bad. And this includes Arc Prediction. BUT... AP does the least damage of all the filtering methods I've tried (in HW and SW).
Although for the general statement you are correct, for the specifics you can not be.
This at least counts for 16/44.1 material which just doesn't contain enough sample points to be right. But please don't fall into obvious pitfalls here;
When you use any random DAC which is not explicitly NOS (the very first were NOS too but not much recognized as such), there's always something in the DAC which "upsamples". However, as you know this this the actual filtering, or better : the means to allow for filtering. So, without upsampling no filtering is possible.
Now, I will confuse you by turning this the other way around : the filtering as we know it (like FIR, PIR) is *explicit* filtering. That is, software changes the wave shapes and you could say that it "filters out" the wrongnesses. This is not what Arc Prediction does; AP merely upsamples only (so, now we could talk about explicit upsampling instead of filtering), but because upsampling will need a proper interpolation means, it does that in the mean time. And that by itself (when done right) acts as the filter again, altyhough nothing is explicitly filtered.
To put this all in other words : when it would be so that even for 16/44.1 AP upsampling makes it worse, well, you would say that vanilla 16/44.1 sounds the best.
Do, do you ?
Nah, I don't think so, and which is in between the lines in your post I just quoted from. But, it may be good to realize that it is not this black and white.
Similarly we must be careful when I say that I even like 192 to be upsampled to 384. First of all, I wouldn't be playing any poor hires so it is not about that. Thus, only the good ones are in for judgement, but now think further what the good ones will be ... This will again be unfiltered results from whatever recording (could be masters), and
theoretically they'd need filtering again !! But, only for content which would be there beyond 96KHz, so it is a kind of far sought;
When I'd have done my own hires recordings - and did that the best way possible, I would never "upsample" that again.
All 'n all I hope it is clear that "upsampling" should be done on 16/44.1 definitely (when playing on real NOS), and the more of it, the better it will be. BUT, only when the first step of it is for the better to begin with (and something like SoX is not).
Uhm, "the better it will be" ? this is not per definition true, because here too (AP), the more upsampling is applied, the less the smaller squares will stay like that (when intended of course). So, there's a tradeoff somewhere (and it is a subject I can't deal with well, unless higher frequencies are dealt with differently from lower ones).
Leading edges are over-emphasized and I'm sure extra HF content is 'created' that may not be there in the original analogue.
Although undoubtedly this will be the case, be careful that you don't judge too soon. Not to undermine your findings, but you *are* too soon to really judge;
I'd say your second break-in period is just about to begin. Give it two more weeks and then it's okay to judge this (and your findings will most probably not differ, but still).
You know, I would love to be able to report that the NOS1 with 8xAP applied to 16/44.1 is just perfect. BUT that's simply not how I'm hearing it.
Simple ... it can't be so. It can be better than (wrongish) native hi-res, it will be better than not applying anything at all, but let's not forget : it is not real.
One last thing maybe ...
Before people will think that comparing (with) vinyl is a measure, IMHetc.O it is not. Of course, a DAC should be able to do that in the first place or oterwise it's "nothing" to our standards. But I think we must be careful when it's taken for an absolute quality measure. There is so much less dynamics in vinyl, sub-low is not possible and all sounds gray-ish to begin with - that it just would not be justified to say "my vinyl sounds the same, so now I have a decent DAC". Oh, I love to hear it from the NOS1 obviously, but I never have been a commercial guy much, and if I (or we) were trying to mimic vinyl, I'd better had made a fine turntable or something.
I hear people say "but if you'd listen to my rig !" ... yes. And you just as well may be right, but then I didn't hear it. The only thing I base this upon (apart from my own poor turntabe) is the getting togethers which are organized each year here, ahead of that mr. Van den Hul, with the finest turntables and cartridges, and really it is nowhere. Still my personal opinion, but just as well something to maybe hunt for. Don't blame me when I'm wrong afterall, but it is unimagineable, the distance being soo large.
Well, let's say all of the above is for some perspective. I'm trying to learn as much as I can myself.
Peter
