not trying to be a contrarian, but these days with so much streaming content out there for the taking, my expectation would be that about 95% of any vinyl you could rip would be out there accessible from a streaming platform; and likely be tape based, so at least having equivalent performance to your particular vinyl set-up.
Mike, when you say 95% of any vinyl is out there and accessible from a streaming platform, can you elaborate?
IME, some of the best quality vinyl is / has been made available via Mobile Fidelity, Impex, Analog Productions, etc., most of which isn't remastered and available to purchase digitally, disc and/ or streaming. To me, that's the real benefit of vinyl these days is to get the best version of rock and jazz in particular, being remastered to yield the best sonics via aforementioned companies and others. That best version, I think others would be interested in ripping to digital for convenience's sake.
I hear significant improvements as the sample rate rises, by which I mean the digital rip sounds ever closer to the native vinyl in my system the higher the sampling rate goes. I've done tests where I rip and 768/24 and then downsample (so there is only one "read" of the record) to the various sample rates and tested. And I've done multiple rips of the same record at various rates all the way up to 768/24 (accepting there may be variation between plays) and I hear the same thing both times - the higher the sample rate the closer it sounds to native vinyl.
You're right. I made a couple of rips using VinylStudio Pro and compared it to Izotope RX record function. VinylStudio might be just a smidge clearer. More importantly there is a noticeable difference between 768/32 and 96/24. There is a slight air and freshness with 768KHz that lower sampling rates cannot capture. Unfortunately it loses all that magic when you resample to lower, 96KHz and/or dithered to lower bit rate. 4X DSD recording and playback isn't as good as 768KHz either.
This also proves any digital process is audible and smears the SQ. Thanks for the tip.
For me, when doing these comparisons, the most obvious differences are in the air and the space around the instruments. As the resolution climbs, the air, the acoustic space, becomes more obvious - you start to see the space between instruments clearer, and finally into the depths and behind them - it all becomes 3D. And this is critical information too. As the resolution drops the acoustic space suffers first, you lose the "visible" air and space, and when you jump back to 44.1 it's obvious - the acoustic space is smeared out and *everything* becomes "out of focus" and "wider" - for instance vocals - at high resolutions you can almost see the lips of the singer focussed in space in front of you, at 44.1 the voice becomes wide and out of focus - a blurred presentation of a singer.
As to why? Well, my conclusion is that has to be reconstruction. If digital reconstruction was perfect you would be able to take a higher resolution and downsample it and still reconstruct it back to the original sound. This works in sampling theory, but not yet in physical DACs. Information gets lost at lower sampling rates - and we can clearly hear this. This is why HQPlayer and Chord MScaler throw millions of taps at the problem to try and get better reconstruction, which subjectively they do, but still not perfect reconstruction. So while we wait for implementation to catch up with theory we can assist by using higher resolutions which helps make reconstruction less error prone...
And if we are brute forcing sampling rates to solve reconstruction, then I feel that 768 is also not quite there. Perhaps 1536 would be 100% transparent?!
And if we are brute forcing sampling rates to solve reconstruction, then I feel that 768 is also not quite there. Perhaps 1536 would be 100% transparent?!
IMHO/IME 768K isn't transparent enough and there is still a huge difference in comparison to original analog. The difference is higher than the difference between sample rates. I don't think 1536K will be %100 transparent.
This is why HQPlayer and Chord MScaler throw millions of taps at the problem to try and get better reconstruction, which subjectively they do, but still not perfect reconstruction. So while we wait for implementation to catch up with theory we can assist by using higher resolutions which helps make reconstruction less error prone...
I use Audirvana and it makes a huge difference at upscaling/upsampling using a minimum phase filter. There is still a slight, but noticeable difference between 96K upsampled to 768K and native 768K. I tried HQPlayer before but I liked Audirvana more. It sounds livelier to my ears. BTW it's Audirvana 3.50 not studio.
IMHO/IME 768K isn't transparent enough and there is still a huge difference in comparison to original analog. The difference is higher than the difference between sample rates. I don't think 1536K will be %100 transparent.
I use Audirvana and it makes a huge difference at upscaling/upsampling using a minimum phase filter. There is still a slight, but noticeable difference between 96K upsampled 768K and native 768K. I tried HQPlayer before but I liked Audirvana more. It sounds livelier to my ears. BTW it's Audirvana 3.50 not studio.
Thanks for putting it into perspective, I thought for a moment the holy grail of the perfect analog capture device had been found...have you tried DSD or any early ADC's that didn't use delta-sigma chips?
Thanks for putting it into perspective, I thought for a moment the holy grail of the perfect analog capture device had been found...have you tried DSD or any early ADC's that didn't use delta-sigma chips?
