Upsampling the way to go? ?

I recently upgraded my Leveinson #39 to the new #390s. The #39 was an excellent sounding player, but the 390s is a big improvement. The 390s increases the sample rate to 352.8 kHz or 384 kHz and is the main reason for the improvement.

I think that both 16 bit and upsampled 24 bit can sound very good. Most of the "problems" that we hear are not with either format, but with the amount, type and how the filtering is used in each design. I've tried it both ways and i prefer upsampling. Then again, my upsampler doesn't have any filtering so that may be part of the reason that i like it so much. The non-oversampling ( 1X ) Audio Note's also don't have any filtering and that may be why so many people think that they sound so good also. Sean
>

The upsampling is only as good as the implementation. If the implementation isn't done with care the added circuitry can easily overwhelm any benefits of upsampling. Better to buy a well-executed non-upsampling DAC than a mediocre upsampling DAC. Also, it's important to understand that upsampling is not analogous to interpolation where the word length is increased to 24 bits or so--they are two separate things and should not be lumped together. I'm drunk and going to sleep now--night night.

Tim

Tim, I thought interpolation *was* "upsampling", whereas increasing word length was accomplished by adding randomly generated dither below the signal level, so it sounds to me like *you* might be confusing two separate things. But then again, what the heck do I really know about this stuff?! Not a lot, let me tell you!!

To address the main question, I just yesterday received a new-to-me MSB Gold Link upsampling DAC bought here on the 'Gon, to compare with my reference Theta DSPro Basic IIIa (Theta also seems to disagree with the theory). I've just started auditioning, and will post my findings when I've had enough time to make a meaningful decision about which unit I will keep. But I can tell you right now that the MSB with upsampling engaged definitely sounds different not only from the Theta, but from itself with the upsampling turned off, so it ain't doing nothing. Finding an upsampling-capable unit and auditioning it against your reference may well prove to be an educational experience - I think it will be for me.

Zaikesman--I'm not an engineer and I could be wrong, but upsampling is upsampling and is totally separate from increasing the word length, which I believe is done via an interpolation algorithm. I was told this by Rick Schultz at EVS who has designed his own DAC and has modified a few in his time. Many DACs claim to upsample to 24/96 or even 24/192 but do not really interpolate to 24 bits. The Crystal chip does do both and may be what is used in the MSB DAC--I'm not sure. Hey, if it's working it's working, whatever it is. Better systems are used by Audio Aero(STARS) and dcs, but they obviously come at a premium.

I still maintain that an upsampling DAC does not necessarily sound better than a non-upsampling unit--it depends more on how the whole DAC is designed and implemented, and there are a LOT of variables there(i.e. Sean's point about filtering, damping, shielding, parts quality, power supply, circuit design, etc.). I'd say just pick the DAC/CD player that sounds best to you--if it happens to upsample great, if not that's fine too. Personally I think Sam Tellig has gone way overboard on this upsampling thing and is misleading many to think it's the second coming when it's only one small link in a very long chain.

Tim