Is it really useless Upscaling 16/44.1 music to 24

Trcns,

You've gotten very good feedback. I certainly agree with Dtc, and also adding this as another example.

My PC audio goes from a highly optimized PC running Windows Server 2012 in core mode, with Audio Optimizer, JRiver, into Audiophilleo with PurePower, into Metrum Octave. The Metrum is non-oversampling so any upconversion needs to be done at the computer. In this setup I like it better to keep things native. So no upconversion.

Recently I tried Acourate (software for digital room correction, amongst other things) and to convolute it needs to turn everything to 24 bit within JRiver. The combo of convolution/correction plus upconversion I like better than the above.

A few months ago I purchased an exaSound e22 which gets rave comments from the femto clocks and its ability to play DSD256 natively, and all the fuss about DSD. This unit retails for 3.5k, while my AP+Metrum retailed for combined 2k two years ago.

- Feeding the DAC at the native rates I liked the AP+Metrum combo better.
- Upconverting to DSD in JRiver and feeding that to the e22 vs natively feeding into the Metrum, I like the Metrum better.
- Upconverting at the PC with HQPlayer (instead of JRiver) to DSD and feeding the e22 sounded as good as the Metrum. Maybe...maybe a touch better. But not enough to justify the price difference. So I sold the e22.

Goes to show that implementation is the key. Not ditching exaSound. I think it's a very good product, targeted for people who don't have such an optimized PC, not running WS2012 with Optimizer, etc.

And regarding my preference to keep things native with the Metrum, I often wonder if my speakers aren't the bottleneck to hearing the difference (B&W 804S). Another example of implementation being the key!

I would much rather have the hard wear upsample and not the softwear. JRiver (SOFTWARE) upsampling added noise and distortion, in my rig. I upsample 44.1 TO 196K with an esoteric G25 U and it sounds much better than anything Jriver could upsample in my system. I cannot use JRiver for upampling or dsp. i feel the dsp just tries to compensate for other issues further down the audio chain.

I didn't care for JRiver upsampling either. That is why I feed the DAC at the file's native resolution. However, I realize there are many factors to consider. Some DACs might sound better at higher sampling frequencies, some computers might be noisier, and some might not want to spend in a G25 - for example. But generalizations are tough to get right.

I am running a 10 meter spdif cable from my PC (separate room) into my sound room so the G25U is kind of like a signal amp for me. I guess today, its an upsampler/reclocker in the digital world. Its worked splendidly. I think if I had the puter in the same room (i would blow my brains out) I may feed the hi res signals natively and bypass the g25u but it does a good job with intrnal bypass circuit, stepping put of the way when anything higher than 44.1 comes through it. 44.1 sounds absolutely amazing. Of course there is a variance in quality amongst 44.1 recordings but for the most part 44.1 is imensely enjoyable.

I think if I had the puter in the same room (i would blow my brains out)

I assume you mean because of the noise the computer makes? Fair point, with standard computers. My computer is designed to be in the listening room, only playing music. It uses very little power, so generates very little heat, so passive cooling is good enough (with a purpose-designed case, of course). The PC has no motors, fans or HDD. Makes absolutely no noise.

And a linear power supply helps in not injecting electrical noise, despite being plugged into a different AC circuit just in case.