I have my DAC8 DSD running with Roon for two years now - and love it! This device really brought my music collection to a new level.
But one thing bothered me always… everybody is writing how much better the DAC8 DSD would sound if everything is upsampled to DSD 512. So after some thinking I recently spent some money for an i7 machine that could handle the upsampling. Everything connects fine, works great, I’m using the recommended Roon settings from Larry_HV he posted in this forum (thanks!). CPU speed is fine, no crackling, no dropouts.
But… is it me or what? I cannot hear a difference between the original samplerate and the upsampled DSD512 version. I tried different tunes, different sample rates, from 44.1 kHz to DSD64. I tried different speakers, from Canton Karat to Yamaha studio monitors and also headphones (Sennheiser HD800S). I do hear differences between different DACs, different mastering versions of music etc. so I think I do have a good ear for this. But I still can’t hear a difference between the original and the DSD 512 upsampled version.
I don’t think you re missing much at all actually!
Upsampling is a bit of a turd polishing job, but even a shiny trd is still a trd. It all falls back to the mixing and mastering of your prefered music, if the guys and gals have made agood job there, you will get full pleasure form a humble 16/44.1 file…
That said, there are some DACs that perform better when they are served a prepared meal, rather than the raw ingredients.
I don’t think there’s any right or wrong with either side of this, it’s a matter of taste, interest and what tribe you want to identify with.
Cant hear any significant difference? Spend your money and time elsewhere, and don’t let anyone tell you what you hear and experience. (not even me!
I run the same DAC-8 DSD with HQ Player through Active Monitors. Certainly can hear a fairly obvious difference between DSD512 upsampled snd non-upsampled 44.1/16 content through the same setup.
Why is it different, it’s tightens up the wave timing so much you can hear the underlying reverb in recordings. (And if you haven’t done it yourself please don’t offer opinions on why it can’t make any difference
Worthwhile getting this working for sure. Something is probably not enabled correctly in Roon. Perhaps try HQPlayer in the signal chain.
To me it’s about how the music flows, not about highs/mids/lows being better. The dac just seems to make the music sound less mechanical and smoother with DSD512
Thanks everyone. So the opinions reach from “don’t bother” to “use HQPlayer, it will make a big difference”
Thats is exactly what I heard when I switched from my previous DACs to the T+A DAC8 DSD at native samplerates (no upsampling)! Thats what I love about this DAC! I’m hearing the reverb from the recording and getting this great “3D effect” with my speakers where you have the feeling the musician is playing right in your living room. But I get this all without any upsampling… so I wonder if upsampling is really that much better?
I’ll give HQPlayer a try and will let you know. What is the recommended settings for HQPlayer, I heard it has TONS of filters and algorithms…?
Yes the DAC-8 DSD sounds very good anyway with normal 44.1khz PCM content, but is a big enough step up when upsampling to DSD512 which is why most tend to go that way. It wouldn’t hurt to trial HQ Player and check yourself.
Indeed, you’re right. I can barely find a setting that plays without glitches. Weird thing is, the CPU usage never goes above 50% and still I hear dropout (especially bad on the EC variants). Is an i7-9700 not powerful enough for this?
I have to admit, these filters DO sound different then the DAC8 DSD or the Roon upsampling. But I cannot say if its better or not with all the glitches and dropouts. I’m trying more over the weekend.
Blockquote
Weird thing is, the CPU usage never goes above 50% and still I hear dropout
I believe if its around 50% that’s basically “100%” as its split across cores. “Closed Form” isn’t too CPU intensive (5% 'ish, if I recall) poly-sinc-xtr-mp would be. Just experiment and have fun.
HQP 4.x embedded is supposed to be pretty decent jump in SQ
OK, after some extensive testing I decided not to use HQ Player and enjoy the sound of the DAC8 DSD as is. I couldn’t find an HQ Player setting that does not glitch in any way AND sound significantly better than the DAC8 at native sample rate. The difference is not that great to spend even more money on a faster CPU for the EC filters (I’ll buy a second DAC8 DSD instead for my office ).
I understand that some people prefer the HQ Player upsampling, but this is my personal taste. YMMV.
Thanks for the help and tipps, at least I gave it a try.
Hi @dabassgoesboomboom,
yes, finally after more than 6 years the production of the DAC8DSD has come to an end, the reason being that some of the key components are not available any more.
Currently I am not allowed to talk about details, but so much for now: a new T+A DAC will be launched in the very near future.
About the “best” output impedance or damping factor a lot of different opinions exist. From a technical standpoint I agree that as low as possible output impedance is desirable. This is why we at T+A in the 1990’s have built amplifiers with damping factors of over 1000 (4 mOhm output impedance). These amplifiers had very good measured data but they often were characterized as a bit sterile and “analytical” sounding. About 20 years ago we changed the design of our amps. The new amps now have much lower damping factors and by far the most of the listeners like their sound better. Of course damping factor is only one thing that infuences the sound but imho the damping factor should not be too low. This is why we made the DF adjustable in the HA200 and most users of this amp I talked to use output impedances of 12 Ohms and above as an optimal value for their phones.
I think you should once listen to the DAC200 before you judge.
Regarding USB inputs:
The DAC200 uses the same new proprietary USB receiver as the SD/SDV3100HV and the HA200.
These receivers are fully UAC3 compliant (of course also UAC2).
They support native DSD up to 1024 from Windows and Linux (kernel > 4.4 / 5.x kernel recommended) based sources.
With MAC OS devices DoP is supported up to DSD 256.
