The sound presentation of my Directstream at 16/44.1 is vibrant, punchy, forward, big and clear, but as the bit depth and sample rate increase something is both gained and lost. It seems like the DS provides a really big canvas for 16 bit, but for 24 bit has to shrink the higher resolution image to fit into the same space. But the biggest change is in sample rate (more with multiples of 48khz, for some reason). 24/96 still sounds pretty good if a bit less precise, but by 24/192 the DS sounds a bit muddled, lacking punch, and washed out and I have to turn up the volume control on my preamp.
With DS I can hear clear benefits going to 24 bit, 96, 192 and occasionally 352khz: better resolution, smoother and more solid sound sound, maybe a little more air or space. These are the same things I notice with other DACs when switching to higher resolution. However, it does seem like some drawbacks in the presentation become apparent the bigger the file format.
If there is a difference in the way the DS handles 16 and 24 bit or 44.1/48khz vs higher sample rates can you guys address these in a future update? Thanks.
Your observations might also point to your source or digital interconnects: higher frequencies radiate more “crap” that may affect one’s system. As an experiment (and if your source supports it) you might try a TOSLink cable and unplug all other inputs.
There are also some sources that treat 48, 96 and 192 differently than 44.1, 88.2, 176.4… Sometimes they only have one precise clock and use a frequency conversion chip (or something equivalent) to generate the other series.
FWIW Pairs of frequencies (I.e. 44.1 and 48, 88.2 and 96, 176.4 and 192) are processed identically in the DS and in many systems the clarity, cohesion of the sound stage, etc. gets better as frequencies go up since there’s more information present that supports the sound stage.
Snowmass definitely helped 44.1 more than, say, DSD. All frequencies got better but 44.1 (and other lower frequencies) had more room for improvement.
[Edit: I should also have mentioned that if you are using upsampling or downsampling to test different rates things are more complicated. If you are running your experiments by upsampling the logical conclusion is that the DS is doing a better job at upsampling. On the other hand if you downsampled some material for these kind of experiments your downsampler (with the settings used) may just be better at 44.1]
Very interesting. Thanks for your detailed response, Ted!
I’ll have to try a TOSLink converter+cable. Right now I’m going Mac Book Pro>Audirvana 3+> Shunyata Venom USB cable. Probably the weak link is the Apple USB buss. However, my Mytek Manhattan seems less sensitive to the system noise (it’s also a more austere and pro-audio type of DAC). I’m guessing a dedicated server/network solution might yield better results.
My observations have come from a sampling of different recordings at their native sample rate and it depth, but I double checked by down-converting one of my own 24/192 recordings into different sample rates and bit depths within my DAW. I guess it’s possible that the DAW SRC uses different filter cutoffs for 44.1 and 48 multiples.