Dsd - how far does one go?
- Thread starter Audiophile Bill
- Start date
Ouch - harsh words. I thought you had the integrated Aavik on loan? This is different product - it is a preamp with PCM/dsd/phono stage
The better algorithms require more CPU power. Jussi said that his algorithms he uses today can be preformed on a FPGA, but the cost of the FPGA is $20000 just for the raw chip at wholesale prices. He has algorithms already prepared for CPU's years down the road. His idea is to release the new algo's as CPU's get more powerful. To keep up with moore's law. Most people usually update their servers every 3-4 years anyways, so works out perfect.
Last edited:
You had the C-300 that was just unveiled at CES? As for the U-300, I don't blame you. 24/192 PCM max capable, and mediocre Pascal Class D amps. Not my cup of tea.
Oh right, I had the U-300 - but it is not only a dac, also an integrated amp, and I tried it direct, with Lampi in, and with a Sony SACD player in. Not for me
Yeah I agree, The Hypex NC-500 amps in the standard $1300 Nord NC-500 based stereo amp are far superior to the Pascal amps in the U-300. Same reason Bel Canto ditched the Icepower modules and have gone with the NC-500's in their $4990 Ref600M's
Nord:
http://www.iqspeakers.co.uk/#!hypex-amps/avuyk
Bel Canto Ref600M:
http://www.belcantodesign.com/home/eone/ref600m-amplifier/
Last edited:
I am interested in the original question still. How far to go with DSD? Sooner than I expected now someone makes a 16x DSD or DSD1024 machine. I understand the possible benefit DSD128 has over 64 as it pushes the noise further from 20 khz. After that what are the higher rates accomplishing that the lower rates don't do adequately? Anybody have some idea?
Most say after quad DSD the benefits are inaudible. Better to work on loose ends elsewhere in the chain.
Better attack transients for me. And that has many implications for overall sound, including timbre accuracy and soundstage.
Takes the noise further out of the audible band. I can hear the difference in my system.
Yes, but there's different levels of filter algorithms that require different levels of processing power. So, just because it is a higher sample rate, doesn't mean it took more computing power to preform the task. For example Hqplayer has a pile of different algorithms for the upsampling/modulators. some require little power, and some much more. But the best sounding ones require the most power. That's the drawback with using an FPGA. they can offer firmware updates, but before long the FPGA maxes out and makes it impossible to further upgrade. This is why in the EMM Labs DAC2X they had the MDAT1 DSP, and now they have the MDAT2, which required a complete new $25000 DAC, with more powerful FPGA. If the DAC was just made to process the DSD filtering end only, upgrading to the new filter algorithms would just be a free update from HQplayer. Eventually you will need to update your computer, but most people do that every 3-4 years regardless. Not only that you will never be able to process the level of algorithms on a FPGA, as an powerful intel processor. The algorithms that run on the EMM Labs DA2 for 16x DSD upsampling wouldn't even make an Intel processor flinch.
I haven't heard it yet, but makes sense to me. I always read reports that DSD sounds more 'soft' and 'rounded' than PCM, something that many found pleasant, apparently, but an idea that I intensely disliked, because it goes against my preferences and my experience of what live unamplified music often sounds to me. Then I read reports that the DSD 256 NADAC sounds energetic and punchy, something very much to my liking. I'll hear it soon myself.
Here's the obvious questions:
Does loading the FPGA device with more "work" means chances that it places more demands on the power supply of the dac/player, thereby could potentially affect sonic performance?
Do FPGA devices generate noise radiation - and that running at higher demand on its capabilities generate more "noise"?
Does it also means that upsampling 44.1kHz pcm to dsd may actually end up affecting noise and power supply loading issues?
Well it may, but a well engineered product will ensure that it is well within spec throughout the entire operating range. However, in my experience, the simpler you can make a DAC, the better. I'm more of a fan of the software based SRC/SDM approach. All in 1 boxes are built for convenience, not the ultimate in sound quality.
I've spent the last couple years exhaustively trying every approach and this is where I'm at now anyways.
1. Soft is not the same as fluid.
2. On my system, PCM does not lack fluidity compared to live music. But of course, different people's perceptions differ.
Similar threads
