Exactly. Scroll down to Jussie’s post 551 for a clearer understanding.
Especially if on MacOS.
GUSTARD DAC-R26 Balanced Decoder R2R+1Bit Dual Native Decoding Music Bridge
- Thread starter m-i-c-k-e-y
- Start date
Especially if on MacOS.
Please refer and try to understand Jussis post number 551 with measurements to show you how wrong you are.
https://community.roonlabs.com/t/which-hqp-filter-are-you-using-2015-2023/6061/551
Last edited:
FWIW it was the damping directly on the FPGA chip on the R2R ladder that made the most difference. Just trusting my ears AB-ing at different bit settings. I posted about my experimentation up thread. What would be fascinating would be to try to measure the very audible effects of signal grounding and vibration damping. I'd wager Wolf's SINAD and jitter spurri graph might look a little better...
Maybe
But with the hundreds if not thousands youve spent tweeking your Dac/system
one could simply purchase a holo spring 3 (SINAD 116.5) or Holo May (SINAD 118.5) to increase resolution way past the R26.
sajunky
Headphoneus Supremus
Please do not repeat yourself, it is insulting. Jussi has no scientific authority, he is computer programmer.Please refer and try to understand Jussis post number 551 with measurements to show you how wrong you are.
https://community.roonlabs.com/t/which-hqp-filter-are-you-using-2015-2023/6061/551
I have spent rather a lot on tweaks now (hundreds on Tungsten cubes alone), and the thought of upgrading the R26 has crossed my mind more than once. One day when our exchange rate is less dire. Thankfully none of the tweaks are permanent... all relocatable to other or new components where they may or may not have the same effect (though most do from experience).
this is the stupidest thing I’ve read all week!! Congrats! The creator of HQplayer who has worked with tons of audio companies and dac manufacturers, licensing his work to said companies is not an audio authority of HQPlayer?!?
Yet Sajunky the no name troll who circles random dac components thinking he knows what they are, posts fantasy and make believe opinions, is some kind of authority?
Please leave Head-Fi and never return.
We will all be better for it.
JackJohnson316
Head-Fier
Quick question regarding 15 bits. Are you saying when the R26 is in NOS mode, I shouldn’t send more than 15 bit music to the unit? I use Audirvāna at the moment and just send the music via the lan port bit perfect, usually 96k/24 bit or 196k/24 bit. Are you saying I should down sample to 96k/15bit and 196k/15bit in Audirvāna? Not even sure if that’s possible, but thanks for the help.
sajunky
Headphoneus Supremus
Please read carefully, highlighted. Jussi has no scientific authority, he is computer programmer.
The creator of HQplayer who has worked with tons of audio companies and dac manufacturers, licensing his work to said companies is not an audio authority of HQPlayer?!?
Yet Sajunky the no name troll who circles random dac components thinking he knows what they are, posts fantasy and make believe opinions, is some kind of authority?
Please leave Head-Fi and never return.
We will all be better for it.
He can talk about SINAD calculator, as such measured value is directly linked to the amount of added dither in a DAC, so users can estimate amount of dither without listening tests.
But when he do asociate SINAD calculator to the real accuracy of the ladder, it is indicative that his education is far away from the theory of signal processing. Jussi is a complete ignorant in this area, it is a fact confirmed by Jussi.
And please don't talk about trolling as my response was strictly giving answer the question asked by the user.
Yes if Audirvana can.
sajunky
Headphoneus Supremus
Not. Number of output bits is a different things. Just select 24-bit.
Do you really need to revisit the very topic @GoldenOne embarrased you on?
Would’ve thought that would have stuck.
Or you can go back a few hundred pages here and see where I gave you the same.
Please go.
Last edited:
sajunky
Headphoneus Supremus
Neither you nor goldenyou has ever managed to embarrass me. You never speak yourself, you talk goldenone, this is other things. As you don't address the issue, but attack the messenger, it is end from my side. Thank you for conversation.
When using an R2R DAC in NOS you should indeed select the correct number of bits. Not doing so will not reproduce small signal content as accurately.
If you send 24 bits, you'll get poorer performance than you'd get by setting it correctly at say 15 bits, because the noise shaper is operating beyond the linear range of your DAC and as a result it's effectively not working correctly. It's based on maths that assumes your DAC can accurately reproduce what it's outputting.
As an example, using a DAC that hasn't been discussed yet just for the sake of avoiding argument, here's the Hifiman EF400 when being fed data with the noise shaper running at 15 bit, playing a -120dB sine:
Now here's it doing the same but with the noise shaper set to 24 bit:
Significant distortion, and in fact the amplitude of the signal itself is also now not reproduced accurately.
To find the correct number of bits, the safe method is to just take the THD+N figure in dB, say -94 for example, and divide by 6. So -94dB divided by 6 is 15.66. So you'd choose 15 bits.
If you have a very low noise ADC, you can instead just test manually to find the limit. Play a 1khz sine through HQP, set vol to -120dB, and increase noise shaper bits by 1 until you suddenly see distortion as shown above. That means the value you set exceeds the linear range of your DAC, so go back 1 and then you're good.
With oversampling DACs or delta-sigma DACs they will be running their own noise shaper internally anyway so this setting won't really have much (or usually any) effect, and you can just set it to 24 for ease
If you send 24 bits, you'll get poorer performance than you'd get by setting it correctly at say 15 bits, because the noise shaper is operating beyond the linear range of your DAC and as a result it's effectively not working correctly. It's based on maths that assumes your DAC can accurately reproduce what it's outputting.
As an example, using a DAC that hasn't been discussed yet just for the sake of avoiding argument, here's the Hifiman EF400 when being fed data with the noise shaper running at 15 bit, playing a -120dB sine:
Now here's it doing the same but with the noise shaper set to 24 bit:
Significant distortion, and in fact the amplitude of the signal itself is also now not reproduced accurately.
To find the correct number of bits, the safe method is to just take the THD+N figure in dB, say -94 for example, and divide by 6. So -94dB divided by 6 is 15.66. So you'd choose 15 bits.
If you have a very low noise ADC, you can instead just test manually to find the limit. Play a 1khz sine through HQP, set vol to -120dB, and increase noise shaper bits by 1 until you suddenly see distortion as shown above. That means the value you set exceeds the linear range of your DAC, so go back 1 and then you're good.
With oversampling DACs or delta-sigma DACs they will be running their own noise shaper internally anyway so this setting won't really have much (or usually any) effect, and you can just set it to 24 for ease
Last edited:
sajunky
Headphoneus Supremus
Logical with small error, as it assume that there is no dithering applied in a DAC and you refer to a real linear range of a DAC. Replace with:
That means the value you set including amount of dither applied in a DAC exceeds the linear range of your DAC.
... and will be fine.
Users who are viewing this thread
