1 bit DSD DAC implementations

[lateboomer]

Summarizing the papers presented against DSD, the core of the argument is that you cannot dither a 1 bit format. You need more than 1 bit for dither to work otherwise you are randomizing your one and only bit! Without dither, the device cannot be linear, and without (correlated) quantization noise. Doubling or quadrupling sample rate doesn't fix this problem.

What is the purpose of dithering? So it is flawed in term of 1 bit DSD cannot be dithered. But is it flaw in term of sound reproduction, which is the most important part? I am common peasant trying to understand complicated digital with you smart people.

 

[Mivera]

What is the purpose of dithering? So it is flawed in term of 1 bit DSD cannot be dithered. But is it flaw in term of sound reproduction, which is the most important part? I am common peasant trying to understand complicated digital with you smart people.

As a common peasant, what's been your take on Hqplayer upsampling?

 

[lateboomer]

I like it a lot. Music is more easy flowing and less digital like. But has to combine with fiber optics and good power supply to extract the best result. One crucial part is also what kind of dac being used. So far I have tried V90 dac, M2tech Young, BDA-2 and Phison DIY dac.

 

[Mivera]

I like it a lot. Music is more easy flowing and less digital like. But has to combine with fiber optics and good power supply to extract the best result. One crucial part is also what kind of dac being used. So far I have tried V90 dac, M2tech Young, BDA-2 and Phison DIY dac.

Your opinion matches most. You mean Mirand DIY DAC. The Phison line is a different line of products.

 

[lateboomer]

Yeah, it is Mirand DIY dac, also I have got M1 DAC. One thing, HQplayer with very good and competent NAA hardware is a must to achieve good sound.

 

[Mivera]

Yeah, it is Mirand DIY dac, also I have got M1 DAC. One thing, HQplayer with very good and competent NAA hardware is a must to achieve good sound.

Yes the hardware makes a big difference. Although the galvanic isolation/reclocker on the PD2, takes a lot of that importance out of the equation. My 1st client is going to be testing the new Sonore Microrendu with it as soon as it's released. We will see how much of a difference it makes as they are touting it as the ultimate NAA.

 

[amirm]

What is the purpose of dithering?

Dither is what makes digital become analog! As we all know, digital has steps that it captures from analog. A 4-bit format for example can represent just 16 levels. Clearly that is far less than our analog waveform. If we don't get rid of them, they will create pretty bad distortion. Dither is just random noise. If we add that *prior* to digitizing the signals, we convert the distortions to noise. We trade noise for distortion which analog does the same.

Now we don't want a ton of noise so we apply that to the lowest order bits of our digital samples. With 16 or 24 bit PCM, that is just fine as losing a bit or two is not a big deal. In a 1-bit format, we don't have any low order bits to donate to dither.

That said, if the system is high enough resolution we can take advantage of natural noise in our incoming analog signal to act as dither. Indeed there are people who make the argument that when converting 24 bit PCM to 16 bit, we don't need to add dither because the 24 bit PCM never has useful data in its low-order bits because no ADC can capture to that fine of a granularity. So DSD gets a free pass just the same.

 

[lateboomer]

Thank you Amir for the explanation. It certainly help me to understand dither better than just only refer to wiki info which is much more complicated to me to sort out essential points.

 

[amirm]

My pleasure. That is one of the core missions of this forum . I also found a couple of very nice videos that explain and show the concept and put them here: http://www.audiosciencereview.com/forum/index.php?threads/what-is-audio-dither.141/