Latest DSD Decoder Project

What people are working on at the moment
User avatar
ed
retired
Posts: 5384
Joined: Thu Jun 21, 2007 4:01 pm
Location: yorkshire
Contact:

#16 Re: Latest DSD Decoder Project

Post by ed »

one thing has always intrigued me about dsd, and I've never found an explanation. How do you record silence with dsd?
There's nowhere you can be that isn't where you're meant to be
User avatar
Nick
Site Admin
Posts: 15707
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#17 Re: Latest DSD Decoder Project

Post by Nick »

Continuous stream of 1 and 0 would do it. Looking at the noise floor of DSD, it may be you get the odd 1 1 0 0 and so on. Depends on the filter used to create it. I have never found an actual definition of the encoder.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
Ant
Shed dweller
Posts: 2332
Joined: Mon Jul 31, 2017 6:45 pm
Location: Yorkshire

#18 Re: Latest DSD Decoder Project

Post by Ant »

pre65 wrote: Fri Jun 21, 2019 3:16 pm
Ray P wrote: Fri Jun 21, 2019 3:11 pm The tracking data shows that the replacement Cronus reclocker arrived at Heathrow on Monday - it's still there! That's longer than it took to get from the US to the UK. Grrr...
That's because a multi national task force (US + UK + Parcelthief) are trying to work out how much to charge you for import duty and VAT on a FOC item. :lol:
They charged me 100 quid import duty on something that was a return from the us...... Because they missed the tick in the box that marked it as a return.
Arseholes
Also starring Rex Hamilton as Abraham Lincoln

www.bte-designs.weebly.com
User avatar
Ray P
No idea why I do this anymore
Posts: 6294
Joined: Thu Nov 22, 2007 5:18 pm
Location: Somerset

#19 Re: Latest DSD Decoder Project

Post by Ray P »

pre65 wrote: Fri Jun 21, 2019 3:16 pm That's because a multi national task force (US + UK + Parcelthief) are trying to work out how much to charge you for import duty and VAT on a FOC item. :lol:
Yes, I think so Phil. They stitch you up every way. If you put zero value on the customs form they refuse to offer insurance, if you put a value they charge you for it! In this instance, Twisted Pear are out of pocket because they're having to supply a replacement so it has value to them and they want to be insured. To me, I've already paid for the item (and the duty and Royal Mails extortion charge) so it should be zero value. Grrr again.
Sorry, I couldn't resist!
User avatar
ed
retired
Posts: 5384
Joined: Thu Jun 21, 2007 4:01 pm
Location: yorkshire
Contact:

#20 Re: Latest DSD Decoder Project

Post by ed »

Nick wrote: Fri Jun 21, 2019 4:36 pm Continuous stream of 1 and 0 would do it. Looking at the noise floor of DSD, it may be you get the odd 1 1 0 0 and so on. Depends on the filter used to create it. I have never found an actual definition of the encoder.
yes, I thought that as well...for a long time...then lying in bed one night I pondered the abrupt silence in the middle of a crescendo whereby there is instant silence with no tail off......say from 110db to 0db between samples....surely the 110db would continue...I'm working on the premise that there is a sound that stops without a decay.
There's nowhere you can be that isn't where you're meant to be
User avatar
Ray P
No idea why I do this anymore
Posts: 6294
Joined: Thu Nov 22, 2007 5:18 pm
Location: Somerset

#21 Re: Latest DSD Decoder Project

Post by Ray P »

ed wrote: Fri Jun 21, 2019 9:04 pm ....surely the 110db would continue...
No, what you hear is the silence...
Sorry, I couldn't resist!
User avatar
Nick
Site Admin
Posts: 15707
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#22 Re: Latest DSD Decoder Project

Post by Nick »

ed wrote: Fri Jun 21, 2019 9:04 pm
Nick wrote: Fri Jun 21, 2019 4:36 pm Continuous stream of 1 and 0 would do it. Looking at the noise floor of DSD, it may be you get the odd 1 1 0 0 and so on. Depends on the filter used to create it. I have never found an actual definition of the encoder.
yes, I thought that as well...for a long time...then lying in bed one night I pondered the abrupt silence in the middle of a crescendo whereby there is instant silence with no tail off......say from 110db to 0db between samples....surely the 110db would continue...I'm working on the premise that there is a sound that stops without a decay.
In the digital world there is no instant. Remember that before encoding the input signal is band limited. So the trailing edge of the signal you are talking about will return to zero at a rate limited by the low pass filter its gone through.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
Max N
Old Hand
Posts: 1453
Joined: Thu Aug 16, 2007 6:10 pm

#23 Re: Latest DSD Decoder Project

Post by Max N »

Which I guess is one of the reasons higher sample rates or bit rates sound better?
User avatar
Ray P
No idea why I do this anymore
Posts: 6294
Joined: Thu Nov 22, 2007 5:18 pm
Location: Somerset

#24 Re: Latest DSD Decoder Project

Post by Ray P »

Nick wrote: Fri Jun 21, 2019 11:54 pm In the digital world there is no instant.
I'm not sure there is in the everyday real world either.

Taking Ed's example of the orchestral crescendo with an abrupt silence, although we hear it as an instantaneous silence it won't be really, not least because even the best drilled musicians will have tiny timing differences so the crescendo will actually decay to a silence, albeit very quickly, and that's without considering that the instruments themselves don't just suddently stop vibrating, or the sound waves in the venue simply vanish.

So I guess the question is simply about whether the digitial (or analogue) recording/playback system can accurately track that decay in the context of our hearing capability.
Sorry, I couldn't resist!
User avatar
Ray P
No idea why I do this anymore
Posts: 6294
Joined: Thu Nov 22, 2007 5:18 pm
Location: Somerset

#25 Re: Latest DSD Decoder Project

Post by Ray P »

Anyway, back to the DSD project.

The DIY Audio member who has been developing the DSD decoder project is very much in the opensource camp and has placed gerber files for the latest version in the public domain; it's basically the same as mine but with integrated isolation and reclocking for the outputs from either an Amanero or Beaglebone Black (for which he's developed a DSD specific linux-based software distro) that sit as sub-boards on the DSD board. There's more info here;

https://puredsd.ru/

and here;

http://www.audiodesignguide.com/DSC1/index.html

Very early days so no promises but I've been thinking about getting a small batch of boards assembled - would there be any potential interest here? I think the cost would likely be in the region of £150-£250.
Sorry, I couldn't resist!
User avatar
Nick
Site Admin
Posts: 15707
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#26 Re: Latest DSD Decoder Project

Post by Nick »

Max N wrote: Sat Jun 22, 2019 9:22 am Which I guess is one of the reasons higher sample rates or bit rates sound better?
Well, its much more complex with DSD than PCM. All PCM has to do is go from one level to a zero in the space of two samples. DSD has to emit 0's to ramp the output down over a number of samples. Of course the DSD samples are occurring more often that the the PCM samples. But it does show how DSD is lossy. PCM 44k1 could go from full scale 0x7fff, to zero 0x0000 in one sample (a jump of 32767). A simple one bit modulation would have to run at 44100 * 32767 Hz to do the same (1.44Ghz). The filtering inside DSD means that that transition would never happen.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
User avatar
ed
retired
Posts: 5384
Joined: Thu Jun 21, 2007 4:01 pm
Location: yorkshire
Contact:

#27 Re: Latest DSD Decoder Project

Post by ed »

Nick wrote: Sat Jun 22, 2019 10:29 am
Max N wrote: Sat Jun 22, 2019 9:22 am Which I guess is one of the reasons higher sample rates or bit rates sound better?
Well, its much more complex with DSD than PCM. All PCM has to do is go from one level to a zero in the space of two samples. DSD has to emit 0's to ramp the output down over a number of samples. Of course the DSD samples are occurring more often that the the PCM samples. But it does show how DSD is lossy. PCM 44k1 could go from full scale 0x7fff, to zero 0x0000 in one sample (a jump of 32767). A simple one bit modulation would have to run at 44100 * 32767 Hz to do the same (1.44Ghz). The filtering inside DSD means that that transition would never happen.
well that was my point really, although a bit naive I'll admit. How many 0 bits to squeeze into 2.8mhz to mimic an abrupt silence, and if/when you do get there then there is no provision in dsd for inky black.....you'd need a quantum bit, neither 0 nor 1, or both......

naive or not, my money is still on pcm for the future, although I'd be the first to fess up that I couldn't tell the difference.
There's nowhere you can be that isn't where you're meant to be
User avatar
Nick
Site Admin
Posts: 15707
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#28 Re: Latest DSD Decoder Project

Post by Nick »

you'd need a quantum bit, neither 0 nor 1, or both......
Well, neither is it in a dithered stream.

DSD's point is that you will never get a full-scale to 0 transition in that short a space of time, so the filtering is designed to cope with what is expected to happen.Its a thread about someone building DSD decoder so its not the place for me to go off on a rant about it. Generally IMHO there is a lot of rubbish out there about digital . My biggest complaint is the scare story about "pre-echo" in digital filters and how its unnatural as if there is some form of time travel going on. mutter mutter grump grump.
although I'd be the first to fess up that I couldn't tell the difference.
I think I can, but maybe I have not heard enough DSD decoders.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
User avatar
Ray P
No idea why I do this anymore
Posts: 6294
Joined: Thu Nov 22, 2007 5:18 pm
Location: Somerset

#29 Re: Latest DSD Decoder Project

Post by Ray P »

It's a shame I was too much of a numpty to find the problem with my reclocker in time to get it finished for Owston; it would have been interesting to demo it at DSD256 and/or DSD512 and see what people think.

My personal experience is that I prefer DSD, which is why I'm working on this project.

I may get it working quicker than I expected though as the replacement reclocker board was delivered a little earlier and I didn't have to pay anything so I take back some of my earlier moaning; not all though as the tracking info still shows it at Heathrow.

I'll install and test the new reclocker and if all is well I'll put the rest of the project back together and see if the software works as expected.
Sorry, I couldn't resist!
Wolfgang
Old Hand
Posts: 229
Joined: Mon May 07, 2018 3:08 pm

#30 Re: Latest DSD Decoder Project

Post by Wolfgang »

I hope I don’t upset anybody or go too much “off-topic” with this comment but in my limited understanding the strength of the Soekris R-2R DAC lies in the possibility to adjust the balance between low pass filter linear phase and (unavoidable) passband ripples which can be described as pre- and post echoes determined by the impulse response of the filter. I think that’s what the “filter brewing thread” in diyaudio is trying to achieve. A simple low pass filter of a DSD DAC doesn’t have this option.And the practical listening experience is much more affected by these" filter issues " than by some limitations of DSD to reproduce silent passages accurately.
Post Reply