DIYHiFi.org

For the sake of audio
It is currently Mon Dec 18, 2017 4:20 pm

All times are UTC




Post new topic Reply to topic  [ 291 posts ]  Go to page 1, 2, 3, 4, 5 ... 20  Next
Author Message
PostPosted: Tue Apr 01, 2014 9:09 pm 
Offline
Cow

Joined: Tue Apr 01, 2014 1:47 am
Posts: 131
Location: Knoxville, Tn
Hello all. Newbie to the forum here. Thorsten suggested I sign up. I have been blowing up his email with questions for the last few weeks. I am sure he (and Vince) will be glad that I am not blowing up their inbox with 20 emails a day! :D

On to the question. I know that many of you have signed NDA's with various companies, so in that interest I will be talking in hypotheticals about black boxes.


Assume a certain DAC, any DAC, has a proprietary multi level delta sigma modulator. It converts all formats to a single format best suited for its modulator. It also provides volume control and digital filtering on this signal including DSD.

I have a laymans "theory" on how this works (specifically the volume control) and need to know how accurate or inaccurate (most likely) it is.

DSD is a one bit differential signal, correct? Differential in that a stream of equal ones and zeros will equal a 50 percent amplitude. What if instead of representing full on with one bit, I use 32. I could then conceivably create a 25 percent signal, without ever changing the sample rate or changing the differential nature of the system, by simply using 16 bits of the 32 bit pulse for "on". Or whatever level you choose. You are simply limiting the voltage of the full on pulse.

Now, I have no idea if this is how it works, or if this is fundamentally flawed. In a simplistic way, it seems to work. Of course, this processed signal would be in our hypothetical DAC put through the proprietary delta sigma modulator before output.

Please, let me know what is correct or incorrect about my ideas, and what may really be going on in DACS that process DSD before conversion.

Thanks, and I look forward to learning as much as I can!!

Andrew


Top
 Profile  
Reply with quote  
PostPosted: Wed Apr 02, 2014 10:42 am 
Offline
Muriel
User avatar

Joined: Thu Jul 26, 2007 5:45 pm
Posts: 1168
Andrew,

ball3901 wrote:
Assume a certain DAC, any DAC, has a proprietary multi level delta sigma modulator. It converts all formats to a single format best suited for its modulator. It also provides volume control and digital filtering on this signal including DSD.


The NDA covered datasheet contains nothing useful to decipher the internal workings...

What we know from public domain sources is this:

1) ES901X accepts 32Bit PCM, DSD or SPDIF
2) SPDIF must be decoded to PCM
3) The internal processing uses an ASRC which in turn incorporates a digtal filter
4) The digital filter is claimed by ESS to work at 32 Bits
5) The digital Volume control is claimed by ESS to work at 32 Bit Precision
6) The digital filter and ASRC can be bypassed, if you do this DSD no longer works

ESS have a public domain product brief:

http://www.esstech.com/PDF/ES9018%20ES9012%20Product%20Brief.pdf

It includes this diagram:

Attachment:
Untitled-1.jpg


I am sure following all the above one may still doubt what really goes on inside these black boxes.

However I am comfortable to suggest that everything take in is first converted to 32BIt PCM at whatever internal sample rate is produced by the system clock (at a much lower rate than the system clock, e.g. Fsys/256 which would explain why 384KHz needs a 100MHz clock) and is then filtered, volume controlled etc. as such a 32 Bit PCM signal, after which it is rendered as Multibit Delta Sigma into the output switches.

If I am really grievously wrong, I would invite Martin Mallinson or Dustin Forman to correct me. They may even e-mail me personally and I will happily eat crow and post the rebuttal myself...

And if all this digital jiggery pokery amounts to more than a heap of beans - I have a number of commercial ES9018 designs at hand, I have implemented other ESS Chips as well, they are very nice DAC's and sound quite decent. There is much worse out there.

The way things have been implemented makes it VERY easy to design and build a DAC with something like an XMOS interface that allows SPDIF, USB PCM and DoP PCM using minimal extra parts. Just follow the App notes and reference designs and you are good to go, no effort, knowledge etc. required, results are pretty much guaranteed to be very good (measured at least). If the guys at AMR and iFi would let me, I'd "design" not but ESS based DAC's, come to work at 11 and not come back from the Pub after lunch and still get all the designs scheduled done...

Kind regards T


You do not have the required permissions to view the files attached to this post.

_________________
"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." Richard Feynman


Top
 Profile  
Reply with quote  
PostPosted: Wed Apr 02, 2014 4:29 pm 
Offline
Cow

Joined: Tue Apr 01, 2014 1:47 am
Posts: 131
Location: Knoxville, Tn
I believe I realized the fundamental flaw in my reasoning. The signal is converted back 6 bits (I think) to meet the needs of the delta sigma modulator. Therefore any high resolution volume control implemented in the manner I describe would in the end be no more than a 6 bit system with 36db of range in 6db steps!!

The chip in question certainly has a more resolute volume control than that! So I would think that yes, the signal must be converted to something that is absolutely not a differential DSD signal.

Thanks for the input as always!

Andrew


Top
 Profile  
Reply with quote  
PostPosted: Wed Apr 02, 2014 4:36 pm 
Offline
Cow

Joined: Tue Apr 01, 2014 1:47 am
Posts: 131
Location: Knoxville, Tn
And yes, I am going to try a new hobby. A DIY DAC build. I had already come to the conclusion that an ESS chip is the place to start. Seemed very easy to implement. Buy a decent XMOS USB to I2S converter to feed it and I think you have mostly a complete design.

(I also had a crazy idea implementing the USB board in a separate enclosure with its own power supply. Very much like a USB to SPDIF converter, except I was going to create a proprietary I2S output, using some kind of frankensteined cable... Cat 6 or something like that... may be a completely nutty idea filled with problems. But it sounds good in my head)


Top
 Profile  
Reply with quote  
PostPosted: Wed Apr 02, 2014 4:43 pm 
Offline
Cow

Joined: Tue Apr 01, 2014 1:47 am
Posts: 131
Location: Knoxville, Tn
And furthermore, I saw one graph of a DSD encoded squarewave from the chip in question. I am not really sure what it was supposed to look like, but it did NOT look like other graphs of DSD encoded signals that I have seen.


Top
 Profile  
Reply with quote  
PostPosted: Wed Apr 02, 2014 5:14 pm 
Offline
Benjamin
User avatar

Joined: Mon Dec 13, 2004 3:29 pm
Posts: 4333
Location: Somewhere other than here.
T..................you have my sympathy.


Top
 Profile  
Reply with quote  
PostPosted: Wed Apr 02, 2014 6:08 pm 
Offline
Cow

Joined: Tue Apr 01, 2014 1:47 am
Posts: 131
Location: Knoxville, Tn
Well, I wouldn't offer him too much sympathy on my account. At least I thoroughly enjoy his designs, and will pay good money to keep buying them.

Offer him sympathy for the 1000 questions that I am sure are coming in daily about his product's "measurements", according to one of those other geeky companies..




Jocko Homo wrote:
T..................you have my sympathy.


Top
 Profile  
Reply with quote  
PostPosted: Wed Apr 02, 2014 7:48 pm 
Offline
Muriel
User avatar

Joined: Thu Jul 26, 2007 5:45 pm
Posts: 1168
Andrew,

ball3901 wrote:
And yes, I am going to try a new hobby. A DIY DAC build.


Do you remember those medieval maps? For areas they did not really know, they tended to picture fabulous beasts and anotate "that way be dragons".

As someone who had to explore these unknown regions I can confirm what the Map's say.

That way be dragons.

ball3901 wrote:
I had already come to the conclusion that an ESS chip is the place to start.


The best choice for the ES9018 is to use a Transformer, 1:1 - as analogue stage, if you want to run 100MHz clocks and also crank up the analogue supplies.

If you know Caledonii Pejorative's special trick for the AD797 (which AFAIK is used by that resonant something crowd) you can make a passable Op-Amp analog stage, but ONLY if you keep the clock speed low.

Other ways exist.

Currently I like best a vintage pair of Stevens & Billington TX-103 Silver MC Step-up's with a handmade silver wire (Audio Note enameled silver) non-inductive wirewound I/U conversion resistor and followed by a 5687WB White Follower, with DC servo so no coupling cap. Still not quite right.

I usually play the iDSD mini prototype in my Office system instead.

ball3901 wrote:
I believe I realized the fundamental flaw in my reasoning. The signal is converted back 6 bits


Not really. Based on what old Mr. HP 'scope (150MHz/2GS) and new Mr. Rigol 'scope (300MHz/2GS) tell me it is converted into 64 non-identical and parallel delta Sigma Datastream's. But I may be wrong.

ball3901 wrote:
And furthermore, I saw one graph of a DSD encoded squarewave from the chip in question. I am not really sure what it was supposed to look like, but it did NOT look like other graphs of DSD encoded signals that I have seen.


If you want a similar graph with any funloving decent PCM DAC, set J-River to convert DSD to PCM and adjust the filter bandwidth to "permissive". Even your iDSD can do it if you do that.

But that misses the point of having DSD in the first place (presuming there is one).

If you see a DSD graph without all the tell-tale noise (+6dBFS above 1.4MHz, rising from around 10KHz - if no external or digital filter used) you know it has been filtered aggressively and likely been turned to PCM with a high order lowpass in the digital domain.

Ciao T

PS, I'm jammed out, this is as much time as I have...

_________________
"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." Richard Feynman


Top
 Profile  
Reply with quote  
PostPosted: Thu Apr 03, 2014 1:53 am 
Offline
Cow

Joined: Tue Apr 01, 2014 1:47 am
Posts: 131
Location: Knoxville, Tn
Thanks for your time, T. I think my idea of a volume controlled DSM would be great though. Any chance that we would see a 32 bit, 4 million plus level modulator anytime soon? lol

As you can tell, I am realizing even more how ridiculous it must sound.


Top
 Profile  
Reply with quote  
PostPosted: Thu Apr 03, 2014 3:31 am 
Offline
Cow

Joined: Tue Apr 01, 2014 1:47 am
Posts: 131
Location: Knoxville, Tn
And one more thing.

My next major audio upgrade is a DAC. I have been holding out for a long, long time. DSD is/was a major factor. But, I am warming to the idea of ditching DSD.

What would you say about DSD converted to 16/44.1 and played back on the DP-777 in bit perfect mode? Sounds like heresy. But speaking of sound, I am wondering if in actual listening, it would sound sublime?


Top
 Profile  
Reply with quote  
PostPosted: Thu Apr 03, 2014 10:53 am 
Offline
Muriel
User avatar

Joined: Thu Jul 26, 2007 5:45 pm
Posts: 1168
Hi,

ball3901 wrote:
Any chance that we would see a 32 bit, 4 million plus level modulator anytime soon?


Not 32 bit, if you can live with 20-24 Bit's, that has been around for ages. It is called PCM. :finga:

Get a Pacific Microsonic Model 2 if you want to hear what PCM can actually do.

Or download the HrX demo tracks from Reference Recordings and have a listen using your iDSD nano...

ball3901 wrote:
What would you say about DSD converted to 16/44.1 and played back on the DP-777 in bit perfect mode?


Actually, I use 176.4KHz/24Bit for DSD playback. Using Bitperfect 1 Mode, DSD filter in J-River set to "permissive". Still prefer real PCM, but is preferable to any DSD DAC I have at hand playing DSD... The iDSD mini prototype I have around since it came back from Japan is a pretty nice machine too though.

Personally, for my own money and without employee discount and all that I'd probably have the iDSD mini over the DP-777. Not that the DP-777 is not notably better, but the difference in price is much, much larger.

Ciao T

_________________
"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." Richard Feynman


Top
 Profile  
Reply with quote  
PostPosted: Thu Apr 03, 2014 1:02 pm 
Offline
Cow

Joined: Tue Apr 01, 2014 1:47 am
Posts: 131
Location: Knoxville, Tn
ball3901 wrote:
Thanks for your time, T. I think my idea of a volume controlled DSM would be great though. Any chance that we would see a 32 bit, 4 million plus level modulator anytime soon? lol

As you can tell, I am realizing even more how ridiculous it must sound.


4 billion. billion. lol. Late night (early morning posting) is good for mistakes.


Quote:
Personally, for my own money and without employee discount and all that I'd probably have the iDSD mini over the DP-777]


Thats the product I have been eyeballing. And waiting on. Indeed, the DP-777 would mean a few less meals. But, I am an impatient dude. :)


Top
 Profile  
Reply with quote  
PostPosted: Thu Apr 03, 2014 4:23 pm 
Offline
Muriel
User avatar

Joined: Thu Jul 26, 2007 5:45 pm
Posts: 1168
Hi,

ball3901 wrote:
Indeed, the DP-777 would mean a few less meals. But, I am an impatient dude. :)


Who am I to blow against the wind?

Hey, I run my TV sound through my DP-777.

You can take it from me right after my AK-47.

And the only way you get that is if I'm flat out of ammo and likely breathing through a few extra holes in my chest. I always used to carry at least 6 clips taped jungle style back in the paramilitary, plus loose rounds. Hate running out of ammo. :eek:

Only thing worse than being shot at is being shot at and having to throw your field rations as retaliation in the hope they eat them and it poisons the fatherless ones, because you did not bring enough ammo to the fight. :finga:

So how can I tell you you should not have one? Go for it bad boy. You deserve. Who needs to eat anyway? :D

Ciao T

_________________
"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." Richard Feynman


Top
 Profile  
Reply with quote  
PostPosted: Fri Apr 04, 2014 1:58 am 
Offline
Cow

Joined: Tue Apr 01, 2014 1:47 am
Posts: 131
Location: Knoxville, Tn
Quote:
Actually, I use 176.4KHz/24Bit for DSD playback]



may I ask why you choose this rate over say, 88.2khz ?

Thanks

Andrew


Top
 Profile  
Reply with quote  
PostPosted: Fri Apr 04, 2014 3:08 am 
Offline
Muriel
User avatar

Joined: Thu Jul 26, 2007 5:45 pm
Posts: 1168
Hi,

ball3901 wrote:
Quote:
Actually, I use 176.4KHz/24Bit for DSD playback]


may I ask why you choose this rate over say, 88.2khz ?


Double the time domain resolution. You can easily compare DSD converted to either rate on your iDSD.

Ciao T

_________________
"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." Richard Feynman


Top
 Profile  
Reply with quote  
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 291 posts ]  Go to page 1, 2, 3, 4, 5 ... 20  Next

All times are UTC


Who is online

Users browsing this forum: No registered users and 2 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB® Forum Software © phpBB Group