World's first NOS 24/384 filterless DAC

[Telstar]

Here's another example : Faith from George Michael. I mention this one, because it occurred to me that it wouldn't work as I wanted before. You could also say that I didn't get the feelings with it, I recalled from when being a tad younger, women involved. It was too sterile and had no slam. It didn't have the power the tough guy on the cover expresses, no matter how gay he is.

In short George Michael is not Boy George 

[pedal]

Lately, my main music source are hi-rez files from various disks and downloads. Sampling frequencies are mainly 96k, but also 88k2, 176k4 and 192kHz. It strikes me that a filterless DAC is perfect for hi-rez files, because they have a natural roll-of with very little(!) amplitude left above, let's say, 40kHz. So no *need* for low pass filtering, because the aliasing distortion will be very low anyway.

[Calibrator]

Looking forward to "official" photos of ready-to-ship product.

The official release pictures will be out shortly, but here's a sneak peak of Peter working on an early version. It clearly shows he's been working on the experimental NOS in a compact form for some time now.























 

[PeterSt]

Lately, my main music source are hi-rez files from various disks and downloads. Sampling frequencies are mainly 96k, but also 88k2, 176k4 and 192kHz. It strikes me that a filterless DAC is perfect for hi-rez files, because they have a natural roll-of with very little(!) amplitude left above, let's say, 40kHz. So no *need* for low pass filtering, because the aliasing distortion will be very low anyway.

Hey hey ... ho ho ...

Who informed you that a filterless DAC is "perfect" for hires ?
(but let me softly add that a while back we talked offline about these things IIRC, and maybe there I said or implied something like that ?)

So no, this is not what I want to express, and I hope I never did somewhere. And as a matter of fact, the whole discussion since (0.9z) hires upsampling (filtering !) is that it should NOT happen. IOW, no filter ! But as an exception Arc Prediction is allowed, because it doesn't imply a roll off; it only increases (fake) resolution.

But since you brought up the subject anyway, let's be confusing.
I think it would be true that applying an Oversampling DAC onto hires material is extra-wrong. This is not so much about the "filter", but merely about the oversampling which just is not needed for hires. So here you go again, because now it is not only the wrong means of filtering which *will* be applied for OS, but also the destroying oversampling.

On this very last, try to see this : As long as no "upsamplers" can be found which definitely make the sound better because of better (but fake) resolution, any means of upsampling is wrong. So, for hires this for sure counts, because you will be manipulating the data while it is not necessary at all. So yes, this is where NOS comes into play, because it doesn't touch anything. That is, if you don't apply any means of filtering, which includes (no) upsampling. But :

As long as no "upsamplers" can be found which definitely make the sound better because of better (but fake) resolution, any means of upsampling is wrong.

with Arc Prediction as the exception, unless suddenly everybody disagrees with this.
But then still ... Arc Prediction is made for letting it loose on a DAC which doesn't again does something to the data, and here we end up again at the necessity of NOS 24/192 (or better). 16 bits can't do it and 96 can't do it.

I hope this is somewhat more clear than confusing ...
Peter

[pedal]

Hey hey ... ho ho ...

Who informed you that a filterless DAC is "perfect" for hires ?
(but let me softly add that a while back we talked offline about these things IIRC, and maybe there I said or implied something like that ?)

Thanks for your interesting reply.

-Actually it came to my mind while using my new software Adobe Audition 3, doing a spectrum analyze of several hi-rez files. Then I realised there is not much(!) HF content above, lets say, 25kHz. Close to the half of the samplingfrequency (48, 88 or 96) there is practically speaking only noisefloor.
[Even if the digital recording technology can capture signals >50kHz, the mics does not. If we are speaking about music recorded in the haydays of analog, then most multitrack mastertape machines rolled of above 20kHz, too].
 
Conclusion: *Nothing* left to filter at sf/2 when the music file has 88k2Hz or higher sampling frequency. The less lowpass filtering, the better. I look forward to hear you DAC!

[goldkenn]

Hi all, I am new to this forum, I found here just when searching in google for anything about NOS DAC, and this project I really have great interesting.

I have not yet read all posts, it is a bit difficult for me since I am from Hong Kong, English not my mother language.

I would like to ask is this project fionally will open sell as kit or built?

May I ask the designer for a favour, mind to keep me update via email? I am eager to buy one, thanks 

[PeterSt]

Hi there - Thank you kindly for your interest !

Although I started this myself as a DIY project, there will be a built version only. Let's say this is because it is meant to be the best (for sound quality) ever, and I rather guarantee that by our own building and measurements here, than leave it up to you with the possibility of making mistakes ending up in unsolveable hum or noise, or worse : no hum or noise but just poor sonic performance, while you don't know it is more poor than intended (you don't have the reference). And not to forget : nothing is normal in this DAC, and if it doesn't sound good I'd rather have that you blame me, than that I blame you for an unknown mistake somewhere while it still could be our mistake here.

I don't maintain something like an email (news) list, but I think it suffices to keep a watch on this topic; When something significantly changes you will read it in here, although I must admit that sometimes things pop up in other topics "by accident".


A general remark for everyone :
Last Monday I was notified by the company which builds the cabinet that the promised proto which should be ready around today, is postponed until the week of July 19. Next they foresee that because of vacations the final cabinet won't be ready before 5 weeks after that (notice that this is related to a couple of "external" rounds like painting and printing).
In the end this is my own fault I think, because I postponed the assignment for making the cabinet "forever", but which by itself was related to the unknown of what had to be in there in detail (and this is final for only a few weeks now). And so I dropped myself right in the middle of the vacation period. Well, at least this gives some space to the other (3) suppliers who are all late now, while the assembly of the parts concerned will take around one week anyway (for the first 100 DACs).

Peter

[goldkenn]

Thank you Peter, a built version will be fine but if a full, or semi kit, or PCB available will be more fun, since I can have my choice of components such as caps, and maybe saving cost by using some my stock parts, as a qualified electronic engineer, I have confidence to face any problem you worry about, although I usually stuck with analog circuit but I do design amplifiers, and nothing low end.

Anyway, I will stuck with the post and attention for any update.

As told I am from Hong Kong, do you having any preparation for international payment such as via PayPal?

[pedal]

Peter:

Today I use PC > Fireface400 > spdif coax cable > DAC. [which is superior SQ compared to USB adaptive mode, BTW].

But you are using PC > DIY modification > I2S cable > DAC. (Is that correct?)

I don't know if you have mentioned it in the past, but do you consider your DIY 12S solution superior to SPDIF?
Eventually, what have you done to get the I2S output from your PC?

(Probably you DAC will accept several digital inputs, but the whole point about your DAC is to get the best sound quality. Right?)

All the best.

[AUDIODIDAKT]

I2S is always superiour to spdif,

I2S is a so called chip-to-chip connection, the output ic from the souncard gets directly connected with the receiver chip in the dac.
This means that data and bit- and wordclock signals stays separated were spdif merges these together and get separated again at the dac side.
I2S has a 3 wire configuration while spdif just uses 1. (this is strange because a cd mechanism outputs I2S)
This skips the need for PLL at the DAC side, wow, less jitter better sound, but this is just half the story, add a good master word-clock and the fun just begins.
Although I think the length of the I2S connection should remain as short as possible.

A simple mod to for e.g. an ESI Juli@ soundcard, will give you I2S output from your pc. (if you have a dac that can input I2S ofcourse)

Cheers,

Roy

[PeterSt]

Today I use PC > Fireface400 > spdif coax cable > DAC. [which is superior SQ compared to USB adaptive mode, BTW].

But you are using PC > DIY modification > I2S cable > DAC. (Is that correct?)

Yes, that is correct.

Eventually, what have you done to get the I2S output from your PC?

Generate it somewhere outside of the PC.
But did that in a fashion that no clock (software or hardware) in the PC plays any role.

I don't know if you have mentioned it in the past, but do you consider your DIY 12S solution superior to SPDIF?

SPDIF sounds completely raw in comparison. "Raw" indeed is the best description, although you will perceive it like that once you heard "the other way".
Notice that we can't speek of "an i2s connection" in general, because something like that can't exist. This by itself is because it isn't a standard for external connection, and thus it is your "implementation" which makes or breaks it. What Roy suggested works too but it can only be far from optimal. Still better than SPDIF though.
i2s is just the description of 3 wires, while two of them already aren't related to the logical connection in the sense "must come from one source". This leaves one other of which you methematically can't say "must come from one source".

[pedal]

Generate it somewhere outside of the PC.
But did that in a fashion that no clock (software or hardware) in the PC plays any role.

Would you like to share the recipe with us? 

[PeterSt]

Thank you Peter, a built version will be fine but if a full, or semi kit, or PCB available will be more fun, since I can have my choice of components such as caps, and maybe saving cost by using some my stock parts, as a qualified electronic engineer, I have confidence to face any problem you worry about, although I usually stuck with analog circuit but I do design amplifiers, and nothing low end.

Anyway, I will stuck with the post and attention for any update.

As told I am from Hong Kong, do you having any preparation for international payment such as via PayPal?

Yesterday I didn't want to suggest to replace caps and everything, but I felt you were heading that way. So here it is :

The main PSU board (the one creating the sound) explicitly consists of a configuration where you can mount 10 smaller caps per channel, 2 big ones, or any combination (but the big ones take space for the smaller ones, so 10 small + 2 big is not possible). It comes with 1x 40000uF audio grade low ESR per channel.
There are 33 other (electrolytical) caps to replace, devided over 4 functional areas. None is in any signal path, and all matter for sound. They should be all the best already and take 3-4 days to break in (so no black gates), although I'm fairly sure that after a month you will still see improvement which you will find significant at going back to new ones after 3-4 days breaking in (I just did that for fun a couple of days back, and while the one month old made me say "this is just live music", I can't say that (yet) from the new ones).
Resistors are Vishay 0,1% non-magnetic throughout (for the non-SMD stuff).

PayPal is no problem !
Peter

[PeterSt]

Generate it somewhere outside of the PC.
But did that in a fashion that no clock (software or hardware) in the PC plays any role.

Would you like to share the recipe with us? 

Later.

[pedal]

Later.

This thread has become a DICK DAC teaser...