[405line]

Okey dokey then.... http://www.computeraudiodesign.com/bristol-sound-vision-best-sound-of-show-award-winner/
Distortion is cool if you adjust to for DC correct offset.

 

[00940]

Well, it might sound good. But it is for sure accompanied by marketing BS.
 
An example: Computer Audio Design claims to use R2R DAC chips but then when you read more closely, you find out that they use the TD1543. Which is not a R2R DAC but a rather poor multibit chip (it was designed by Philips as an economy part, as explained in its datasheet). By claiming the R2R part, they try to capitalize on the reputation of true high end R2R DAC, such as the PCM1704 or AD1865.
 
I also wonder how they manage to have, I quote, "less than 1 cm from the USB board to the DAC chips" while using 16 DAC chip in DIP8. Well, sorry, but that seems pretty impossible in terms of size.
 
The audio optimization guide is also a paradox in itself. This guy design a DAC with asynchronous USB input and galvanic isolation (making it independant of the quality of the USB stream coming from the computer) and then goes on to suggest to disable everything on your computer to improve the quality of the USB bus. Paranoia comes to mind.
 
I also cannot help but notice that all the reviews are by the same reviewer...
 
This being said: the TDA1543 has a long history in the diy community. It's very cheap, it's easy to use, it has a strong output (making it easy to use passive I/V), its distortions, while very high by today's standards, are pleasant. Paralleling a boatload will reduce noise and distortions. No doubt it is nice to listen to.

 

[leeperry]

lol a bunch of TDA chips for 7K£, I guess the super duper power cable is the real game changer here.

 

[Baxide]

Which DAC chip would you suggest that he had used that would have satisfied you in terms of the asking price?
Too many times people focus on the DAC chip used in order to validate the price of the finished product. But you are paying for far more than just a chip. There are also screws, resistors, packaging, support, future R&D, etc. etc.
 

 

[00940]

Of course you pay more than the chips... obviously even more so at 11000$. But here the ratio is just ridiculous.
 
To put things in perspective, it's the same price as the DCS Debussy DAC.

 

[Currawong]

I used to own the Metrum Octave, which is a bit different from a typical NOS DAC. No issues with ultra-high frequencies as I measured it to have steep roll-off up to Nyquist. Feeding it high-res files (which effected the same thing as over- or up-sampling) improved the sound a little (and also changed the cut-off filter from what I tested).

 

[leeperry]

I believe they keep the model numbers of the DAC chips used in the Metrum Acoustics units secret so it's impossible to read any datasheet and check whether there is actually some oversampling applied internally.......quite frankly not oversampling is not technically possible AFAIK. Even the TDA chips oversample(as per their datasheet proves) as this is mandatory for a D/A conversion, I believe Mr Lavry confirmed it clearly when I raised the subject in the OP. Not oversampling whatsoever is really caveman technology and would throw THD figures up through the roof.

 

[405line]

With regards to the Nyquist issues has anyone tried interpolating aka upsampling in the PC i.e feeding NOS with 88.2Khz or 176.4Khz, in theory this should push the images and any aliasing to twice (or more) the "normal" frequencies. I have had no problems using the SOX resampler in foobar for this purpose. My dac-ah uses an fairly ancient CS414 receiver so 88.2Khz is as far as I can go..i feel it may well be enough as it will effectively give me 44Khz playback bandwidth before any native NOS artifacts become apparent if indeed they actually are of any significance in the first place, although I am certain the treble is a shade more extended when A-B'ing.
 
This post could be called "how to double or quadruple the frequency response and of a filterless NOS dac and shift aliasing to 66-68 Khz"
I hope this explains better than I can.
 
free".
 
 

 
A picture paints a thousand words.

 

[zilch0md]

OK, I just finished reading this short, 23-page thread, which is certainly one of my favorites, here on Head-Fi - I enjoyed reading it, immensely.
 
Dan Lavry checked out on page 13, but I have to hand it to him for how well he explained "the facts" across many posts - taking the time to teach us a LOT about how DACs work, while doing a really good job of controlling his frustration.  His passion and conviction are admirable, to be sure.  I believe him when he says NOS DACs don't measure well, and I even trust his explanations for why they fail to measure well. In almost the same breath, I can nevertheless say that I'm really looking forward to the Metrum Octave MkII NOS DAC that's due to arrive, tomorrow.  

 
 
I'm in the camp that says, if it sounds good to my ears, I don't care how it measures -and- my ears like some warmth and coloration, I confess. Give me my Audeze LCD-2 rev.1 fed by a neutral, transparent signal or my HD800 fed by a warm and ever so slightly fuzzy signal - that's what I like - and I really don't care if I get the "imperfections" I seek from the recordings that find their way to my favorites playlist, from my DAC, my amp, my headphones, or from some combination of these actors. I might have a champagne wallet and beer taste, but I know what I like.  No measurements necessary.
 
From all the great subjective reviews I've read, I'm really looking forward to trying the Octave MkII in the following chain. I'll soon find out if this DAC can take the HD800 where I want it to go, despite the neutral and transparent personality of both the amp and headphone:
 
FiiO X5 coaxial out > Metrum Octave MkII > OPPO HA-1 balanced out > unmodified HD800
 
Mike

 

[lamode]

I feel sorry for Dan. I don't think anyone really understood or even read his post. If you can understand how a Digital to Analog Converter works, you'd understand how absurd the idea of a truly non oversampling DAC is.

 
+1
 
A properly implemented oversampling DAC will just reduce noise and push it further from the audio bandwidth. There are only benefits to doing this.

 

[Yuri Korzunov]

I wrote educational articles about DACs with easy (as I hope) explanation:

1. R2R Ladder DAC vs Sigma-Delta DAC vs DSD DAC [with infographic. It is recommended to read first, before second article]
   https://samplerateconverter.com/educational/r2r-ladder-dac-vs-sigma-delta
   
   
2. NOS DAC. Non-Oversampling DAC [advantages, disadvantages, sound quality issues]
   https://samplerateconverter.com/educational/nos-dac

If there are questions, write here.

 

[studer]

Dan Lavry from Lavry Engineering recently wrote a post in regards to Non-Oversampling DAC.

To sum it up, Dan said, "So the NOS DA is either a piece of very inferior gear, or it is an up sampling DA that some sales guy calls NOS while it is not."



I'm very interested in this topic regarding NOS DAC, any comment guys?

An interesting analogy would be Photoshop work where one would "upsample a standard image to a higher resolution and then comparing the two. Try and see what you prefer (many variables here)
Do you see (hear) the jagged edges? From what distance?
This may be a simplistic analogy.
The theory behind NOS vs OS DACs is complex. When testing and comparing, a multitude of factors come into play;

That is why I suggest the statement above ("So the NOS DA is either a piece of very inferior gear, or it is an up sampling DA that some sales guy calls NOS while it is not.") seems quite abrupt..

 

[Yuri Korzunov]

An interesting analogy would be Photoshop work where one would "upsample a standard image to a higher resolution and then comparing the two. Try and see what you prefer (many variables here)
Do you see (hear) the jagged edges? From what distance?

In my opinion, optical analogy to sound is spectral (colour) one rather. Because optical phenomenon is wave like sound (https://samplerateconverter.com/content/ultrasound-ultraviolet-audio-optics).

Oversampling, high resolution (high sample rate) aim is solving of issue of non-steep analog filter: intermodulations, dynamic range reducing.

Is there geometrical analogy with analog filter in this example?

 

[studer]

In my opinion, optical analogy to sound is spectral (colour) one rather. Because optical phenomenon is wave like sound (https://samplerateconverter.com/content/ultrasound-ultraviolet-audio-optics).

Oversampling, high resolution (high sample rate) aim is solving of issue of non-steep analog filter: intermodulations, dynamic range reducing.

Is there geometrical analogy with analog filter in this example?

Interesting link and very informative; however, my analogy aimed to other aspects of the "sound" besides frequency response.
A sinus-shaped signal with (44.1 kHz) that looks (measures visually) like a jagged line may not translate as such in our hearing syatem.
Passing through the amplification line and then the headphones(speakers) and then being received by our hearing system leads to "filtering".
The better by architecture transient response of NOS is a major advantage, translating into perceived "realism" (arguably).
As one who listenes to analog as well as digital sources, I experienced the advantages and limitations of both designs (having the analog sound as a comparison starting point).

 

[Yuri Korzunov]

Passing through the amplification line and then the headphones(speakers) and then being received by our hearing system leads to "filtering".
The better by architecture transient response of NOS is a major advantage, translating into perceived "realism" (arguably).

NOS have more probability of audible intermodulations than oversampled/digitally filtered DAC.

Currently I don't viewed yet researches of the ringing audio impact to human perception. Also a musical signal has few sharp transients between neighbour samples in many cases.