The tired of the obfuscation on Digital Audio article.

[Orb]

Yeah I agree with a staircase statement rather "the staircase", which I now notice I typed doh
Cheers
Orb

 

[DonH50]

DAC articles in the Technical Forum on WBF show the stairsteps, as well as discuss the implications, and also I showed the NTF for DS DACs.

http://www.whatsbestforum.com/showthread.php?2829-Don-s-Tech-Series

Sorry for the semi-advert interruption... - Don

 

[xiphmont]

It's worth noting that although stairsteps of some variety or another occur as an implementation detail in DACs, the exact analag doesn't occur in ADCs; there is a track and there is a hold, but things are already conceptually discrete time at the point of the hold (the hold feezes time at the point of the sample).

You also can't have an antialiasing-less ADC the way the AudioNote DAC gets away without a reconstruction filter. Unless the input to the ADC is bandlimited, all Hell effecively breaks loose and there's no recovering. It's no longer a matter of being a questionable engineering decision, it simply doesn't work as you can't undo the aliasing distortion later after samping.

 

[LL21]

You also can't have an antialiasing-less ADC the way the AudioNote DAC gets away without a reconstruction filter.

I am such a non-techie, i can barely follow this thread, but remain intrigued by those here who clearly know things about digital i do not. What is your opinion of the AudioNote DAC...in particular the 5th Element which they recently released? On the one hand, the price seems silly. On the other hand, time and again, i do hear from the few that have heard it...it is in another league above other digital they've heard. Do you have any views from the technical perspective on the merits of how AudioNote designs their DACs? Thanks for any guidance.

 

[xiphmont]

What is your opinion of the AudioNote DAC...in particular the 5th Element which they recently released?

My opinion is technical. I've not heard it, though nothing in the reviewed measurements suggests any audible sonic flaws. Not in the DAC itself anyway...

NOS DACs, more importantly DACs with no reconstruction/anti-imaging filter, are going to present a serious challenge to anything downstream. If nothing else is performing the anti-imaging lowpass either, the vast majority of the signal power will be above the audible band extending into RF. My own previous testing shows that injecting high power RF into a power amplifier can produce unexpected and unpleasant results. You'll be getting the full brunt of the amplifier's IMD, RF rectification [depending on type/topology], and effects of substantial signal power beyond the GBP. ...not to mention what it will do to tweeters not rated for the amplifier's full power output.

So, my opinion is: Theoretically it's fine; we can't hear the ultrasonics. Practically, downstream components aren't designed to deal with this much ultrasonic/RF and they'll be in danger of mild to serious misbehavior... unless, I suppose, you buy them from AudioNote too. I'm going to go out on a limb here* and call this a snake oil fad, done purely for false differentiation, and an actively bad idea.

OTOH, anyone who's listened to a NOS DAC and thinks it sounds good no longer gets to complain that '44.1kHz isn't enough'; if it wasn't, AudioNote buyers would be flinging themselves out windows to escape the pain. Obviously, they're not. In fact, the acceptance of the AudioNote is a singlehanded debunking of several earlier snake-oil claims. Could this simply be a masterful troll? If so, it is _brilliant_.

 

[opus111]

On the other hand, time and again, i do hear from the few that have heard it...it is in another league above other digital they've heard.

I concur with most of what Monty's saying about RF energy - its abuse to amps and contributes IMD which IME is audible as loss of dynamics and greyness of tonality. However the fact that it sounds so good given all the abuse just points to the rubbishiness of the other kinds of non-NOS DACs. The common-or-garden kind of S-D DAC is introducing more audible crud than amps do when subjected to RF abuse - go figure.

 

[xiphmont]

I concur with most of what Monty's saying about RF energy - its abuse to amps and contributes IMD which IME is audible as loss of dynamics and greyness of tonality. However the fact that it sounds so good given all the abuse just points to the rubbishiness of the other kinds of non-NOS DACs. The common-or-garden kind of S-D DAC is introducing more audible crud than amps do when subjected to RF abuse - go figure.

Have we seen a review of an AudioNote NOS DAC on something other than an AudioNote amp? Especially, oh let's say... something with a JFET input stage? ;-)

 

[Gregadd]

approximately constant

 

[bbb]

Have we seen a review of an AudioNote NOS DAC on something other than an AudioNote amp? Especially, oh let's say... something with a JFET input stage? ;-)

Don't really need to concentrate on AudioNote DAC5 as most 16 bit NOS dac (TDA15**, AD1865) don't have a reconstruction filter. All of AudioNote's current dac lineup are filterless. IIRC, even the Metrum Octave and totaldac (can select FIR filter though) are run filterless and there are plenty of reviews for the Metrum. Or you can look up Lab47's dacs which was 1st designed by the so called father of NOS, Ryohei Kusunoki. There are plenty written about Kusunoki and his theory on NOS, which will likely be dismissed by the objective type. Also, you can pick up some very cheap NOS dac on ebay ala Opus if you really want to have a go at it.

<edit> totaldac's designer Vincent is a member of this forum so maybe you can ask him.

 

[bbb]

Here is some of Martin Colloms' thoughts on Digital and NOS for anyone interested http://www.hificritic.com/Forum/yaf...HEX-balanced-has-arrived-at-HIFICRITIC.aspx?=. BTW, he's the one proclaiming AN 5th element DAC to be the best of the best.

 

[xiphmont]

Don't really need to concentrate on AudioNote DAC5 as most 16 bit NOS dac (TDA15**, AD1865) don't have a reconstruction filter.

Sure. I was discussing it as an example where we have detailed third party measurements already mentioned in the thread. and I agree that theoretically it's not doing anything that should damage its audible quality. The practical Concerns (warrants a capital C) would be the same for any NOS DAC without an anti-imaging filter.

There are plenty written about Kusunoki and his theory on NOS, which will likely be dismissed by the objective type. Also, you can pick up some very cheap NOS dac on ebay ala Opus if you really want to have a go at it.

We 'objectivists' [hate that label, but it fits] generally dismiss things for transparent, documented reasons. The world is full of facts aren't merely 'a preference'.
In any case, I'm busy wasting a ton of money on a rather more fun [to me anyway] audio project that is even more pointless, so I shall decline ;-)

totaldac's designer Vincent is a member of this forum so maybe you can ask him.

I think I'll at least ping him about this thread for his commentary.

 

[opus111]

Here is some of Martin Colloms' thoughts on Digital and NOS for anyone interested http://www.hificritic.com/Forum/yaf...HEX-balanced-has-arrived-at-HIFICRITIC.aspx?=.

He seems to be down on digital filtering for some reason there. I'm betting that the 'piano hardness' he hears is correlated with some aspects which are changing when he puts a digital filter in the chain (like with an oversampling filter - running the DAC faster therefore more glitching) and not caused by the filtering itself. I'm with 'RK' - the main reason NOS sounds better is elsewhere from (lack of) digital filtering.

 

[xiphmont]

He seems to be down on digital filtering for some reason there. I'm betting that the 'piano hardness' he hears is correlated with some aspects which are changing when he puts a digital filter in the chain (like with an oversampling filter - running the DAC faster therefore more glitching) and not caused by the filtering itself. I'm with 'RK' - the main reason NOS sounds better is elsewhere from (lack of) digital filtering.

And I'm not yet convinced it does sound better, but I'm also not willing to go down that testing path right now, so I'm out.

 

[Phelonious Ponk]

And I'm not yet convinced it does sound better, but I'm also not willing to go down that testing path right now, so I'm out.

I've had a couple of NOS DACs in on loan, in my system. They sound different. Let's just say the lack of Collumns' "hardness" of piano is somebody else's "softness." They sound more "analog," I'll give you that.

Tim

 

[Orb]

My opinion is technical. I've not heard it, though nothing in the reviewed measurements suggests any audible sonic flaws. Not in the DAC itself anyway...

NOS DACs, more importantly DACs with no reconstruction/anti-imaging filter, are going to present a serious challenge to anything downstream. If nothing else is performing the anti-imaging lowpass either, the vast majority of the signal power will be above the audible band extending into RF. My own previous testing shows that injecting high power RF into a power amplifier can produce unexpected and unpleasant results. You'll be getting the full brunt of the amplifier's IMD, RF rectification [depending on type/topology], and effects of substantial signal power beyond the GBP. ...not to mention what it will do to tweeters not rated for the amplifier's full power output.

So, my opinion is: Theoretically it's fine; we can't hear the ultrasonics. Practically, downstream components aren't designed to deal with this much ultrasonic/RF and they'll be in danger of mild to serious misbehavior... unless, I suppose, you buy them from AudioNote too. I'm going to go out on a limb here* and call this a snake oil fad, done purely for false differentiation, and an actively bad idea.

OTOH, anyone who's listened to a NOS DAC and thinks it sounds good no longer gets to complain that '44.1kHz isn't enough'; if it wasn't, AudioNote buyers would be flinging themselves out windows to escape the pain. Obviously, they're not. In fact, the acceptance of the AudioNote is a singlehanded debunking of several earlier snake-oil claims. Could this simply be a masterful troll? If so, it is _brilliant_.

I have been away working on a project so 1st time I can join back in the conversation, finally get to rest woot
One consideration though is the sinc roll-off which nearly all NOS DACs do not implement correction for Sin(X)/X; so higher frequencies well have much less attenuation, which is also helped from my understanding that modern good wideband pre/amps have a bandwidth to around 300khz (and consideration for IM filter/RF contamination).
However agree that technically it is not desirable, but may explain why NOS DACs have not caused problems for reviewers (and the many pre/power/integrated they have been used with) and no mention of practical issues/failure from forum members on various sites.

Cheers
Orb

 

[Orb]

.....Snip
I played with the idea of making a little javascript dither playground tool for the website, but it would eb kind of boring. The vid did show a low vs high amplitude sinewave without dither. With dither it's unremarkable... the noise floor just stays static.

Xiph, yeah the only way to make it interesting is by my meaning/context of low level signal; meaning a 1khz signal at -90dB level for 16bit.
In practical terms this makes it very clear the benefit of dither as the sinewave is very jagged/stepped without dither, furthermore combined with this should be spectrum plot that shows distortion effects as well (and this is very important and missed by many who argue on various forums about PCM and dither).
Good example of this is some links I have provided in the past:
Jim Lesurf spectrum distortion and waveform -90dB no dither and dithered shows it best: http://www.audiomisc.co.uk/inadither/Page2.html

John Atkinson -90dB sinewave undithered and dithered for some DACs:
It becomes interesting for two reasons; a) how undithered sinewave is handled by product and b) separately shows as I mentioned quite clearly difference between undithered and dithered (not NOS related audio measurement-reviews).
Fig8 undithered 1kHz waveform of NOS Audionote: http://www.stereophile.com/images/712an.AN41fig08.jpg
Fig5 undithered 16bit 1kHz waveform (near perfect representation of undithered at 16bit) of Simaudio Evolution SD DAC: http://www.stereophile.com/images/1111Simfig05.jpg
Fig6 undithered 24bit 1kHz waveform (shows that even at 24bit needs to be dithered) of Simaudio Evolution SD DAC: http://www.stereophile.com/images/1111Simfig06.jpg

Interestingly the MSB multibit DAC has even better performance for undithered 1kHz 24bit waveform: http://www.stereophile.com/images/1012MSBfig09.jpg

So I think it can be added successfully to your site showing visually how and why dithered needs to be considered and also why a staircase in this context is applicable to only this stage in the complete cycle of Analogue-Digital-Analogue (important emphasis due to other types of a staircase related function-behaviour), and critically that dither is added to all music at the studio (so is a stage-function within the complete cycle).

Cheers
Orb

 

[LL21]

I've had a couple of NOS DACs in on loan, in my system. They sound different. Let's just say the lack of Collumns' "hardness" of piano is somebody else's "softness." They sound more "analog," I'll give you that.

Tim

If you ever heard the Stahl-Tek Vekian DAC, let me know what you think. I am going to take a guess (wild a-s guess) you might like it given your comments about digital and analog. while i prefer my Zanden, i was impressed with Stahl-Tek.