[3X0]

Unless your DAC is R-2R ladder DAC type it would then be Delta-Sigma which by very nature cannot do None oversampling to start with, regardless what effect or filter it says it has, sorry.

Yup, you're right in that regard -- the defeatable NOS switches as used in the PWD and some Peachtree Audio products aren't truly NOS due to the nature of the chip. But conversely, the article you posted earlier is misleading as modern R2R DACs still have a reconstruction filter external of the chip (delta-sigma DACs have this built-in). Since reconstruction filters of NOS DACs is not standardized by any means, the engineer is at liberty to do all kinds of questionable reconstruction filtering to the signal at his/her whim. Some (many?) NOS DACs might be regarded as sound processors rather than bona fide DACs. Exaggerated smoothing in the high-frequency domain with a generous low-pass filter may not be so uncommon.
 
And even when there is no explicit digital/analog filter for reconstruction, the NOS DAC/processor will accomplish this by some means of implementation -- this makes the effect on the signal relatively unpredictable.
 
The artifacts of NOS DACs are measurable and repeatable even from (in)famous publications like Stereophile. The classic NOS DAC trades in errors of the timing domain (i.e. jitter) for a higher noise floor and some distortion products. I have found more evidence to suggest sensitivity to such artifacts than evidence supporting sensitivity to nominal levels of jitter.

 

[mechgamer123]

I had the chance to hear two pairs of original SR-Lambdas tonight...
Set 1: In good working condition driven connected to a really expensive setup (I forgot the exact model of the DAC, but the speaker amp was a  Final Audio Music 6 preamp/speaker amp powered by batteries) through some speaker transformer (not the SRD-7) and I didn't like them as much. The definitely had very accentuated upper treble that made the soundstage feel bigger, but also made anything other than prefect recordings sound too bright and harsh.
The second pair are in far worse shape, the foam is falling apart around the driver, the pads are pretty much toast (and on backwards to boot) and the connectors look like they've been jimmy rigged to keep a connection - they'll randomly quit making sound when I move my head, and they're driven by the SRD-7, which the owner of both sets says sounds worse than the other amp. But for some reason this set isn't anywhere near as bright as the other set, its tonality is much better, although the midrange may be more recessed, I'd need more listening to figure that part out. They do sound cleaner and more refined than my SR-202 though, which I find surprising given their condition...

 

[complin]

Actually many of the top R2R ladder DAC's don't use a conventional DAC chip at all but are built with discrete components i.e. the TOTALDAC from France
 
I believe Headfier Hun7er has or had one of these?

 

[astrostar59]

Yup, you're right in that regard -- the defeatable NOS switches as used in the PWD and some Peachtree Audio products aren't truly NOS due to the nature of the chip. But conversely, the article you posted earlier is misleading as modern R2R DACs still have a reconstruction filter external of the chip (delta-sigma DACs have this built-in). Since reconstruction filters of NOS DACs is not standardized by any means, the engineer is at liberty to do all kinds of questionable reconstruction filtering to the signal at his/her whim. Some (many?) NOS DACs might be regarded as sound processors rather than bona fide DACs. Exaggerated smoothing in the high-frequency domain with a generous low-pass filter may not be so uncommon.

And even when there is no explicit digital/analog filter for reconstruction, the NOS DAC/processor will accomplish this by some means of implementation -- this makes the effect on the signal relatively unpredictable.

The artifacts of NOS DACs are measurable and repeatable even from (in)famous publications like Stereophile. The classic NOS DAC trades in errors of the timing domain (i.e. jitter) for a higher noise floor and some distortion products. I have found more evidence to suggest sensitivity to such artifacts than evidence supporting sensitivity to nominal levels of jitter.

I find this all very odd. Measured noise floor on the Metrum Hex is 113DB, how much do you want?

As for filter after the chip, the conversion from digital to analogue is bit for bit conversion. The key difference with Delta-Sigma is they are upsampling then converting. To my ears
this is where it gets smeared and the treble takes on a very obvious hifi or digital quality. Try a good NOS DAC with RedBook (with no brick wall filter) in your system, then come back and restart this discussion. Or just stick to your magazines instead.

 

[milosz]

Did some ask for DAC recommendations for their Stax?
 
There are some low cost DACs out there that you might like.  You can buy a used one and try it, maybe buy another lower-cost used DAC and compare them, etc.  to get a feel for what - if any- differences you find between them.  Maybe later, buy a fancy expensive one.
 
I suggest trying the original Musical Fidelity V-DAC.  http://www.stereophile.com/content/musical-fidelity-v-dac-da-processor-measurements  I have one and to me it sounds a trifle "sweet" or kind of "tube-like."  This may be a projection or aural imagination on my part, I have not done any A/B comparisons with this DAC.  But it measures pretty well and quite a few people have enjoyed using it.  I like mine.
 
Cambridge DAC MAGIC is also another good budget DAC you can get used for not much money.  And there are any number of good DAC / DAC-Dynamic Headphone Amp / DAC-Preamp   units from Audio-Gd that are very well made and offer good sound with excellent value. Audio-Gd  DACs come in two sigma-delta flavors:  ES9018  and WM8741, and one R2R flavor - PCM1704UK.  Sigma-delta and R2R are two kinds of DAC technologies, or digital-to-analog conversion technologies, with ES9018, WM8741 and PCM1704UK being the specific DAC chips used.  Each have their strengths.
 
HOWEVER  if you are listening to garden-variety MP3 files as your music source, all the finesse that a better DAC can offer is going to be wasted.  MP3 files -  or low bitrate AAC or WMA files-  are *NOT* high-fidelity audio. An MP3 with a 240 KHz or higher bitrate can sound OK depending on the music, but bitrates lower than 240 KHz tend to be kind of crappy; not bad for background music but not worth spending $hundreds on electrostatic 'phones to listen to.
 
You are much better off with LOSSLESS FILES - an actual CD, or FLAC, ALAC, OR WMA LOSSLESS. Some people even go so far as to say that HIGHER RESOLUTION files- like 24-bit/96 KHz files as can be purchased from HDTRACKS.COM or PONO MUSIC are SO MUCH BETTER THAN CD that they hardly bother listening to lower-res music any more. (I've done some A/B/X tests with high res files using myself and also some pro musicians as test subjects and we couldn't tell the difference, so personally I am not convinced that going beyond Red Book standards really makes an audible difference, but your mileage may vary)

 

[Lonely_Rider]

When I was choosing a new DAC about four years ago, one I listened was Red Wine Audio Isabellina NOS DAC. Well, I didn't liked it at all: sound was murky, and there was grainy distortion in high frequencies. My choise was Electrocompaniet ECD-1. Maybe NOS DAC's are better today, than four years ago, I don't know.
 
Those, who fear "digital sound" (like we all), I recommend to listen Lynx Hilo. I was sceptical about DAC's made for professional use, but not anymore.

 

[James-uk]

Did some ask for DAC recommendations for their Stax?

There are some low cost DACs out there that you might like.  You can buy a used one and try it, maybe buy another lower-cost used DAC and compare them, etc.  to get a feel for what - if any- differences you find between them.  Maybe later, buy a fancy expensive one.

I suggest trying the original Musical Fidelity V-DAC.  http://www.stereophile.com/content/musical-fidelity-v-dac-da-processor-measurements  I have one and to me it sounds a trifle "sweet" or kind of "tube-like."  This may be a projection or aural imagination on my part, I have not done any A/B comparisons with this DAC.  But it measures pretty well and quite a few people have enjoyed using it.  I like mine.

Cambridge DAC MAGIC is also another good budget DAC you can get used for not much money.  And there are any number of good DAC / DAC-Dynamic Headphone Amp / DAC-Preamp   units from Audio-Gd that are very well made and offer good sound with excellent value. Audio-Gd  DACs come in two sigma-delta flavors:  ES9018  and WM8741, and one R2R flavor - PCM1704UK.  Sigma-delta and R2R are two kinds of DAC technologies, or digital-to-analog conversion technologies, with ES9018, WM8741 and PCM1704UK being the specific DAC chips used.  Each have their strengths.

HOWEVER  if you are listening to garden-variety MP3 files as your music source, all the finesse that a better DAC can offer is going to be wasted.  MP3 files -  or low bitrate AAC or WMA files-  are *NOT* high-fidelity audio. An MP3 with a 240 KHz or higher bitrate can sound OK depending on the music, but bitrates lower than 240 KHz tend to be kind of crappy; not bad for background music but not worth spending $hundreds on electrostatic 'phones to listen to.

You are much better off with LOSSLESS FILES - an actual CD, or FLAC, ALAC, OR WMA LOSSLESS. Some people even go so far as to say that HIGHER RESOLUTION files- like 24-bit/96 KHz files as can be purchased fromHDTRACKS.COM or PONO MUSIC are SO MUCH BETTER THAN CD that they hardly bother listening to lower-res music any more. (I've done some A/B/X tests with high res files using myself and also some pro musicians as test subjects and we couldn't tell the difference, so personally I am not convinced that going beyond Red Book standards really makes an audible difference, but your mileage may vary)

Agree mostly but 256 AAC and 320 orbis are transparent on 99.9% of recordings and nobody yet has proven otherwise. I used to laugh when people steamed and used apple downloads then I decided to educate myself and test it for myself. As to the dac recommendations you can't go wrong with an ODAC if USB is all you need. Also just wanted to say I'm currently in a position where I have reached end game with dynamics so I'm going to buy some stats in the near future . Great thread, I'm learning alot , thank you all.

 

[3X0]

Measured noise floor on the Metrum Hex is 113DB, how much do you want?

As for filter after the chip, the conversion from digital to analogue is bit for bit conversion. The key difference with Delta-Sigma is they are upsampling then converting. To my ears
this is where it gets smeared and the treble takes on a very obvious hifi or digital quality.

Actually the effective noise floor of the Hex is somewhere between -100 and -90 dB but this is still mostly good enough for Redbook and our ears. I suspect you quoted the effective SnR of the core DAC in the circuit, while the fully implemented device as a whole behaves differently.

The quantization errors produced by oversampling DACs is very low in level and pushed well above the Nyquist frequency. The analog filtrations designed into some NOS DACs perform necessarily worse as it doesn't have as much room to work with, introducing the aforementioned artifacts into the audible exist.

I suspect your upstream decisions might be influenced by the quality of the recordings. Especially when mixed for speaker playback, mastering can wind up a bit hot in the treble.

@complin I am even more skeptical of such fully discrete DACs unless the manufacturer has the resources to properly control effective performance. Lack of standardisation and risk for "creative" (i. e. fidelity-destroying) decisions shouldn't be something that one voluntarily pays a premium for. I have concerns against such DACs being considered "top" if their performance is not properly validated. Topology adulation shouldn't be reasonable justification.

 

[negura]

Astrostar59/Negura, could you take a couple of pix of your Stax next to your KGSSHv so that I can get a sense of scale, please?

I'm particularly interested in how compact Geoff has made his builds including his newer square block design.

 

 

[astrostar59]

When I was choosing a new DAC about four years ago, one I listened was Red Wine Audio Isabellina NOS DAC. Well, I didn't liked it at all: sound was murky, and there was grainy distortion in high frequencies. My choise was Electrocompaniet ECD-1. Maybe NOS DAC's are better today, than four years ago, I don't know.

Those, who fear "digital sound" (like we all), I recommend to listen Lynx Hilo. I was sceptical about DAC's made for professional use, but not anymore.

No wonder, that Red Wine DAC looks to have one of the most basic power supplies and output amp I have ever seen. That is NOT where most NOS DACs are at.

Does this look better as regards implementation?

 

[realmassy]

To be honest one of the biggest difference between my two DACs (Metrum Hex and Ayre QB-9 DSD) is the background, which is TOTALLY black on the Ayre and 'grey' (relatively speaking) on the Metrum. They both have great qualities, but with the QB-9 instruments really pop out of the black. This helps a lot with instrument separation which are more solid and easy to pinpoint. The Hex has different abilities, but to my ears is no better than the QB-9.

 

[Michgelsen]

This feeling that instruments pop out of the black I have found to be caused by good dynamics, so this can mean that the Ayre has better dynamics than the Hex. This in turn MIGHT be due to a better power supply.

 

[astrostar59]

Actually the effective noise floor of the Hex is somewhere between -100 and -90 dB but this is still mostly good enough for Redbook and our ears. I suspect you quoted the effective SnR of the core DAC in the circuit, while the fully implemented device as a whole behaves differently.

The quantization errors produced by oversampling DACs is very low in level and pushed well above the Nyquist frequency. The analog filtrations designed into some NOS DACs perform necessarily worse as it doesn't have as much room to work with, introducing the aforementioned artifacts into the audible exist.

I suspect your upstream decisions might be influenced by the quality of the recordings. Especially when mixed for speaker playback, mastering can wind up a bit hot in the treble.

@complin I am even more skeptical of such fully discrete DACs unless the manufacturer has the resources to properly control effective performance. Lack of standardisation and risk for "creative" (i. e. fidelity-destroying) decisions shouldn't be something that one voluntarily pays a premium for. I have concerns against such DACs being considered "top" if their performance is not properly validated. Topology adulation shouldn't be reasonable justification.

I have 2 DACs, the Metrum Acoustics and the Audio Note 4.1, and BOTH have no filter, it has been removed from the chip. I have heard the 4.1 with and without the filter, and I can assure you, it sounds a lot better without it. I have horn speakers and Stax 009s, and no, nothing has blown up as some say it will.

The filter kills the decay and ambience in the music, and the oversampling smears the sound and creates a digital treble.

My taste may not be like everyone, I want organic music, as though my DAC could be a good turntable rig (as I owned before I went digital). At this level with the 009s or a top flight speaker setup, the treble quality is everything.

 

[astrostar59]

Hey Negura, that looks real nice. Your KGSShv compact looks a lot smaller
than the Perfect Wave DAC. I will post some photos next week of mine. We decided for speed of build reasons
to do the heatsinks in Silver. It looks great in either black or silver IMO.

I am using mine with a Metrum Octave II so RCA/XLR convertor plugs. Thus the volume is about 12.oclock most
of the time. With XLR is is about 9 oclock.

It sounded great from first use as Geoff had run it for 60 hours (no music). It did lift a bit after a few days, more transparent
and even smoother treble. The bass power with the 009s in incredible, I was shocked. And the extra detail (or brightness as some
call it with the 009s is not an issue with the KGSShv.

I think Geoff does a great build. I also believe not all KGSShv's sound the same, it is the same design as either an off-board
or on-board type, but the quality of the transformer (and the over spec size of it) plus the resistor types and the capacitor quality
all affect the final SQ IMO. Audio Note have been doing the same amp designs for years, but put better and better parts in that design
and call it levels. I have heard my 300B Monoblocks at level 3 (my amps) and at level 5, and it is not subtle.

The casework is also nice. One thing Geoff got right IMO is also the mounting of the 8 power regulators (think that's what they are).
They are glued and screwed onto a solid 25 mm square section bar that is bolted to the inside of the heat sinks. It gets rid of all that heat.

 

[paradoxper]

Joining the HV party.