Teac UD 501, "NOS mode"; measures worse, sounds better?

[Orpheus]

I bought the Teac UD 501 about a year ago, for computer audio. It's a great and flexible DSD DAC, and very sweet sounding. It also offers multiple inputs, including asynchronous USB.

 

It also has a lot of flexibility with what it can do with the signal. It can upconvert to 24/192 with any PCM signal, and you can also chose between two filters, or no filter.

 

I've been listening to a great, lively, but occasionally harsh redbook CD of Jodi Mitchell's "Hits", and fiddling around with filters and up sampling.

 

Basically, I like the PCM upsampled to 24/192, and the filter turned off.

 

Having reached this conclusion, I set about looking for theoretical support for the views of my ears.

 

Rather discouraging. The performer of the tests in the link I've included seems to find that by far the worst measurements are achieved with the filter off, though (I think) the upsampling helps.

 

On the other hand, I think he is saying that this is effectively a NOS mode, and that the measurements are very similar to some of the older chips, such as TDA1543.

 

http://archimago.blogspot.ca/2013/05/measurements-teac-ud-501-pcm-performance.html

 

Interested to hear some views about this.

[Orpheus]

Is that fellow right when he says the TEAC offers NOS mode when the PCM filter is removed?

[Guest myrantz]

Is that fellow right when he says the TEAC offers NOS mode when the PCM filter is removed?

He's a member here.. Hopefully he'd chime in.. So far you like the DAC?

[Orpheus]

He's a member here.. Hopefully he'd chime in.. So far you like the DAC?

Yes, I think it's a great DAC.

 

Very pure and musical.

[acg]

If you are interested in the nos aspect of your dac, then I recommend that you try XXHighEnd software and apply a state of the art digital filter and upsampling. I think you may be surprised by further gains in sq.

Anthony

Edited April 17, 2014 by acg

[Orpheus]

If you are interested in the nos aspect of your dac, then I recommend that you try XXHighEnd software and apply a state of the art digital filter and upsampling. I think you may be surprised by further gains in sq.

Anthony

Well I suspect that the TEAC comes with a state of the art digital filter and state of the art upsampling. 

 

As I said, I find the upsampling beneficial, but I am not sure that the filtering is beneficial, even though it measures better with it, in most respects. Interestingly, intermodular distortion is lower without the filtering. However that manifests itself!

[acg]

I sincerely doubt it uses a zero ringing digital filter because as far as I know xxhe has the only one in existence. It is possible that some of what you like about the nos sound is caused by the reduced load on the dac chip...which can be eliminated by xxhe whilst maintaining the native sample rate of the dac chips and a true sota digital filter...best of both worlds.

Edited April 17, 2014 by acg

[Orpheus]

I sincerely doubt it uses a zero ringing digital filter because as far as I know xxhe has the only one in existence. It is possible that some of what you like about the nos sound is caused by the reduced load on the dac chip...which can be eliminated by xxhe whilst maintaining the native sample rate of the dac chips and a true sota digital filter...best of both worlds.

Where do I find these things? I suppose I should do a search of the digital sub-forum.

[hochopeper]

I did have a read of the blog last night but it's long and lots of measurments so I didn't have time to fully digest it all sufficiently to answer the question in your OP. I can't get to the blog because it's blocked by my work ... I'll have another look later some time and see if I can help answer the question

 

A general note on filters though - there is more than one purpose for a 'filter' in a DAC system. One type of filter is in the 'digital' side of the DAC and works on the conversion of sample rates (upsampling/oversampling etc) The other type of filter is the analogue reconstruction filter that is used to remove the out of band signals from the output of the DAC modulator, this can be either a passive analogue filter at output of a voltage output DAC or built into an intermediate step in the current to voltage conversion of a current output DAC.

 

What acg is suggesting is to use the upsampling filter on the computer and remove filtering from the DAC and run it at highest supported sample rate. This means that the out of band signals are of such a frequency that some would say they no longer require the reconstruction type filter at all.

 

I dissagree and suggest that what this is doing is in effect making the interconnect carry a far wider bandwidth signal and the power amplifier input filter acts as the reconstruction filter. What is missed is that anyone that analyses these things never looks at it with a spectrum analyser of sufficient bandwidth to see that the filter is just moved to 500kHz or 1MHz or something and that to compare apples with apples the place to look at the equivalent graph would be after the power amplifier's input stage, which is impractical (as is getting a wide bandwidth spectrum analyser but that is mostly a $$$ thing).

 

That might have muddied the waters a little more rather than help but hopefully it clears up that using the word 'filter' we need to be careful which filter we're talking about!

 

Chris

Edited April 17, 2014 by hochopeper

[Orpheus]

I did have a read of the blog last night but it's long and lots of measurments so I didn't have time to fully digest it all sufficiently to answer the question in your OP. I can't get to the blog because it's blocked by my work ... I'll have another look later some time and see if I can help answer the question

 

A general note on filters though - there is more than one purpose for a 'filter' in a DAC system. One type of filter is in the 'digital' side of the DAC and works on the conversion of sample rates (upsampling/oversampling etc) The other type of filter is the analogue reconstruction filter that is used to remove the out of band signals from the output of the DAC modulator, this can be either a passive analogue filter at output of a voltage output DAC or built into an intermediate step in the current to voltage conversion of a current output DAC.

 

What acg is suggesting is to use the upsampling filter on the computer and remove filtering from the DAC and run it at highest supported sample rate. This means that the out of band signals are of such a frequency that some would say they no longer require the reconstruction type filter at all.

 

I dissagree and suggest that what this is doing is in effect making the interconnect carry a far wider bandwidth signal and the power amplifier input filter acts as the reconstruction filter. What is missed is that anyone that analyses these things never looks at it with a spectrum analyser of sufficient bandwidth to see that the filter is just moved to 500kHz or 1MHz or something and that to compare apples with apples the place to look at the equivalent graph would be after the power amplifier's input stage, which is impractical (as is getting a wide bandwidth spectrum analyser but that is mostly a $$$ thing).

 

That might have muddied the waters a little more rather than help but hopefully it clears up that using the word 'filter' we need to be careful which filter we're talking about!

 

Chris

Will be interested to hear your thoughts when you look at the blog again.

 

What sounds best to me is for the TEAC to upconvert to 24/192, but not to apply a filter to the result.

 

 

 

 

[DoggieHowser]

Will be interested to hear your thoughts when you look at the blog again.

 

What sounds best to me is for the TEAC to upconvert to 24/192, but not to apply a filter to the result.

 

If you are using JRiver, you can also try using your computer to up convert to DSD or 2xDSD. I've found that to sound best with my Playback MPS5. 

 

Probably need to dig the 501 out from storage to give that a go

[Orpheus]

If you are using JRiver, you can also try using your computer to up convert to DSD or 2xDSD. I've found that to sound best with my Playback MPS5. 

 

Probably need to dig the 501 out from storage to give that a go

Do you have a Teac UD501?

 

It is a very competent machine. I've had a dCS Puccini, dCS stack (Elgar and Verdi), and I still have a Plinius CD101, which I'm very fond of, and the TEAC is right up there for me.

 

Dit it out and compare it to your Playback. You might be surprised how close they are (not that I've heard the Playback).

 

If I had more analytical speakers, perhaps I would find it easier to separate them (I have Quad ESL2905s, which I am very fond of, but absolute resolution is not their strong point).

[DoggieHowser]

I did put it in my media room, augmenting the EMM Labs XDS-1 because that didn't have DSD USB capability but I preferred using PCM into the AES input of the XDS-1 in that system.

 

The Teac is a decent performer but it wasn't quite as resolving in complex passages as the EMM Labs was.

[Orpheus]

I did put it in my media room, augmenting the EMM Labs XDS-1 because that didn't have DSD USB capability but I preferred using PCM into the AES input of the XDS-1 in that system.

 

The Teac is a decent performer but it wasn't quite as resolving in complex passages as the EMM Labs was.

What speakers do you have? Perhaps they are more resolving than mine.

 

But perhaps resolution of that kind is not ultimately what I'm after as well, or at least I'm not prepared to accept what usually comes with it.

[DoggieHowser]

Thiel CS3.7 with Nemo amplification and ARC Ref 5SE

[Orpheus]

Thiel CS3.7 with Nemo amplification and ARC Ref 5SE

Yep, the Thiel speakers I've heard have been very detailed indeed.

 

My friend had a pair, along with a dCS Scarlatti stack and ARC ref 300 (?) mono blocks. 

 

He eventually replaced the speakers, buying Wilson Maxx speakers, which are exceedingly large, and incredibly detailed.

 

I can't afford that sort of gear!

 

I also prefer something a bit different, as my system indicates.

 

But I do appreciate the pinpoint imaging and incredibly competent reproduction of what his system is fed. It is just that it is so good, it somehow calls attention to the fact that it is not real, ironically, because it nearly is.

[davewantsmoore]

What sounds best to me is for the TEAC to upconvert to 24/192, but not to apply a filter to the result.

 

Edit:  nevermind.   It's been a long day  :/

 

 

Edit2:    Yes.  That is what I would also do or further.   This type of thing is the sole benefit of higher sample rates IMHO.    Allowing us not to do any further harm to the signal.

 

What Chris said is a relevant concern though.

Edited April 17, 2014 by davewantsmoore

[DoggieHowser]

The UD501 is a pretty good performer but it's not IMHO a giant killer. I thought the Mytek was more resolving as well but I did prefer the sound of the UD501. Maybe it's the 1795 sound - which is also used in my Bel Canto DAC3.5VB and the original Ayre QB9.

[acg]

[quote name="hochopeper" post="1091309"

What acg is suggesting is to use the upsampling filter on the computer and remove filtering from the DAC and run it at highest supported sample rate. This means that the out of band signals are of such a frequency that some would say they no longer require the reconstruction type filter at all.

I dissagree and suggest that what this is doing is in effect making the interconnect carry a far wider bandwidth signal and the power amplifier input filter acts as the reconstruction filter. What is missed is that anyone that analyses these things never looks at it with a spectrum analyser of sufficient bandwidth to see that the filter is just moved to 500kHz or 1MHz or something and that to compare apples with apples the place to look at the equivalent graph would be after the power amplifier's input stage, which is impractical (as is getting a wide bandwidth spectrum analyser but that is mostly a $$$ thing).

That might have muddied the waters a little more rather than help but hopefully it clears up that using the word 'filter' we need to be careful which filter we're talking about!

Chris

Yes Chris, digital filter only, I doubt that there is a way that Orpheus could change or remove the analogue reconstruction filter even if he wanted to...I dare say the Teac in nos mode would not be like some pure nos dacs that do not have any analogue filtering so his dac should output the same bandwidth as it normally does so no danger there.

Orpheus, you can just download the free trial of xxhe and try it for yourself.

Anthony

[DoggieHowser]

I finally got around to hooking up the UD501 recently. 

 

I used a cheap longish USB cable (a cheapish NuForce impulse USB) to connect my newly built (almost silent PC linked here ) to run JRiver MC 19 and set it to output in 2xDSD to the Teac.

 

Initial impressions are that it's very good with my PCM material vs the time I used my Mac Mini with Audirvana Plus (Integer and Direct mode enabled) to the same DAC.

 

I thought the PCM up sampled DSD2x sounded more natural with a deeper soundstage depth and I was able to make out more of the background instruments. But it does sound a wee bit harsh when it is just plugged in. I am going to leave it playing on repeat for a bit longer before I do some critical listening. 

Edited April 24, 2014 by DoggieHowser

[Orpheus]

I finally got around to hooking up the UD501 recently.

I used a cheap longish USB cable (a cheapish NuForce impulse USB) to connect my newly built (almost silent PC linked here ) to run JRiver MC 19 and set it to output in 2xDSD to the Teac.

Initial impressions are that it's very good with my PCM material vs the time I used my Mac Mini with Audirvana Plus (Integer and Direct mode enabled) to the same DAC.

I thought the PCM up sampled DSD2x sounded more natural with a deeper soundstage depth and I was able to make out more of the background instruments. But it does sound a wee bit harsh when it is just plugged in. I am going to leave it playing on repeat for a bit longer before I do some critical listening.

Interesting. Look forward to hearing more.

[DoggieHowser]

Spent the day listening to the Thiels. Was quite sad to hear that the new owners have divested the plant and equipment in Lexington. I wonder if we'll even be able to get replacement drivers. I don't often get to give these spokers a proper dose of music at my (acceptable) listening levels with the missus in the house but she was out today.

 

I was trying to test out my new quiet PC running JRiver to output 2xDSD to the Teac UD-501. You may recall I had my Mac Mini using iTunes/AudirvanaPlus (Direct/Integer mode) to the Teac a while back but it was a little underwhelming. I went back to using the Mac hooked to a Weiss INT202 sending AES to the XDS-1 (PCM). And the Teac was unplugged and left on a shelf for months.

 

But a recent test with the Playback Designs MPS-5 getting fed 2xDSD with the quiet Windows PC and then Mac OS X JRiver got me intrigued if I could achieve better results with the Teac.

 

I had hooked up the Teac to the PC a day ago and initial impressions were encouraging. But the sound was a bit harsh to me with the unit just plugged in from cold so I left the PC and the Teac on repeat for the last day or so. 

 

This afternoon, I went through my music library putting through a whole bunch of songs from totally different genres - from Celine Dion to Rammstein, from Chopin to Michael Giacchino, from Korean to Taiwan, 60s-70s rock to modern material, from Depeche Mode to Daft Punk electronica. 

 

Have to say I was pretty impressed, both with the Thiels and the Teac. The soundstage from the CS3.7s is really something, extending beyond the 8 feet they straddled, in some cases, I was a little surprised that I thought I heard the effects going to my wide speakers - had to double check they were off

 

The Teac also handled everything with aplomb. I was amazed at how good PCM up sampled to DSD sounded with JRiver. I actually preferred the PC/Teac/DSD combo to the Mini/A+/Weiss INT202/EMM Labs combo for computer audio. 

 

I did eventually put a CD inside the XDS-1 for comparison. Sadly, the EMM Labs with a spinning disc still sounded better than the Teac with the computer performing up sampling (I was able to make out more instruments/details/layers with the CD spinning than the same track on the computer) but it has convinced me to leave the Teac permanently plugged in. 

 

[Orpheus]

Very interesting. I certainly think it is pretty competent.

 

The absolutely best computer audio I have heard was when I had the dCS Puccini and u-Clock.

 

With the dCS set-up, I really didn't hear a difference between CD/SACD and computer audio.

 

But really, I couldn't justify keeping it, at that time.

 

Thiels are excellent speakers. I think they would more easily allow you to differentiate between different digital sources than my Quad 2905s.

[LogicprObe]

I'll be starting a thread when I do a reformat and load up the XXHighEnd software.

Hopefully this week. (lots of tax to do but the computer might fall over before then)

[Scratcher]

I am very happy with the UD-501, best dsd capable DAC I have heard in my system so far.

It's very well designed and flexible with options a plenty.

I am running a fanless passive 2012r2 server with AO via a Vertere d-if USB cable to the Teac.

It also has the option to change xlr hot pin to 2 or 3, very handy for Accuphase gear

I find the upsampling very pleasing with poorly recorded/mastered music.

Too me PCM nos mode seems very good, the slow and sharp filters are noticeably different but not better to my ears.

The DSD filters seem very similar besides the obvious gain differences.

I can't hear any difference using the IFi Iusb so I'm guessing the DAC draws it's power from onboard.