Orpheus Zero Review (CD Player)

[JohnPM]

This reminds me that the REW generator 16 bits Jitter test signal was wrong last time I tried.

REW doesn't have a 16-bit J-test signal, the signal is 24-bit.

 

[JW001]

Fine ... but though I still buy CDs I have actually played one in several years.

Yes! Buying a new CD, opening EAC, half an hour and I can play it everywhere in the house without a damn Orpheus Zero

 

[NTTY]

This looks strange.
It should rather look like this:

View attachment 396625

This reminds me that the REW generator 16 bits Jitter test signal was wrong last time I tried.
(24 and 32 bits are correct)

I'll check again.

Hi,

Using Multitone from @pkane and creating a WAV from the JTest @16bits 44.1kHz, this is what I get from loopback with my interface, but also from the SPDIF out of the Orpheus after having burnt the WAV on a CD:

It means it's created with dither.

And this is the analog output of the Orpheus (XLR) with the same test file (from CD of course):

It sounds better and gives a better picture. Do I keep it?

Thanks

 

[AnalogSteph]

If you could do me a -16dBFS sine with shape dither, it could be a direct comparison.

Your wish is my command.
(You did see the -20, -40, -60 and -90 dB files I also provided originally, right?)
(I'm using a very simple Foobar2000 converter preset to encode these FLACs, so the reverse should also work. I've had some issues with Audacity in this regard. If you want to be 100% sure, FLAC on the command line or with a FLAC frontend should always work. I couldn't get WAVs nearly as small.)

In fact I tried to replicate following your instructions to create a shape dithered file but I always end up with very low level of distorsion in the file, so it’s not as good as yours.

Hmm. I'm literally just generating a 997 Hz 0.999 amplitude tone in Audacity 2.4.2 with default sample format = 24 bit, then using Amplify to reduce the level as required and exporting as 16 bit WAV (with a high-quality dither type setting of "Shaped", obviously).

As for your J-Test problem - HA to the rescue.

Would there be a simple way for me to test that?

Do you have any "dumb" RCA to 1/4" TS cables? Then the computer just needs to be grounded, or preferably just the interface, otherwise results could be unrealistically bad.

But I’m not sure of what you meant about contact cleaner need.

Chips in sockets are generally more prone to contact issues than those soldered in. If you get an '80s home computer in that's acting funny and has tons of socketed chips, the first thing to do would be reseating those, especially if the sockets are known to be of the cheap and nasty kind. Even PC expansion cards with their gold-plated contacts are not entire immune. (You can do better, of course. I don't think I've ever heard of anyone having issues with Intel LGA775 sockets.) Now I assume the TEAC has always been kept in your living spaces, plus Switzerland isn't exactly known for its humid climate, so the odds of there being any issues right now after not even 20 years are fairly slim.

 

[NTTY]

Your wish is my command.

Coooool, thank you!

(You did see the -20, -40, -60 and -90 dB files I also provided originally, right?)

Of course, they are now part of my test CD
And that's one more to add for the to be version 6.5

(I'm using a very simple Foobar2000 converter preset to encode these FLACs, so the reverse should also work. I've had some issues with Audacity in this regard. If you want to be 100% sure, FLAC on the command line or with a FLAC frontend should always work. I couldn't get WAVs nearly as small.)

Ok, thanks for the info.

Hmm. I'm literally just generating a 997 Hz 0.999 amplitude tone in Audacity 2.4.2 with default sample format = 24 bit, then using Amplify to reduce the level as required and exporting as 16 bit WAV (with a high-quality dither type setting of "Shaped", obviously).

Got you, I was nearly there. I just did not know Audacity offered a generator. I gave it a try again, and I got it (0.999 amplitude, not 1, thanks!)::

As for your J-Test problem - HA to the rescue.

This is great. The 16bits/44.1kHz JTest there has no dither and looks really nice.

I'll go with this one.

Do you have any "dumb" RCA to 1/4" TS cables?

Yep, and Motu recommends to use TRS with the ring floating, and so I did cables like that too. It indeed improves a little noise and distortion, but way below the CD player, so that does not change the measurements, especially with TPDF dither.

Then the computer just needs to be grounded, or preferably just the interface, otherwise results could be unrealistically bad.

Ok, I think you already mentioned that to me. I'll try too.

Chips in sockets are generally more prone to contact issues than those soldered in. If you get an '80s home computer in that's acting funny and has tons of socketed chips, the first thing to do would be reseating those, especially if the sockets are known to be of the cheap and nasty kind. Even PC expansion cards with their gold-plated contacts are not entire immune. (You can do better, of course. I don't think I've ever heard of anyone having issues with Intel LGA775 sockets.) Now I assume the TEAC has always been kept in your living spaces, plus Switzerland isn't exactly known for its humid climate, so the odds of there being any issues right now after not even 20 years are fairly slim.

Ah ok I understand now. Indeed, the Orpheus stayed in my living room for the last 17 years or so. And true, it's never been exposed to humidity.

 

[AnalogSteph]

Got you, I was nearly there. I just did not know Audacity offered a generator. I gave it a try again, and I got it (0.999 amplitude, not 1, thanks!)::

You can't go to exactly 1 because of the asymmetry in digital integer values, e.g. 16 bits goes from -32767 to +32768. An amplitude 1.0 sine would extend from -32768 to +32768, which you obviously don't quite have on the negative side. Hence slight clipping and associated distortion. (You may have seen RG peaks of 0.999969 before, that's -32767.) It gets worse once you introduce fancy dither which might peak at +/-11 or so, so worst-case you need that much space around the sine on top. That leaves us with a maximum of 0.9996 and change. 0.999 is safe and only -0.0087 dBFS anyway (or -0.0084 dBFS, depending on how you define 0 dBFS).

This gets really funny once you try dithering to 8 bits. By the time you've got enough headroom for fancy dither, you're at like -0.8 dBFS. Still worth it though.

I bet this has something to do with why mastering folks didn't like the sound when using dither on traditional (NOS) brickwall limiters and thus generally preferred leaving it off. (Congratulations, you just won a fraction of a dB on top - or maybe two if we're being generous - and lost like 24 at the bottom. ) Good riddance to those things anyway, they're a prime example for nonlinear operations in the digital domain not being inherently band-limited and causing lots of aliasing. Anything of the sort works much better when oversampled to give it more "breathing room" in the spectrum, from brickwall limiters to regular old compressors and tape simulators.

 

[Rja4000]

REW doesn't have a 16-bit J-test signal, the signal is 24-bit.

Hi John
Thanks for your feedback.

But in the REW generator, you may generate and save J-test signal to a file with 16 bits selected. And you indeed get a 16 bits file.
That's the one shown here.

So, at least, that is confusing.

 

[restorer-john]

Yes! Buying a new CD, opening EAC, half an hour and I can play it everywhere in the house without a damn Orpheus Zero

This is a review of the Orpheus Zero CD player, not some "rip-it and stream it" pile of software, PCs and network hardware...

 

[NTTY]

You can't go to exactly 1 because of the asymmetry in digital integer values, e.g. 16 bits goes from -32767 to +32768. An amplitude 1.0 sine would extend from -32768 to +32768, which you obviously don't quite have on the negative side. Hence slight clipping and associated distortion. (You may have seen RG peaks of 0.999969 before, that's -32767.)

In Two's complement representation, yeah I see.
32767 / 32768 = 0.999969.

I did not know about replay gain, so I searched and found, thanks

BTW, amplitude 1 gives this (with shape dither, to compare with the one I published before):

And by the way, this is exactly what I got when first creating 0dBFS 997Hz sine in 24bits with REW, and then used Audacity to save it as 16bits WAV with Shaped dither. And this is why I mentioned I could not get the same quality as your files.

It gets worse once you introduce fancy dither which might peak at +/-11 or so, so worst-case you need that much space around the sine on top. That leaves us with a maximum of 0.9996 and change. 0.999 is safe and only -0.0087 dBFS anyway (or -0.0084 dBFS, depending on how you define 0 dBFS).

Got you: 20 x log(0.999) = -0.00869dBFS

This gets really funny once you try dithering to 8 bits. By the time you've got enough headroom for fancy dither, you're at like -0.8 dBFS. Still worth it though.

I bet this has something to do with why mastering folks didn't like the sound when using dither on traditional (NOS) brickwall limiters and thus generally preferred leaving it off. (Congratulations, you just won a fraction of a dB on top - or maybe two if we're being generous - and lost like 24 at the bottom. ) Good riddance to those things anyway, they're a prime example for nonlinear operations in the digital domain not being inherently band-limited and causing lots of aliasing. Anything of the sort works much better when oversampled to give it more "breathing room" in the spectrum, from brickwall limiters to regular old compressors and tape simulators.

So much wasted for a fraction of a dB, as you say. Thanks for the insights, one more time!

 

[NTTY]

@Rja4000 @MC_RME @AnalogSteph @sam_adams I updated the JTest results (first page) from the recommended source by @AnalogSteph.

So it went from that:

To that:

The test file (shown in red here and captured from the digital output of the Orpheus) is now indeed better. Blue trace is the output from analog XLR.

Cheers

 

[AdrianusG]

Hello Everyone,

This is a review and detailed measurements of the Orpheus Zero (Mk1).

View attachment 396518

As I already wrote in several of my previous reviews, I like testing CD Players, especially older ones. This one is "only" 20 years old, though

Stereophile briefly talked about this player but did not measure it. Let me fix that.

My firs review of the Onkyo C-733 here contains information about my measurements which I align with the AES standard (to few exceptions). It means that, over time, you can compare the devices I reviewed.


Orpheus Zero - Presentation

Released in 2005, this is a Swiss made CD player, high quality that means. The price was very high at the time, and it was reading CD Audio only.
This is the firs version, there was an MKII version with a blue display.

View attachment 396520

It features a professional look, with its rackable 1RU size, except it does not have the necessary holes
It is a top loader using a Philips CD Pro drive. Conversion is done by a Wolfson DAC WM8740. All of that is relatively simple and housed in a heavy enclosure.

Back panel shows XLR and RCA analog outputs, with Coax and XLR digital outputs:

View attachment 396521

Despite the slim look, it weight more than 22lbs (10kg), it's always a surprise when I need to carry it.

When opening (manually) the top door, two red led give it a nice touch at night, I like a lot:

View attachment 396523

There are not so many pictures of this player on the web, so let's quickly check the inside:

View attachment 396525

This is well organized, no surprise. At the bottom of this picture, you can see the large servo control card with digital output, and just above it, on the right, the small D/A card with the WM8740.

I acquired this player in 2007 and it's been my main player for all these years. I compared it to many other players. I tried to use external DACs in the foolish hope I could improve the sound, but no. I continue to use it as a CD player today, not as a transport. So that player and I have shared a lot of music, and so many personal events... It has a special place in my heart.

For the last 2 years, I trained myself to measure CD players, and I was really afraid of the day that the Orpheus would go under (my) review. What if it did not perform? All these years of thinking I could not find better, was I wrong?

Time has come to verify. Follow me

View attachment 396527


Orpheus Zero - Measurements (Analog outputs - From CD)

From now on, I will be consistent with my measurements as I described them on the Onkyo C-733 review. So over time, this will help comparing the items I reviewed.

From both RCA and XLR, the Orpheus Zero outputs 2Vrms as per its specs (0.01dB less on the XLR though). The balanced outputs of The Orpheus Zero invert absolute polarity. That said, the interface allows to invert the polarity, and it shows as is (below) not to forget about it:

View attachment 397057

Let's start with the standard 1kHz sine @0dBFS (dithered) from my test CD (XLR out):

View attachment 396529

Left and right channels are shown but only one gets evaluated in that view. Both channels have nearly the same performances. Plot is on H2 (-110.8dBr and -110.9dBr).

THD+N is limited (and so is the SINAD) by the dither noise present on my test CD. It's the best we can get. It shows 92.9dBr on the dashboard, which is 1dB better than the best CD Players I previously measured, and despite fact I'm measuring here with 1dB headroom on my interface.

The RCA outputs do better on THD perspective (3dB less) but show a little power supply related noise of no concern (below -110dBr). I'll continue this review from XLR only.

Note as well the very precise 1000.00Hz I get from a 1kHz test tone. Not Clock deviation even after 20 years of nearly everyday utilization. It's the Swiss Precision spirit

For few reviews, I've been adding a view of 1kHz @-6dBFS, so here we go:

View attachment 396530

The distorsion went down a little (-107.4dB), which is very good and better than the specs.

Other results (not shown) are:

• Crosstalk : -137dB at 1kHz and -116dB @10kHz
• SNR : 97.2dB (1kHz @-60dBFS, no dither)
• IMD AES : -84.4dB (18kHz+20kHz 1:1 @-5dBFS)

I suppose you saw a very silent power supply above, and since I usually add this view, let's have a look:

View attachment 396532

Bandwidth measurements(@-12dBFS below) showed a gentle roll of at 20kHz (-1dB) starting very early (10kHz),:

View attachment 396533

The two channels are shown and are perfectly matched (hence only one trace seen), I like to see that (it's not always the case as you can tell from my other reviews).

Let's have a look at the oversampling filter (Overlay of White Noise and 18k+20Hz dual tones):

View attachment 396555

This view shows that It is a relatively slow filter with a decent attenuation (roughly -90dB) from the periodic white noise (blue trace). The aliases of 18k and 20kHz test tones are attenuated below -90dB which is very good.

Multitone test showed no issues (1/10 decade):

View attachment 396534

We have much more than the Audio CD free of distortion.

Updated Jitter test (16bits/44.1kHz) with a better test file, as per your comments (see below and thanks to @Rja4000 for having initiated the discussion):

View attachment 397051

Red trace is what is on the test CD (from the digital output), it can’t be better and is what's recorded from the test file. The Orpheus (blue trace) adds a (very?) little amount of Jitter.

Started with the Teac VRDS-20 review, and on your request + support to get it done (more here), I'm adding the "intersample-overs" test which intends to identify the behavior of the digital filtering and DAC when it comes to process near clipping signals. Because of the oversampling, there might be interpolated data that go above 0dBFS and would saturate (clip) the DAC and therefore the output. And this effect shows through distorsion (THD+N measurement up to 96kHz):

Intersample-overs tests Bandwidth of the THD+N measurements is 20Hz - 96kHz	5512.5 Hz sine, Peak = +0.69dBFS	7350 Hz sine, Peak = +1.25dBFS	11025 Hz sine, Peak = +3.0dBFS
Yamaha CD-1 (Non-Oversampling CD Player)	-79.6dB	-35.3dB	-78.1dB
Orpheus Zero	-80.1dB	-29.7dB	-57.3dB
Onkyo C-733	-79.8dB	-29.4dB	-21.2dB
Denon DCD-900NE	-34.2dB	-30.4dB	-19.1dB

I kept some references and will keep the same for other reviews, so you can quickly compare. The results of the Orpheus Zero mean the oversampling filter has a good level of headroom, meaning that the resolution of the Wolfson DAC was put in good use.

And let me finish with my favourite measurement, the THD (excluding noise) vs Frequency at -12dBFS:

View attachment 396547

The Orpheus Zero has no issue here, it is the best trace that I have in my collection of this specific measurement. This an easy test for advanced 1bit DACs. I like this measurement because it shows lack of linearity already at this level with older R2R architectures, and some lower resolution 1bit architectures too, that I enjoy testing.


Orpheus Zero - Measurements (Digital out)

A number of you are into using CD players as transports, and like to know how the digital output performs.

Well, it is perfect. I’ll keep it simple, with what I believe to be the most representative measurement of the digital output quality, and that is a 1kHz sine at -90.31dBFS which shows the 3DC levels of the smallest digital signal in 16bits sign magnitude representation:

View attachment 396549

I have to note that the signal is extremely stable and would facilitate the job of the DAC's PLL reading it.

The second view below shows what's on the test CD (from the standard 1kHz @0dBFS measurement):

View attachment 396551

It can't get better than that with my test CD.


Conclusion

Well, I was so anxious to test it, I should have done that way before! Fact is that this player equals the best nearly everywhere, and does better in some cases (ie intersample-overs resistance).

After all, if I never found better, it's simple because this player does not disrupt the CD Audio content. That was kind of a statement from Orpheus Laboratories (now OrpheusLab). It had a price, yes indeed, but definitely not snake oil.

I hope you enjoyed the review as much as I enjoyed writing it. Let me know how to improve and if you have questions. I have recorded all the 46 measurements. If you want me to publish others or run one of your choice, feel free to ask. I will add some new measurements later for the fun of it about some noise shaping test files that @AnalogSteph sent me, and this is quite interesting.

Enjoy your weekend!

--------
Flo

Great test and review

I like those Top loader DAC's especially with this Philips pro mechanism, it shows that they really were at the top with this back then.

 

[Jimmyblues1959]

Hello Everyone,

This is a review and detailed measurements of the Orpheus Zero (Mk1).

View attachment 396518

As I already wrote in several of my previous reviews, I like testing CD Players, especially older ones. This one is "only" 20 years old, though

Stereophile briefly talked about this player but did not measure it. Let me fix that.

My firs review of the Onkyo C-733 here contains information about my measurements which I align with the AES standard (to few exceptions). It means that, over time, you can compare the devices I reviewed.


Orpheus Zero - Presentation

Released in 2005, this is a Swiss made CD player, high quality that means. The price was very high at the time, and it was reading CD Audio only.
This is the firs version, there was an MKII version with a blue display.

View attachment 396520

It features a professional look, with its rackable 1RU size, except it does not have the necessary holes
It is a top loader using a Philips CD Pro drive. Conversion is done by a Wolfson DAC WM8740. All of that is relatively simple and housed in a heavy enclosure.

Back panel shows XLR and RCA analog outputs, with Coax and XLR digital outputs:

View attachment 396521

Despite the slim look, it weight more than 22lbs (10kg), it's always a surprise when I need to carry it.

When opening (manually) the top door, two red led give it a nice touch at night, I like a lot:

View attachment 396523

There are not so many pictures of this player on the web, so let's quickly check the inside:

View attachment 396525

This is well organized, no surprise. At the bottom of this picture, you can see the large servo control card with digital output, and just above it, on the right, the small D/A card with the WM8740.

I acquired this player in 2007 and it's been my main player for all these years. I compared it to many other players. I tried to use external DACs in the foolish hope I could improve the sound, but no. I continue to use it as a CD player today, not as a transport. So that player and I have shared a lot of music, and so many personal events... It has a special place in my heart.

For the last 2 years, I trained myself to measure CD players, and I was really afraid of the day that the Orpheus would go under (my) review. What if it did not perform? All these years of thinking I could not find better, was I wrong?

Time has come to verify. Follow me

View attachment 396527


Orpheus Zero - Measurements (Analog outputs - From CD)

From now on, I will be consistent with my measurements as I described them on the Onkyo C-733 review. So over time, this will help comparing the items I reviewed.

From both RCA and XLR, the Orpheus Zero outputs 2Vrms as per its specs (0.01dB less on the XLR though). The balanced outputs of The Orpheus Zero invert absolute polarity. That said, the interface allows to invert the polarity, and it shows as is (below) not to forget about it:

View attachment 397057

Let's start with the standard 1kHz sine @0dBFS (dithered) from my test CD (XLR out):

View attachment 396529

Left and right channels are shown but only one gets evaluated in that view. Both channels have nearly the same performances. Plot is on H2 (-110.8dBr and -110.9dBr).

THD+N is limited (and so is the SINAD) by the dither noise present on my test CD. It's the best we can get. It shows 92.9dBr on the dashboard, which is 1dB better than the best CD Players I previously measured, and despite fact I'm measuring here with 1dB headroom on my interface.

The RCA outputs do better on THD perspective (3dB less) but show a little power supply related noise of no concern (below -110dBr). I'll continue this review from XLR only.

Note as well the very precise 1000.00Hz I get from a 1kHz test tone. Not Clock deviation even after 20 years of nearly everyday utilization. It's the Swiss Precision spirit

For few reviews, I've been adding a view of 1kHz @-6dBFS, so here we go:

View attachment 396530

The distorsion went down a little (-107.4dB), which is very good and better than the specs.

Other results (not shown) are:

• Crosstalk : -137dB at 1kHz and -116dB @10kHz
• SNR : 97.2dB (1kHz @-60dBFS, no dither)
• IMD AES : -84.4dB (18kHz+20kHz 1:1 @-5dBFS)

I suppose you saw a very silent power supply above, and since I usually add this view, let's have a look:

View attachment 396532

Bandwidth measurements(@-12dBFS below) showed a gentle roll of at 20kHz (-1dB) starting very early (10kHz),:

View attachment 396533

The two channels are shown and are perfectly matched (hence only one trace seen), I like to see that (it's not always the case as you can tell from my other reviews).

Let's have a look at the oversampling filter (Overlay of White Noise and 18k+20Hz dual tones):

View attachment 396555

This view shows that It is a relatively slow filter with a decent attenuation (roughly -90dB) from the periodic white noise (blue trace). The aliases of 18k and 20kHz test tones are attenuated below -90dB which is very good.

Multitone test showed no issues (1/10 decade):

View attachment 396534

We have much more than the Audio CD free of distortion.

Updated Jitter test (16bits/44.1kHz) with a better test file, as per your comments (see below and thanks to @Rja4000 for having initiated the discussion):

View attachment 397051

Red trace is what is on the test CD (from the digital output), it can’t be better and is what's recorded from the test file. The Orpheus (blue trace) adds a (very?) little amount of Jitter.

Started with the Teac VRDS-20 review, and on your request + support to get it done (more here), I'm adding the "intersample-overs" test which intends to identify the behavior of the digital filtering and DAC when it comes to process near clipping signals. Because of the oversampling, there might be interpolated data that go above 0dBFS and would saturate (clip) the DAC and therefore the output. And this effect shows through distorsion (THD+N measurement up to 96kHz):

Intersample-overs tests Bandwidth of the THD+N measurements is 20Hz - 96kHz	5512.5 Hz sine, Peak = +0.69dBFS	7350 Hz sine, Peak = +1.25dBFS	11025 Hz sine, Peak = +3.0dBFS
Yamaha CD-1 (Non-Oversampling CD Player)	-79.6dB	-35.3dB	-78.1dB
Orpheus Zero	-80.1dB	-29.7dB	-57.3dB
Onkyo C-733	-79.8dB	-29.4dB	-21.2dB
Denon DCD-900NE	-34.2dB	-30.4dB	-19.1dB

I kept some references and will keep the same for other reviews, so you can quickly compare. The results of the Orpheus Zero mean the oversampling filter has a good level of headroom, meaning that the resolution of the Wolfson DAC was put in good use.

And let me finish with my favourite measurement, the THD (excluding noise) vs Frequency at -12dBFS:

View attachment 396547

The Orpheus Zero has no issue here, it is the best trace that I have in my collection of this specific measurement. This an easy test for advanced 1bit DACs. I like this measurement because it shows lack of linearity already at this level with older R2R architectures, and some lower resolution 1bit architectures too, that I enjoy testing.


Orpheus Zero - Measurements (Digital out)

A number of you are into using CD players as transports, and like to know how the digital output performs.

Well, it is perfect. I’ll keep it simple, with what I believe to be the most representative measurement of the digital output quality, and that is a 1kHz sine at -90.31dBFS which shows the 3DC levels of the smallest digital signal in 16bits sign magnitude representation:

View attachment 396549

I have to note that the signal is extremely stable and would facilitate the job of the DAC's PLL reading it.

The second view below shows what's on the test CD (from the standard 1kHz @0dBFS measurement):

View attachment 396551

It can't get better than that with my test CD.


Conclusion

Well, I was so anxious to test it, I should have done that way before! Fact is that this player equals the best nearly everywhere, and does better in some cases (ie intersample-overs resistance).

After all, if I never found better, it's simple because this player does not disrupt the CD Audio content. That was kind of a statement from Orpheus Laboratories (now OrpheusLab). It had a price, yes indeed, but definitely not snake oil.

I hope you enjoyed the review as much as I enjoyed writing it. Let me know how to improve and if you have questions. I have recorded all the 46 measurements. If you want me to publish others or run one of your choice, feel free to ask. I will add some new measurements later for the fun of it about some noise shaping test files that @AnalogSteph sent me, and this is quite interesting.

Enjoy your weekend!

--------
Flo

Excellent review! I really enjoy reading about quality vintage gear. My 30 plus year old Naim CDi still
sounds quite good. It's built to last. While Naim no longer services them there are independent shops
that still do.

 

[NTTY]

FYI, I also updated Jitter measurements for the Denon DCD-SA1 and the Denon DCD-3560.

 

[JW001]

This is a review of the Orpheus Zero CD player, not some "rip-it and stream it" pile of software, PCs and network hardware...

No PCs - a Raspberry Pi with a 2TB SSD, attached to the home network via 1Gb/s interface.

Sorry for the distraction, I actually liked the review.

 

[dzerig]

The Wolfson DACs were part of the reason people felt there was a sound associated with the DAC.

I wonder if something like the THD vs frequency at different dBFS is why…

Sinad is a over general term. In speaker reviews and headphone reviews here, Amir notes the strong human ear response to 2-3 kHz tones implies if you have distortion here it sounds worse than power supply noise mostly at low frequencies, which is less offensive to the human ear).

Does distortion vs frequency matter with dacs, as it is so low most is lower than the source especially those of us who listen to music mastered on reel to reel? Probably not, but if people are hearing differing sounds in blind tests, and the frequency response is flat, then I have no other explanation.

Great review, thanks! I love playing LPs and CDs and enjoy music more that way. Something about the tics and such in LPs make music sound better.

 

[AnalogSteph]

Does distortion vs frequency matter with dacs, as it is so low most is lower than the source especially those of us who listen to music mastered on reel to reel?

Back in 2011-ish, there were some buggy Realtek HDA driver versions that would not init the device correctly in 44.1 kHz so the hardware would run at the wrong sample rate and I assume linear interpolation was being used internally, leading to distortion being up to 0.1% (-60 dB) by 10 kHz. That was definitely audible. Anything you find on half-decent DACs should be perfectly fine though. (With a fixed driver, the Realteks sounded fine, too. The usual models have actually been without any audible flaws for over a decade now.)

DACs do tend to differ in how they react to overs (<0 dBFS levels), so results with heavily brickwalled loudness war era CDs may vary. You would think that anything with a digital volume control would have it prior to the filter and possible ASRC and could be bailed out like that, but apparently this is not always the case. (Apparently a current ES9039Q2M is fine, but an old ES9018 is not, and it's currently unclear when the change occurred.)

 

[sam_adams]

I updated the JTest results (first page) from the recommended source

The links posted in that HA thread are from code written and files created by user wakibaki over on DIY Audio four years earlier in 2011:

Jitter Test Signal (J-test Signal) with MATLAB

credit where it is due.

The SoX code snips linked here use the 'play' command from SoX and only result in playback and not files that can be saved for use. Additionally, the 16-bit output is actually 32-bit output. The snips also indicate the use of the 'bc' command which will only be found on macOS or UNIX/Linux systems. The lines below:

JL=$(bc -l <<< "dbbit=20*l(2)/l(10); 16*dbbit")

and

JL=$(bc -l <<< "dbbit=20*l(2)/l(10); 24*dbbit")

define the variable 'JL' in the code snip and will only work on the above systems. The output for both 16-bit and 24-bit from those commands are below:

16-bit:

bc -l <<< "dbbit=20*l(2)/l(10); 16*dbbit"
96.32959861247398246768

24-bit:

bc -l <<< "dbbit=20*l(2)/l(10); 24*dbbit"
144.49439791871097370152

Substitution could be made in the snip . . .

To create files to be saved and played or burned to a CD, the 'play' command cannot be used since the output cannot be redirected to a file. Instead, the 'sox' command must be used and the code snip below can be used to create files (16-bit/44.1 KHZ macOS example):

16-bit 44.1 KHz .wav:

JL=$(bc -l <<< "dbbit=20*l(2)/l(10); 16*dbbit")

/usr/local/bin/sox -V -D -b 16 -r 44100 -n /path/to/file/jtest1644.wav synth 60 square 11025 square 229.6875 square 0.00001 remix 1v0.5,2p-$JL,3i-$JL channels 2

using the '-b 16' flag forces the bit depth to 16-bits. The '60' after the 'synth' flag sets the file length to sixty seconds. The bitdepth, bitrate, and sample rate can be verified afterwards by running the command:

/usr/local/bin/sox -V /path/to/file/jtest1644.wav -n stats

/usr/local/bin/sox: SoX v14.4.2
/usr/local/bin/sox INFO formats: detected file format type `wav'

Input File : '/path/to/file/jtest1644.wav'
Channels : 2
Sample Rate : 44100
Precision : 16-bit
Duration : 00:01:00.00 = 2646000 samples = 4500 CDDA sectors
File Size : 10.6M
Bit Rate : 1.41M
Sample Encoding: 16-bit Signed Integer PCM
Endian Type : little
Reverse Nibbles: no
Reverse Bits : no


Output File : '' (null)
Channels : 2
Sample Rate : 44100
Precision : 16-bit
Duration : 00:01:00.00 = 2646000 samples = 4500 CDDA sectors

/usr/local/bin/sox INFO sox: effects chain: input 44100Hz 2 channels
/usr/local/bin/sox INFO sox: effects chain: stats 44100Hz 2 channels
/usr/local/bin/sox INFO sox: effects chain: output 44100Hz 2 channels
Overall Left Right
DC offset -0.000015 -0.000015 -0.000015
Min level -0.500031 -0.500031 -0.500031
Max level 0.500000 0.500000 0.500000
Pk lev dB -6.02 -6.02 -6.02
RMS lev dB -6.02 -6.02 -6.02
RMS Pk dB -6.02 -6.02 -6.02
RMS Tr dB -6.05 -6.05 -6.05
Crest factor - 1.00 1.00
Flat factor 6.02 6.02 6.02
Pk count 1.32M 1.32M 1.32M
Bit-depth 16/16 16/16 16/16
Num samples 2.65M
Length s 60.000
Scale max 1.000000
Window s 0.050

After creating the files using the above code snips, you can import them into the RTA window in REW using the 'WAV' button and apply whatever FFT length, window function, and averaging you wish to apply—at the time of import—so you can decide whether-or-not you are looking at the correct spectrum of the file or an illusion.

 

[NTTY]

Your wish is my command.

BTW, it's indeed -16dBFS but 916Hz sine tone And no issue, I created a new one.

 

[AnalogSteph]

BTW, it's indeed -16dBFS but 916Hz sine tone

D'oh. I don't know at what point the project frequency got changed to 48 kHz or why, but it was. Corrected.

997 Hz in a 48 kHz file played back at 44.1 is indeed 915.99375 Hz, as you found out.

 

[JohnPM]

Hi John
Thanks for your feedback.

But in the REW generator, you may generate and save J-test signal to a file with 16 bits selected. And you indeed get a 16 bits file.
That's the one shown here.

So, at least, that is confusing.

OK, I have added a 16 or 24 bit choice for the next build and 16-bit is forced if saving to a 16-bit format.