Science Thread: Review of Audioquest Jitterbug and Uptone Regen USB Conditioners

mansr

New Member
Sep 20, 2015
46
0
0
There seems to be some kind of moral sliding scale to product engineering in hi end audio. It goes like this. The less one charges for a product, the less one needs to ensure accurate representations were made marketing that product. To me, the less expensive worthless tweeks are the most harmful because they end up in the hands of many more folks. Maybe the Regen falls into this category, maybe it doesn't. But saying that Uptone is exempt from making accurate representations about their product because it's only $175, isn't fair, IMO.
You got halfway there. If the product is less than $200, anything goes. If it is more than $10k, ditto. The amount of justification demanded peaks somewhere around $1k-$5k.
 

BE718

New Member
Oct 1, 2015
354
0
0
Hi

I performed some further measurements on my rgen and MDAC this evening. Interesting findings. I too have found that when connected single ended the regen induces what I presume is 8kHz USB packet noise.

BTW this isnt a J test, its just at 12kHz.

MDAC without regen balanced output.



MDAC with regen balanced output



MDAC without regen single ended output



MDAC with regen single ended output........mmmmmmm..........

 
Apr 3, 2010
16,022
0
0
Seattle, WA
Amir, try that with an 1m long IC cable. And do not measure the cable itself (its electrical parameters), but try measuring what comes out of your preamp or DAC (as this is what matters). You will not be able to show any differencies, unless you go for some etreme examples, like vdH The First carbon interconnects, which have much greater resistance than your regular metal cable, or a cables with unussually high capacitance etc.

Try two short ICs of similar design (say copper and silver made AQ cables - you can find ones that will only differ in conductor material used). They will sound distinctly different. But on the preamp out, with the standard set of measurements, will measure exactly the same.
We can't use the rules of one universe and compare a situation in another universe. I am sure you know what happened last they tried that in Ghostbusters by crossing the beams!

The fact that measurements don't show a change is objective data that can't be questioned (not easily anyway). The fact that we think we hear a difference is questioned. We can fix that and once we do, the audible differences go away making the listening and measurements in sync. Again, I am using the rules of one universe, i.e. how we conduct our working professional engineering/research to show consistency there. We can't use unacceptable methods such as ad-hoc sighted listening to challenge work that complies with that domain.
 
Apr 3, 2010
16,022
0
0
Seattle, WA
I have many years of experience in the hobby and do understand that professional engineering and testing has an important place. But it is not the final arbiter, in the end it is the sound that matters...
It is and hence the reason I have been measuring what comes out of the DAC, i.e. the final sound that matters as you say. In that regard, one of these products did nothing, the other showed a bit of degradation. BTW, years of being a hobbyist and five bucks will just get you a cup of coffee :D. I have been in the hobby for 40 years too but rely on little of that to analyze what goes on in these systems. I don't wake up thinking I am a doctor because I have been a patient all my life.

Amateur testing and pseudoscience leave allot to be desired IMO. And the people that often latch onto it with a death grip, leave even more...
Sure. Hence the reason you are in a thread where engineers who understand the system have run tests with industry standard practices and equipment costing $25,000. Let me know how many Amateurs you find similarly situated. Doesn't mean I am automatically right of course. But it does mean your comment is out of line.
 
Apr 3, 2010
16,022
0
0
Seattle, WA
MDAC without regen single ended output



MDAC with regen single ended output........mmmmmmm..........

Hmmm indeed :). Welcome to the forum. Good to see another set of measurements with different DAC showing the type of degradation I found. Interesting that balanced output did not show it. So it must be some kind of common mode noise???
 
May 30, 2010
13,912
5
38
Portugal
Amir,

Time for a technical question.

The picture shown in the Audio Precision page "Audio.TST September 2011: Measuring Jitter with APx, Controlling APx with Python" for a DAC with 51 ps jitter shows a spectra with a noise floor 20 dB bellow the similar measurement you have shown. Can we expect that in such a system and DAC we would see differences that were unnoticed in the test you carried?

Since you took the time to answer Adam question I feel I can also add that Van den Hul reported about 20 years ago that in order to measure differences and correlate sound quality with cables measurements must be carried at the -140 dB level.
 

Attachments

jkeny

Member Sponsor
Feb 10, 2012
3,427
0
0
Ireland
Hmmm indeed :). Welcome to the forum. Good to see another set of measurements with different DAC showing the type of degradation I found. Interesting that balanced output did not show it. So it must be some kind of common mode noise???
ArnyK, on another forum, has reported the suppression of this 8KHz USB protocol noise when using balanced analogue output of a DAC.

Amir, how are you measuring the output of the analogue DAC outputs?
 

Tony Lauck

New Member
Aug 20, 2014
183
0
0
It would be helpful if all FFT plots came with documentation of how the FFT is calculated. This would include, as a minimum, the number of FFT bins and the window used. Without this information, it is impossible to estimate the "FFT gain" and without this gain information one can mistakenly compare apples to oranges. Other essential information includes the sampling rate and the bit depth. For example, a 44/24 silent track, converted to 44/16 using iZotope dither (set for TPDF) has an FFT noise bins at -137 dBfs using Sony Soundforge with 65K FFT with Blackman Window. Changing to 16K FFT it's more like -131 dBfs. changing to a rectangular window has a smaller effect.

One other thing to worry about is how the FFT bins line up with the signals (or spikes). For example, displaying a 1000 Hz sine wave that peaks at -12 dBfs with the Soundforge 65K Blackman FFT the spectrum shows a peak at -45 dBfs, but if the frequency is shifted to 1004 Hz the peak appears at -12 dBfs, as expected. One needs to know the exact sampling rate of the analyzer to understand if these effects might appear in measurements if one wants to be sure of accurate results when looking at noise spikes (e.g. 8 and 16 kHz spikes that have been discussed). This is an example of how a seemingly irrelevant change in test conditions can create a large and misleading difference in measured results.
 

Mitchco

New Member
Dec 5, 2011
28
0
0
re: eye patterns. Back in the 80's I worked on an engineering team that patented a digital dataset (modem) http://www.google.com/patents/US4677646 where we used eye patterns to tune the data transmission length. i.e. how far the dataset could transmit digital signals over copper wires before the eye pattern collapses and transmission stops due to too many transmission errors. Technically called Bit Error Rate.

I have uploaded a few technical documents, for folks that are interested, to better understand what an eye pattern is and what it is used for. For example, have a look on page 16 of the document on “Understanding Eye Pattern Measurements.pdf”. Notice there is what is called a “mask compliance testing” for an eye pattern. It is a named standard. The idea is that in the transmission of digital data, if your eye pattern stays out of the masked area, you have 100% guaranteed bit for bit digital transmission of the data. It’s that simple.

When I was tuning eye patterns, it was to ensure we could get the longest transmission of digital data before running into errors that would overflow the error correction circuitry and then simply stop transmitting. If you look at the “HDMI Eye Audioquest.pdf” document, you can see on page 5 the mask compliance test and how it relates to the test Bit Error Rate.

I have also included “Understanding Data Eye.pdf” (see page 5 mask compliance) and “Mask Testing Agilent.pdf” (see Mask Testing on Page 8). If one does the research, one will find 100’s of technical documents, all saying the same thing related to eye patterns.

My point in posting this is to shed some light on eye pattern testing for those interested. In the case of the mask compliance test, it is a simple pass/fail. If the eye pattern stays out of the masked area, then we have 100% digital transmission with no data errors. From a audio sound quality perspective, there is no correlation to an eye pattern other than if the bit error rate is so bad, we hear that as drop outs when listening to music.

How would I test a Regen or similar device? It's all about what is audible and what is not. I would record the analog output of the DAC, once with the device under test (DUT) in the circuit and once without. Then ABX the two recordings and listen for an audible difference at normal listening levels. The key being an audible difference. When I wrote this article on Fun With Digital Audio – Bit Perfect Audibility Testing the best I could do when comparing one recording with artifacts to the original recording at normal listening levels was around -70 dBFS.

If I performed a simple null or difference test with the two recordings (i.e. one with the DUT, one without) and if the measured difference was -80 dBFS for example, in the course of regular listening levels, it is unlikely that I would be able to hear the difference as the level is below my (tested) audibility threshold. Or put another way, the artifact in the recording is masked by main program level of the music. Look up psychoacoustics and auditory masking to learn more. One can also download the music and difference files from my article linked above to test one's own audibility/auditory threshold. Or take this simple test to get a feel for what auditory masking sounds like: Masking a tone by noise.

Enjoy the music.
 

dallasjustice

Member Sponsor
Apr 12, 2011
2,090
0
0
Dallas, Texas
re: eye patterns. Back in the 80's I worked on an engineering team that patented a digital dataset (modem) http://www.google.com/patents/US4677646 where we used eye patterns to tune the data transmission length. i.e. how far the dataset could transmit digital signals over copper wires before the eye pattern collapses and transmission stops due to too many transmission errors. Technically called Bit Error Rate.

I have uploaded a few technical documents, for folks that are interested, to better understand what an eye pattern is and what it is used for. For example, have a look on page 16 of the document on “Understanding Eye Pattern Measurements.pdf”. Notice there is what is called a “mask compliance testing” for an eye pattern. It is a named standard. The idea is that in the transmission of digital data, if your eye pattern stays out of the masked area, you have 100% guaranteed bit for bit digital transmission of the data. It’s that simple.

When I was tuning eye patterns, it was to ensure we could get the longest transmission of digital data before running into errors that would overflow the error correction circuitry and then simply stop transmitting. If you look at the “HDMI Eye Audioquest.pdf” document, you can see on page 5 the mask compliance test and how it relates to the test Bit Error Rate.

I have also included “Understanding Data Eye.pdf” (see page 5 mask compliance) and “Mask Testing Agilent.pdf” (see Mask Testing on Page 8). If one does the research, one will find 100’s of technical documents, all saying the same thing related to eye patterns.

My point in posting this is to shed some light on eye pattern testing for those interested. In the case of the mask compliance test, it is a simple pass/fail. If the eye pattern stays out of the masked area, then we have 100% digital transmission with no data errors. From a audio sound quality perspective, there is no correlation to an eye pattern other than if the bit error rate is so bad, we hear that as drop outs when listening to music.

How would I test a Regen or similar device? It's all about what is audible and what is not. I would record the analog output of the DAC, once with the device under test (DUT) in the circuit and once without. Then ABX the two recordings and listen for an audible difference at normal listening levels. The key being an audible difference. When I wrote this article on Fun With Digital Audio – Bit Perfect Audibility Testing the best I could do when comparing one recording with artifacts to the original recording at normal listening levels was around -70 dBFS.

If I performed a simple null or difference test with the two recordings (i.e. one with the DUT, one without) and if the measured difference was -80 dBFS for example, in the course of regular listening levels, it is unlikely that I would be able to hear the difference as the level is below my (tested) audibility threshold. Or put another way, the artifact in the recording is masked by main program level of the music. Look up psychoacoustics and auditory masking to learn more. One can also download the music and difference files from my article linked above to test one's own audibility/auditory threshold. Or take this simple test to get a feel for what auditory masking sounds like: Masking a tone by noise.

Enjoy the music.
Hey Mitch, I am glad you posted here. I always learn so much from your comments.

Maybe you can have a go with the Regen too. Maybe when Amir is done with it, he can send my Regen (in his possession) to you so you can do some null/difference testing with your software and the Hilo. That would be interesting for sure.


Michael.
 

Elberoth

Member Sponsor
Dec 16, 2012
1,857
2
38
Poland
MDAC with regen single ended output........mmmmmmm..........

Now we are getting somewhere ! Those jitter spikes that were barely visible on Amir's measurements, can be seen clearly when the noise floor was lowered.

A Q to Amir: the high noise floor on your measurements, is the limitation of your testing rig, or the Meridian DAC ? If the later, can you measure another DAC, with lower noise floor ?

I would still like to see how those spikes look like with a quality, linear PSU.

BTW - another USB regenerator just hit the market - iFi iUSB 3.0. I have one at home right now. I have posted naked images on CA:

http://www.computeraudiophile.com/f...es-regenerator-xxx-pictures-25985/#post470425
 

Tony Lauck

New Member
Aug 20, 2014
183
0
0
Hi

I performed some further measurements on my rgen and MDAC this evening. Interesting findings. I too have found that when connected single ended the regen induces what I presume is 8kHz USB packet noise.

BTW this isnt a J test, its just at 12kHz.



MDAC with regen single ended output........mmmmmmm..........

How do the ugly spurs change when the 12 kHz test signal is altered in frequency or amplitude? What does the unbalanced analog output look like when the input signal is just dither noise with no test signal? This information might help in understanding what is going on.
 
Sep 30, 2015
3,131
0
0
Now we are getting somewhere ! Those jitter spikes that were barely visible on Amir's measurements, can be seen clearly when the noise floor was lowered.

A Q to Amir: the high noise floor on your measurements, is the limitation of your testing rig, or the Meridian DAC ? If the later, can you measure another DAC, with lower noise floor ?

I would still like to see how those spikes look like with a quality, linear PSU.

BTW - another USB regenerator just hit the market - iFi iUSB 3.0. I have one at home right now. I have posted naked images on CA:

http://www.computeraudiophile.com/f...es-regenerator-xxx-pictures-25985/#post470425
Sonore just came out with a hub similar to the IFI as well. Not sure if John Swenson had a hand in developing it too or not.

http://www.sotm.sonore.us/SOtM3.html#1
 
Apr 3, 2010
16,022
0
0
Seattle, WA
Sonore just came out with a hub similar to the IFI as well. Not sure if John Swenson had a hand in developing it too or not.

http://www.sotm.sonore.us/SOtM3.html#1
I think this category is going to explode. I would be shocked if in a year we did not have 20-30 different units if not more. They are so easy to build and seemingly such easy sale that anyone with a soldering iron and screwdriver practically will build one :).
 

Elberoth

Member Sponsor
Dec 16, 2012
1,857
2
38
Poland
Sonore just came out with a hub similar to the IFI as well. Not sure if John Swenson had a hand in developing it too or not.

http://www.sotm.sonore.us/SOtM3.html#1
They used to sell their PCI/PCIe -> USB cards for quite a while. I had both, but eventually I replaced those with the Jcat one, which I liked better.

This device was conceived as an alternative for computers that do not have PCIe slots. It seems thay decided to put it in a box and sell it as stand alone device.
 
Sep 30, 2015
3,131
0
0
They used to sell their PCI/PCIe -> USB cards for quite a while. I had both, but eventually I replaced those with the Jcat one, which I liked better.

This device was conceived as an alternative for computers that do not have PCIe slots. It seems thay decided to put it in a box and sell it as stand alone device.
There's actually 2 new versions. Internal and external.
 
Apr 3, 2010
16,022
0
0
Seattle, WA
They used to sell their PCI/PCIe -> USB cards for quite a while. I had both, but eventually I replaced those with the Jcat one, which I liked better.
Would you like to loan me your Jcat to see how much it improves the performance relative to built-in USB port? Was going to buy one just now but fell off my chair when I saw the 400 Euro price!!!
 

Elberoth

Member Sponsor
Dec 16, 2012
1,857
2
38
Poland
BTW - I found a post by John Westlake, where he describes the computer used druring Regen testing:

"I've been "getting to grips" with the USB interface and Regen - as the Regen will be a competitors product I'll have to be careful so as not to appear critical - I'll have to claim the 5th amendment if some question might force me into being too harsh about its design.

As a USB source I'm using a small Sony Pocket PC (VGN-UX380) and the USB connected via its docking station - running WinXP SP3.

Early days, so I'm just performing simple tests - but I can already see on a fast Analogue scope that the USB Packet it jitter visually effected by what the PC is processing - with eye pattern Jitter increasing from 320pS to 400ps simply by running a Youtube video.

The Screen is a mess and I'm just getting the feel of whats going on - while its easy to see the jitter modulation on the Analogue scope screen, I'm not sure how to post it here.... the screen is a mess....

You can see via the jitter modulation on the scope when the PC is performing other tasks - with big "apparently" random jumps in jitter, I suspect that these jumps are caused by other "hidden" operations being performed by the Windows OS.

I'm not in anyway endorsing the idea behind "optimized" media players (as the are trying to resolve an issue that should not effect the analogue domain in the first instance - IMO they are an indirect poor and dubious bodge fix) - but seeing the OS / CPU "Process" modulation comes as no surprise and opens a whole bag of hurt...

This observations are with the Upstream USB input to the Regen, it will be interesting to see if the USB hub is able to reformat the data packets and thus attenuate the OS / CPU process related jitter modulation."
 
Sep 30, 2015
3,131
0
0
I think this category is going to explode. I would be shocked if in a year we did not have 20-30 different units if not more. They are so easy to build and seemingly such easy sale that anyone with a soldering iron and screwdriver practically will build one :).
Yeah it's the fad of the season. Personally I think much lower jitter Ethernet to I2S/DSD internal renderer boards which are slaved to the DAC master clock are going to replace the need for USB. Especially now that Ravenna has made the Ethernet renderer in DAC's a plug and play sound device, just like USB, so we are no longer bound with DLNA/UpNP limitations.