I've performed some Measurements on SPL Volume2 (Model 2602) volume control

[SIY]

How ?

I haven't tried it yet, but if it can be done, do a cross correlation to determine the delay (it's where the cross-correlation function peaks), then put the delay into the sweep analysis. I'll have to play around with it to see if the latter is possible. The AP does it automatically in the Transfer Function mode.

I've been having an email exchange with the guy who wrote the Virtins software, so I'll ask him as well.

edit: it looks like the method of delay removal is explained in section 2.6.2.8.4 of the manual.

 

[LTig]

I've found some articles for crest factor optimization methods (by computing phases),

Can you post links?

 

[LTig]

By the way, googling a bit, I found this, from our local university
http://www.montefiore.ulg.ac.be/~stan/ArticleJAES.pdf
Not directly related, though.

I know the MLS method for measuring the anechoic FR of speakers in a room. I have working (Linux) code for creating the test signals and processing the results using Fast Hadamard Transformation - an old project done with 2 other audio guys. I see no reason why this method would not work with FFR of audio electronics. Actually it should work better than with speakers due to much lower distortion in the DUT.

 

[Rja4000]

Can you post links?

https://www.researchgate.net/public...synthesize_multisines_with_arbitrary_spectrum

 

[Rja4000]

https://www.researchgate.net/public...synthesize_multisines_with_arbitrary_spectrum

Now we "just" need to apply this iteration method in a programming language, to optimize Crest Factor for 31 pseudo-log-equidistant frequencies... that are aligned with FFT bins, for several sampling rates and FFT window size.
But mostly for 192kHz rate and 512k or 1024k samples.
Well, that doesn't help for the FFR, I'm afraid.
Or maybe it does.

 

[MC_RME]

The multisine method is not about delay, but frequency accuracy. Every bin must fit perfectly.

 

[Rja4000]

The multisine method is not about delay, but frequency accuracy. Every bin must fit perfectly.

I have to admit I didn't get the link with delay.
The need for the frequency to fit the FFT bin for better FFT result accuracy, I do understand.

 

[MC_RME]

I played around with the AP Flex software a bit. Amir, WolfX-700 and SIY might be interested to check the below using their APx555.

AP's method of measuring THD is - these days- quite limited. Instead of simply taking the harmonics and calculating the respective THD values from within a high resolution FFT, like nearly everyone else does, and thus can offer THD measurements down to -160 dB easily, their THD measurement is limited to a lower FFT size, so can't separate noise and harmonics at levels that - again these days - even 200$ DACs reach easily (as proven by this forum).

Looking at the so called Distortion Product Ratio measurement tells it all: a stable 1st harmonic, and if below -120 dB heavily fluctuating harmonics 2 to x. Noise dominates them. So while the APx is famous for its analog high precision generator with about -150 dB THD, the APx itself is not able to measure that.

Which brings me back to my above post - I wanted to see if it is possible to add some workarounds so one can better separate THD and THD+N sweeps, to get a better impression what part is noise and what part is THD. That is indeed easily possible by reducing the captured noise.

Step 1: Using a 1 kHz sine, limit the Analyzer Bandwidth using elliptic high and low pass from 950 Hz to 10 kHz. This includes 10 harmonics, but removes a lot of the noise energy that taints the THD value.

Step 2: Set the Analyzer Settings (the detector) to Average mode instead of Flat. Average reduces noise, but keeps all steady tones on the same level as Flat does.

Pic 1 shows that these changes have no influence on normal, steady state signals and their levels. Note that red (Det AVG) includes the BW Limit.

As with the following pics this is a simple -60 dBFS to 0 dBFS sweep in 31 linear steps. The next pic shows how the ratio curve of THD is lowered by both changes by more than 10 dB. At levels of -10 dBFS THD takes over and they start to merge.

Another way to look at it is THD Level instead of Ratio. This one (IMHO) better visualizes that noise is the dominant and constant factor for most of the graph. It still shows the improvement by bandwidth limitation plus detector averaging to include less noise.

I exported the red curve from the THD Ratio measurement and imported it into the standard THD+N measurement of this DUT (Device Under Test).

While this measurement would still have the OP ask why all these graphs show less distortion at higher levels, I wonder if a measurement with such settings isn't a useful addition to the existing graphs, showing the influence of the distortion more clearly. I also expect this curve to be more low/detailed when not using a simple DUT in loopback as done here but the APx as source and analyzer.

And if you ever asked what we do on Christmas...here is your answer...

 

[Blumlein 88]

I have to admit I didn't get the link with delay.
The need for the frequency to fit the FFT bin for better FFT result accuracy, I do understand.

This reminds me of something I've run into before especially on J-test results. If you use 512K or 1024K FFT's the bins are so small, that clock drift between DAC and ADC can smear the results. If two clocks are pretty close you get nice results, and if they differ by say 20 ppm, they can look much messier due to no longer falling neatly in the very small FFT bins.

 

[MC_RME]

I often measure with HpW in non-window mode (which requires a synced and coherent system). This starts to show issues with ESS DACs due to their ASRC implementation. Not a big deal, but one notices the difference to none ASRC DACs.

 

[MC_RME]

I just noticed that post #48 was posted in the wrong thread. It was intended to be here:
https://www.audiosciencereview.com/...owest-on-high-volume.10464/page-2#post-291517

Maybe mods can shift it to there?

 

[LTig]

I played around with the AP Flex software a bit. Amir, WolfX-700 and SIY might be interested to check the below using their APx555. [..]
And if you ever asked what we do on Christmas...here is your answer...

Love what you do - do what you love. I earn my living writing control software for NMR spectrometers, and here I'm sitting in my well earned Christmas vacation and write software to create FFR test signals ...

 

[LTig]

I wrote a program for creating test files in WAV format (Linux only for the moment). I had a short look into the link and think I can add support for such a signal and provide them for you. Have to find out how to dermine the phase of the individual sinus signals.

Just added support of FFR signals. I used the Schröder formular to calculate the phases and it works nice - crest factor (CF) is always less than 2, so not a problem and IMV it's not necessary to use a very time consuming search algorithm to make CF a little lower.

Actually it's already very time consuming even on my Ryzen 7 1700X (8 cores + hyper threading) to create the FFR test signal for a useful number of tones so I had to implement multi threading to calculate 16 (on my PC) individual tones in parallel (makes it about 5 times faster).

Here is a test signal with 192 kHz samplerate and 24 bits for 64k FFT size, 2 seconds long, just dropped into REWs RTA window. It took 1m 42s to create. The level of the individual tones is at -50 dBFS. CF is 1.735.

EDIT BEGIN:
The individual tones are at a level of -49.6 dBFS each.

• Using the Schröder formular actually creates a sweep as one can see when loading the WAV file in Audacity (and hear when playing it):
  
  
  
  This can be programmed much easier though.
  
  
• Using a constant phase of 0 degree creates an impulse with a very high CF of 173.
  
  
  The individual tones are at a level of -89.2 dBFS each:
  
  
  
  This is unusable.
  
  
• Using a random phase creates a signal which looks and sounds like white noise with a CF of 4.6:
  
  
  
  The individual tones are at a level of -57.9 dBFS each:
  
  
  

I may have to investigate more time into the phase algorithm .
EDIT END

@Rja4000 : my algorithm currently creates all frequencies up to 91% of the nyquist frequency (to ensure that at 44.1 kHz samplerate the highest tone is just a tiny bit over 20 kHz). You wanted to have WAV files with more than 96 kHz samplerate. Tell me more:

• samplerate
• bit depth (I recommend 24)
• FFT size
• length of WAV file (should be long enough for at least on FFT size)
• highest frequency (91% of Nyquist or less?)

 

[jae]

What are some similar high quality volume controls like this (preferably fully balanced)?

 

[Rja4000]

What are some similar high quality volume controls like this (preferably fully balanced)?

I also own a passive Radial Engineering SAT-2.
It has some benefits... and drawback.
I'll post a comparisin, if you're interested.

 

[jae]

I also own a passive Radial Engineering SAT-2.
It has some benefits... and drawback.
I'll post a comparisin, if you're interested.

I'm more so looking for things along the line of the SPL product, products with a large knob/ stepped attenuator that measure well. I'm not too fond of the feel and small knobs on prducts like the RME ADI or topping 90's

 

[Rja4000]

Just added support of FFR signals. I used the Schröder formular to calculate the phases and it works nice - crest factor (CF) is always less than 2, so not a problem and IMV it's not necessary to use a very time consuming search algorithm to make CF a little lower.

Actually it's already very time consuming even on my Ryzen 7 1700X (8 cores + hyper threading) to create the FFR test signal for a useful number of tones so I had to implement multi threading to calculate 16 (on my PC) individual tones in parallel (makes it about 5 times faster).

Here is a test signal with 192 kHz samplerate and 24 bits for 64k FFT size, 2 seconds long, just dropped into REWs RTA window. It took 1m 42s to create. The level of the individual tones is at -50 dBFS. CF is 1.735.
View attachment 43826
EDIT BEGIN:
The individual tones are at a level of -49.6 dBFS each.

• Using the Schröder formular actually creates a sweep as one can see when loading the WAV file in Audacity (and hear when playing it):
  
  View attachment 43837This can be programmed much easier though.
  
  
• Using a constant phase of 0 degree creates an impulse with a very high CF of 173.
  View attachment 43838The individual tones are at a level of -89.2 dBFS each:
  
  View attachment 43839This is unusable.
  
  
• Using a random phase creates a signal which looks and sounds like white noise with a CF of 4.6:
  
  View attachment 43841The individual tones are at a level of -57.9 dBFS each:
  
  View attachment 43842

I may have to investigate more time into the phase algorithm .
EDIT END

@Rja4000 : my algorithm currently creates all frequencies up to 91% of the nyquist frequency (to ensure that at 44.1 kHz samplerate the highest tone is just a tiny bit over 20 kHz). You wanted to have WAV files with more than 96 kHz samplerate. Tell me more:

• samplerate
• bit depth (I recommend 24)
• FFT size
• length of WAV file (should be long enough for at least on FFT size)
• highest frequency (91% of Nyquist or less?)

Re-reading this, I'm not sure I answered.
Sample rate 768kHz
24 bits
FFT what do you think ?
Length: enough
91% of Nyquist

But your 192kHz would do nicely already

 

[LTig]

Re-reading this, I'm not sure I answered.

You didn't, and I forgot totally about it.

Sample rate 768kHz
24 bits
FFT what do you think ?
Length: enough
91% of Nyquist

But your 192kHz would do nicely already

The 192 kHz 24 Bits 64k FFT file with 2 seconds runtime is attached as zip file containing a flac (1 MB).

768 kHz with enough time resolution may need a veeerrrryyy looooooonnnnnnng time to calculate. Let me try it later. I could send you those now via wetransfer.com. I just need a valid email address (per private conversation).

 

[Rja4000]

768 kHz with enough time resolution may need a veeerrrryyy looooooonnnnnnng time to calculate.

I guessed so

 

[Rja4000]

The 192 kHz 24 Bits 64k FFT file with 2 seconds runtime is attached as zip file containing a flac (1 MB).

Thanks !