Beta Test: Multitone Loopback Analyzer software

[Moto]

I used your suggestion and even went 5-24khz in multitone. Snr only went down from 124.8 to 124.6. In REW it is 122.5

 

[pkane]

I used your suggestion and even went 5-24khz in multitone. Snr only went down from 124.8 to 124.6. In REW it is 122.5

I don't know what REW is doing to compute SNR, so can't really tell why the two are different.

 

[Moto]

In an old comment on REW forum before there was a separate snr number in REW, John told a user to just reverse sign on thd+n to get snr.

John’s post.

For an SNR figure in the presence of a signal just negate the THD+N figure. The figure on the distortion panel next to the fundamental frequency is the level of the fundamental, when the fundamental is very low (or just a noise spike as in the image) it will be significantly lower than the total RMS signal level shown in the levels panel.

 

[Moto]

@pkane so snr is computed in multitone using all noise and all harmonics in selected measurement range, not just max 15 harmonics plus noise?
It also does just the measurement range, not nyquist/2?

 

[pkane]

@pkane so snr is computed in multitone using all noise and all harmonics in selected measurement range, not just max 15 harmonics plus noise?
It also does just the measurement range, not nyquist/2?

Correct. The maximum # of harmonics determines what's shown on the display, not used in the computation. Noise is measured across the measurement frequency range. If you want Fs/2 then you'll need to specify it as the upper bound of the measurement bandwidth.

 

[Moto]

I’m sorry to be dense, but in Multitone, why do I see e.g., snr=124.8 and thd+n=-122.8 when all harmonics and noise are included in the snr? Why isn’t snr just the negation of thd+n?

In the latest beta of REW, snr=-N which is a few 10ths of a db better than thd+n. That seems to violate john’s earlier post on snr being the negation of thd+n.

I’m confused.

 

[JohnPM]

From the REW help, emphasis added:

When the Distortion Panel button (keyboard shortcut Alt+D) is selected the analyser calculates and displays harmonic or intermodulation distortion figures for the input, including THD, N+D (A-weighted noise plus distortion), N (noise and non-harmonic distortion), THD+N, SNR (where N excludes distortion), ENOB, HHD (higher harmonic distortion for harmonics from the 10th up to at most the 50th) and the relative levels of the 2nd to 9th harmonics.

 

[Moto]

Thx @JohnPM . So snr=-N always in REW in dbc mode?
It seems like snr in REW should always be better in REW than in Multitone since Multitone includes all harmonics in N. The opposite is true though. Multitone always shows higher snr than REW.

 

[Moto]

@pkane does it make sense that snr in Multitone would be better than snr in REW given john’s clarification of REW’s computation of snr?

 

[pkane]

@pkane does it make sense that snr in Multitone would be better than snr in REW given john’s clarification of REW’s computation of snr?

I can't say, @Moto. Multitone does a fairly straightforward calculation after removing the fundamental and harmonics from the result and then summing the remaining energy for the specified bandwidth. I don't know what REW does.

 

[Moto]

I can't say, @Moto. Multitone does a fairly straightforward calculation after removing the fundamental and harmonics from the result and then summing the remaining energy for the specified bandwidth. I don't know what REW does.

Maybe I misunderstood you earlier. I thought you said in the snr calculation that you used noise plus harmonics in the “noise” part of your computation.

 

[pkane]

Maybe I misunderstood you earlier. I thought you said in the snr calculation that you used noise plus harmonics in the “noise” part of your computation.

No, sorry if that wasn’t clear. Noise calculation excludes harmonics, and, of course, the fundamental.

 

[Moto]

No, sorry if that wasn’t clear. Noise calculation excludes harmonics, and, of course, the fundamental.

Interesting. Then REW and Multitone should agree on snr. Odd that snr in Multitone is about 2db better than REW. Oh well.

 

[pkane]

Interesting. Then REW and Multitone should agree on snr. Odd that snr in Multitone is about 2db better than REW. Oh well.

On another note in 1.0.83 whenever I set the fft when just doing 1000hz sine measurement, it reverts to 256k. What would cause it ti do that?

High sample rate will increase FFT size automatically.

 

[Moto]

High sample rate will increase FFT size automatically.

I had just realized that and edited the post to delete that but you beat me to it with an answer!

 

[pkane]

New Multitone version 1.0.84 is available with just one addition: a Spectrogram viewer.

Spectrogram viewer is a new tab that shows the recorded audio clip on time vs frequency axes with color coding to indicate signal level.

Multiple color palettes are selectable providing somewhat different contrast display to highlight different details in the data. FFT Size (under Chart units) determines the vertical frequency resolution. Larger FFT size will result in a longer computation. The scale on the right provides the mapping of colors to their respective level in the plot, in dBFS.

 

[Grooved]

New Multitone version 1.0.84 is available

Hi @pkane , thanks for this new version

I still had the same problem with one computer and Multitone being locked on dBr, but after deleting the MultitoneDefault.dw file, it seems to be OK now.

 

[Sokel]

Is there a chance we can have a "no PC shutdown" option during measurements?
I keep forgetting it!

 

[pkane]

Is there a chance we can have a "no PC shutdown" option during measurements?
I keep forgetting it!

It's already available: use a better ASIO driver. As an option, you can also buy a new PC

 

[Sokel]

It's already available: use a better ASIO driver. As an option, you can also buy a new PC

No,that is another matter,I'm talking about the energy saving schemes,some like foobar prevent that (as an option).