Beta Test: Multitone Loopback Analyzer software

[DonH56]

Hi Don,

You might be surprised to find that AES has made their own version of many of these standards that don't always agree with IEEE AP has mostly followed AES standards, from what I can tell from their website.

Hey Paul,

Not surprised, maybe a little disappointed, since I was involved with the IEEE versions (and had to adhere to them when I was designing things).

Audio has almost always been a mess of conflicting standards (remember IHF? And don't get me started on HDMI...) I would have hoped basics like INL and DNL were the same but do not know. I have never really done audio testing, not since my college days working as a tech and a little hobby stuff after, so no dog in this hunt.

Life goes on - Don

 

[pkane]

Hey Paul,

Not surprised, maybe a little disappointed, since I was involved with the IEEE versions (and had to adhere to them when I was designing things).

Audio has almost always been a mess of conflicting standards (remember IHF? And don't get me started on HDMI...) I would have hoped basics like INL and DNL were the same but do not know. I have never really done audio testing, not since my college days working as a tech and a little hobby stuff after, so no dog in this hunt.

Life goes on - Don

Yes, audio is a mess. I don't believe there's an AES standard for INL or DNL, but I'm not a member, so can't check

 

[DonH56]

Yes, audio is a mess. I don't believe there's an AES standard for INL or DNL, but I'm not a member, so can't check

I left the AES long, long ago but have actually been thinking of rejoining for "fun". I have been an IEEE member over 40 years now but lately am wondering "why?"

 

[Rantapossu]

Does the "Peak level" measurement show the left channel values for both channels when doing level and frequency sweeps?

Using:

"Peak level" level sweep:

"RMS level" seems to be ok:

"Peak level" frequency sweep, same problem:

 

[Rantapossu]

And are these values missing from the measurement results?
- SinAD
- Peak level

If Peak level is missing and you add it, should the current "Level" be "RMS Level"?

 

[Sokel]

Does the "Peak level" measurement show the left channel values for both channels when doing level and frequency sweeps?

Using:
View attachment 229542

"Peak level" level sweep:

View attachment 229548

"RMS level" seems to be ok:

View attachment 229549


"Peak level" frequency sweep, same problem:

View attachment 229552

Tried it too and I got an "System.IndexOutOfRangeException" error.

 

[Sokel]

Restarted and I don't get the message but I get same Peak values for the connected and the sorted channel.
Which is right,isn't it,peak values are for the recording.

 

[pkane]

And are these values missing from the measurement results?
- SinAD
- Peak level

If Peak level is missing and you add it, should the current "Level" be "RMS Level"?

View attachment 229558

Peak and rms values for Level are already there. I can show you again how to expand the column, if you want

 

[Rantapossu]

Peak and rms values for Level are already there. I can show you again how to expand the column, if you want

View attachment 229573

I promise to practice it!

Roger with the "Level"!

But no matter how much I expand it, "SinAD" doesn't show up...

 

[Blumlein 88]

Okay, have a question about displayed levels. Asked about it once before without getting an answer.

Here is a short level sweep showing RMS and Peak level. As expected there is a 3 db difference.

When I switch to the spectrum view I'd expect that last reading which is a 0dbFS signal to be -3.48 or so db if this shows RMS or -.48 db if this is peak. The .48 is because my hardware puts out a 0 db level at -.48 db into my ADC. Instead it shows about -6.5 db or presumably -6 db plus the .48 db due to my hardware. So it isn't peak and it isn't RMS what and why is this?

And if I just run a spectrum of a 0 dbFS tone I get the same result. My interest is obviously doing low level signals for linearity or other low levels purposes. I cannot read the signal level off the graph of the spectrum and be sure what it is. Of course I'm also curious if this is effecting the distortion and other results by a 3 db or so difference that is not accurate.

 

[Sokel]

Okay, have a question about displayed levels. Asked about it once before without getting an answer.

Here is a short level sweep showing RMS and Peak level. As expected there is a 3 db difference.
View attachment 229613

When I switch to the spectrum view I'd expect that last reading which is a 0dbFS signal to be -3.48 or so db if this shows RMS or -.48 db if this is peak. The .48 is because my hardware puts out a 0 db level at -.48 db into my ADC. Instead it shows about -6.5 db or presumably -6 db plus the .48 db due to my hardware. So it isn't peak and it isn't RMS what and why is this?

View attachment 229616

And if I just run a spectrum of a 0 dbFS tone I get the same result. My interest is obviously doing low level signals for linearity or other low levels purposes. I cannot read the signal level off the graph of the spectrum and be sure what it is. Of course I'm also curious if this is effecting the distortion and other results by a 3 db or so difference that is not accurate.

To me is -1 and -4 respectively and when I switch to spectrum is -7 which is the value of the fundamental.

 

[Blumlein 88]

To me is -1 and -4 respectively and when I switch to spectrum is -7 which is the value of the fundamental.

But why is the fundamental different from peak by 6 db? Normally I expect FFT spectrums to show peak. Either way a 6 db difference doesn't match up with anything unless I'm mistaken about how this works. When you look at the crest factor in MT it shows 1.41 for sine waves as it should. This should be a 3.01 db difference in peak and RMS.

Here I have the cursor on a 0 dbFS tone in Audacity and it shows 0.0 db while the RMS pane shows -3.01 as expected.

 

[Sokel]

But why is the fundamental different from peak by 6 db? Normally I expect FFT spectrums to show peak. Either way a 6 db difference doesn't match up with anything unless I'm mistaken about how this works. When you look at the crest factor in MT it shows 1.41 for sine waves as it should. This should be a 3.01 db difference in peak and RMS.

Here I have the cursor on a 0 dbFS tone in Audacity and it shows 0.0 db while the RMS pane shows -3.01 as expected.

View attachment 229623

View attachment 229624

You get 0db in spectrum if you measure in dbr scale though.
That's relative to level,right?

 

[Blumlein 88]

You get 0db in spectrum if you measure in dbr scale though.
That's relative to level,right?

Yes, but for instance at low levels that doesn't help me because it is arbitrary and not a measure of a known level.

 

[Sokel]

Like this:

level

 

[Sokel]

Yes, but for instance at low levels that doesn't help me because it is arbitrary and not a measure of a known level.

You have a point there...

 

[Blumlein 88]

I knew I had seen this somewhere. Finally remembered where. So in the results tab, you see the level in my case is -.5 dbFS (-3.5 dbFS RMS). Then the fundamental is listed as -6.5 db FS. I'm wondering why that is. Maybe I've forgotten something somewhere. I would think the fundamental should be at -.5 dbFS (with my hardware). If I'm forgetting something obvious thru ignorance just tell me.

 

[Blumlein 88]

I Guess the fundamental value can be used in the results page I just have to adjust for the extra 6 db and level going into my ADC. Here is the 23rd bit, which is -138.46 db + (-.48 db) + (-6.02 db)= -144.96 as my expected result for the 23rd bit by itself. I get pretty close at -144.8 dbFS.

 

[Blumlein 88]

Using 16 averages is a bit further from the target, but not surprising as noise has to be a problem at that level.

A single average is close, but doing it multiple runs it varies a couple tenths of a db either way.

 

[Sokel]

I Guess the fundamental value can be used in the results page I just have to adjust for the extra 6 db and level going into my ADC. Here is the 23rd bit, which is -138.46 db + (-.48 db) + (-6.02 db)= -144.96 as my expected result for the 23rd bit by itself. I get pretty close at -144.8 dbFS.

View attachment 229632

I get 146.7 which is -.2db.Close enough.