• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Beta Test: Multitone Loopback Analyzer software

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,524
Likes
37,057
Here is a result with a Topping D10 balanced feeding the Babyface pro.
1636261422953.png
 

Grooved

Addicted to Fun and Learning
Joined
Feb 26, 2021
Messages
679
Likes
441
Here is a result with a Topping D10 balanced feeding the Babyface pro.
View attachment 163738
So you're getting a little bit better result with the Babyface loopback.

I'm note sure if it's as much important in this test than in a loopback difference checked with DeltaWave, but did you sync both devices with the SPDIF signal (SPDIF optical out from D10b to SPDIF optical in of the Babyface, to lock the Babyface on the D10b) ?

Done or not, could you please compare both cases?
 

Grooved

Addicted to Fun and Learning
Joined
Feb 26, 2021
Messages
679
Likes
441
And here are the first two (IMHO obvious) wishes:

1. Add buttons to show only blue or white (so one can toggle the view between all three states)
2. Add Auto dBr, so that the highest peak is set to 0 dBr and the scale is also auto adjusted. Makes it much easier to get the values (ratio) of noise floor and harmonics from the screen.
...
Would it just push the graph up with the peak being on 0 but getting a wrong result on noise ?
I just checked with one device and was getting the same noise level for a 0, -10 and -20 dB signal level, only tone and harmonics were lower, so if I take the -20dB signal result and push it up to align it with 0dB, the noise would end up 20dB higher while in realty, it was on the same level than when I sent a 0dB signal...
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,632
Likes
10,205
Location
North-East
So you're getting a little bit better result with the Babyface loopback.

I'm note sure if it's as much important in this test than in a loopback difference checked with DeltaWave, but did you sync both devices with the SPDIF signal (SPDIF optical out from D10b to SPDIF optical in of the Babyface, to lock the Babyface on the D10b) ?

Done or not, could you please compare both cases?

There’s a 16ppm clock drift reported in that recording, so the clocks were not synced.
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,632
Likes
10,205
Location
North-East
Would it just push the graph up with the peak being on 0 but getting a wrong result on noise ?
I just checked with one device and was getting the same noise level for a 0, -10 and -20 dB signal level, only tone and harmonics were lower, so if I take the -20dB signal result and push it up to align it with 0dB, the noise would end up 20dB higher while in realty, it was on the same level than when I sent a 0dB signal...

dBr is in units relative to a specific signal. In the case of Multitone, it's relative to the largest amplitude of the signal. Noise is also relative to this watermark. In other words, noise level is not expressed as the absolute level that you might see in dBFS measurements, but relative to the maximum amplitude.
 

Grooved

Addicted to Fun and Learning
Joined
Feb 26, 2021
Messages
679
Likes
441
There’s a 16ppm clock drift reported in that recording, so the clocks were not synced.
Did no check the drift value, no sync indeed. Would be curious to know if there's a difference with another test, synced this time
dBr is in units relative to a specific signal. In the case of Multitone, it's relative to the largest amplitude of the signal. Noise is also relative to this watermark. In other words, noise level is not expressed as the absolute level that you might see in dBFS measurements, but relative to the maximum amplitude.
Yes, but the tool was already adjusting the result, if not the graph, am I wrong ? I did not spend time checking that but it was the impression I got.
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,632
Likes
10,205
Location
North-East
Did no check the drift value, no sync indeed. Would be curious to know if there's a difference with another test, synced this time

Yes, but the tool was already adjusting the result, if not the graph, am I wrong ? I did not spend time checking that but it was the impression I got.

Multitone measures and compensates for clock drift, in a similar way that DeltaWave does. Without a clock sync, the ADC and the DAC will have clocks that can run at slightly different rates. This can shift frequency response and introduce phase differences in the analysis, so it's important to either measure this and compensate for it, or sync the two devices.

When you select dBr option, the tool will automatically scale the result to align the maximum amplitude with 0dB. Everything else in the display will be scaled appropriately, including noise. Without the dBr option, the display is in dBFS, in other words, scaled to the largest possible sample value of 1.0.
 

Grooved

Addicted to Fun and Learning
Joined
Feb 26, 2021
Messages
679
Likes
441
Thanks @pkane ! I missed an important thing reading too much fast, it was with dBr and you have added "Spectrum in dBr" option in the last version.

Back to drift, I'm a little bit surprised that I get no drift at all between Tone2 Pro and Cosmos ADC (without any sync), just a bit of delay which is something normal in this case I guess (no dBr activated in this test) :
RTA 1kHz 0dBFS Cosmos-Tone2 cal-3.5V.jpg
 
  • Like
Reactions: trl
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,632
Likes
10,205
Location
North-East
@restorer-john :

Did you try different sampling rates with WASAPI or just ASIO? Using WASAPI, I get pretty consistent results, difference of less than 1dB between 44.1k, 48k, and 96k.

Can you please share what multitone test signal you were using, and what FFT size when you saw these large variations between sampling rates? And was ASIO used for input or output?

96k:
1636297065954.png



48kHz:
1636297189432.png


44.1k:
1636297337345.png


EDIT: adding 192k result

1636299100713.png
 
Last edited:
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,632
Likes
10,205
Location
North-East
Thanks @pkane ! I missed an important thing reading too much fast, it was with dBr and you have added "Spectrum in dBr" option in the last version.

Back to drift, I'm a little bit surprised that I get no drift at all between Tone2 Pro and Cosmos ADC (without any sync), just a bit of delay which is something normal in this case I guess (no dBr activated in this test) :
View attachment 163864

Drift calculation requires a multi-tone signal, preferably with at least 32 tones or more. You can tell if you're looking at a dBr display because it will have (dBr) in the title, like this:

1636298011962.png
 

Grooved

Addicted to Fun and Learning
Joined
Feb 26, 2021
Messages
679
Likes
441
Drift calculation requires a multi-tone signal, preferably with at least 32 tones or more. You can tell if you're looking at a dBr display because it will have (dBr) in the title, like this:

View attachment 163867
Was looking at the Sinewave instead of Multitone :facepalm: Will check it later today
Regarding the dBr option, good thing to add the setting in the title ;)
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,524
Likes
37,057
For those wondering here is the Topping D10 balanced feeding the Babyface Pro with the clocks locked via SPDIF.

1636312365649.png
 

Hee

Member
Joined
Jan 15, 2020
Messages
32
Likes
23
Thank you. Pkane
I reinstalled the window and succeeded.

'spartan cue' audiointerface, trs out -> trs in
Thank you again.
 

Attachments

  • cue_multi.jpg
    cue_multi.jpg
    359.8 KB · Views: 92

trl

Major Contributor
King of Mods
Joined
Feb 28, 2018
Messages
1,967
Likes
2,523
Location
Iasi, RO
Good work! If you want to replicate my measurements, these are the settings for multitone test:

Sample rate = 192 kHz
FFT size = 256k
Averaging = 8
Not quite the same, so I wonder if I did somthing wrong. I'd like to see the entire dynamic range dropping from near 0dBFS, if possible. @pkane pls hlp!

Multitone_32AP_D90_CosmosADC_.png
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,632
Likes
10,205
Location
North-East
Not quite the same, so I wonder if I did somthing wrong. I'd like to see the entire dynamic range dropping from near 0dBFS, if possible. @pkane pls hlp!

View attachment 164575

Try increasing FFT size, it looks like frequency resolution below 50Hz isn't sufficient. I'll have an updated version soon that will help with that, but for now, just try larger FFT size.
 

trl

Major Contributor
King of Mods
Joined
Feb 28, 2018
Messages
1,967
Likes
2,523
Location
Iasi, RO
Multitone_32AP_D90_CosmosADC_FFT_.png

Increasing FFT size does help indeed, thank you! Looking forward for the next update.
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,632
Likes
10,205
Location
North-East
New version 1.0.3 of Multitone is available. If you already had v1.0.2, just start it and you should get a prompt to auto-update to the latest.

Changes in 1.0.3

  • Added Measurement History window to compare results
  • Added dither option to settings
  • Added clear plots/results button
  • Changed multi-tone generation for better separation in lower frequencies
  • Fixed an error when enumerating audio devices that don’t report their settings
 
Top Bottom