• 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

Grooved

Addicted to Fun and Learning
Joined
Feb 26, 2021
Messages
679
Likes
441
  • Fixed an error when enumerating audio devices that don’t report their settings

Thanks @pkane , totally forgot to report this one since the first version, it was always showing the highest setting previously use, even after lowering the settings
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,626
Likes
10,202
Location
North-East
One more update bringing this up to version 1.0.4:

Changes in 1.0.4

  • Added support for using the same ASIO driver for input and output
  • Added Harmonics display to the spectrum plot when testing with a single frequency
  • Added support for TIM-type test signals consisting of a combination of a square and sine waves
  • Added two Transient Intermodulation distortion test signals to the list
  • Fixed history plots not displaying properly when only a single test tone was recorded
  • Some minor clean-up of functionality and display

The key update in this version is the ability to use the same ASIO driver for input and output. Other features were added by request:

1636838429488.png


TIM test signal:
1636838749303.png


1636838797506.png
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,521
Likes
37,050
And I haven't even gotten around to using version 3 yet. Looks like all good additions/changes.
 

MC_RME

Addicted to Fun and Learning
Technical Expert
Audio Company
Joined
May 15, 2019
Messages
854
Likes
3,564
Same here, currently travelling. 1.04 looks like a big one. Exciting!
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,626
Likes
10,202
Location
North-East
Version 1.0.5 is now available:

Changes in 1.0.5​

  • Added support for saving/removing custom test signals (+ and - buttons next to the test tone edit box)
  • Added additional metrics (frequency, THD, SNR, and ENOB) to a single-tone test result
  • Added current FFT window on the main display above the FFT size
  • Changed “Dirichlet” FFT window name to “Dirichlet (Rect)”

Added documentation on how to create custom test signals. Some examples:

MySine 123.56Hz
SMPTE 30/60/1k/7k/13k 4:4:1:1
MT_Test4 20/100/13k/13.1k/15k 1:4:1:1:1
SQT_Test3 sq:1000Hz/13kHz 4:1


The new metrics are displayed for single-tone test signals, as long as the "Display Harmonics" setting is enabled in settings:

1637349772775.png
 
Last edited:

trl

Major Contributor
King of Mods
Joined
Feb 28, 2018
Messages
1,967
Likes
2,523
Location
Iasi, RO
A bit confusing the ENOB, given the formula (source: Wikipedia):
36ec123dbd02e4ed5dafef30e90f58faecaa694a
; I was hoping to see something around 19.5...20 bits.

Also, the Amir was measuring a SINAD of 120...121 dB, not 123 dB, so maybe this E50 is a bit better?
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,626
Likes
10,202
Location
North-East
A bit confusing the ENOB, given the formula (source: Wikipedia):
36ec123dbd02e4ed5dafef30e90f58faecaa694a
; I was hoping to see something around 19.5...20 bits.

Also, the Amir was measuring a SINAD of 120...121 dB, not 123 dB, so maybe this E50 is a bit better?

The ENOB calculation you're quoting is for a full-scale signal (0dBFS). Multitone makes an adjustment for a test signal that was measured below the full scale. The adjustment is the ratio between full scale and the measured signal. In the above test this is about 3.5dB, which accounts for about a 0.6 bits increase in the ENOB.
 
Last edited:
  • Like
Reactions: trl

Lambda

Major Contributor
Joined
Mar 22, 2020
Messages
1,785
Likes
1,519
@pkane is it open source?
And how are the chances to get it running on Linux?
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,626
Likes
10,202
Location
North-East
@pkane is it open source?
And how are the chances to get it running on Linux?

It's not open source. It might run on Linux with Wine, but I've not tested it this way. Some of my other software, such as DeltaWave, is usable under Linux (see here). Since Multitone uses the same libraries and has similar dependencies, there is a decent chance it might work, but no guarantees :)
 

trl

Major Contributor
King of Mods
Joined
Feb 28, 2018
Messages
1,967
Likes
2,523
Location
Iasi, RO
Multitone makes an adjustment for a test signal that was measured below the full scale. The adjustment is the ratio between full scale and the measured signal. In the above test this is about 3.5dB, which accounts for about a 0.6 bits increase in the ENOB.
Thanks for the replay, Paul. You mean -3.5dB, right?

What about the THD+N of -123.1 dB instead of the -120 dB measured by Amir? Is it related to the same adjustment as above?
 

Hee

Member
Joined
Jan 15, 2020
Messages
32
Likes
23
Hi!
Look at the device in the picture above, the sampling rate setting is different.
If the sampling rate is the same, td+n is lowered. (about 119.6td+n 2ch mode)
In a few days, I will compare this program with REW.
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,626
Likes
10,202
Location
North-East
Thanks for the replay, Paul. You mean -3.5dB, right?

What about the THD+N of -123.1 dB instead of the -120 dB measured by Amir? Is it related to the same adjustment as above?

The ratio is that of full-scale to measured level, so it's actually +3.5dB. This is from an Analog Devices paper on measurements:
1637987811743.png

THD+N is also a ratio, but I can't say if that's the reason for the difference from the screen shots posted.
 
  • Like
Reactions: trl

trl

Major Contributor
King of Mods
Joined
Feb 28, 2018
Messages
1,967
Likes
2,523
Location
Iasi, RO
In a few days, I will compare this program with REW.
REW and ARTA and all other software will show a worse THD+N when lowering the amplitude of the DAC.

However, I am still puzzled about the above measurement, as DAC was running at full amplitude and the -3.5 dB is a result of the analyzer scale being different. I would day that FullAmpl/InputAmpl should be a negative number, like (0dBFS/-3.5dBFS), unless I'm on a completely different wavelength. :)
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,626
Likes
10,202
Location
North-East
REW and ARTA and all other software will show a worse THD+N when lowering the amplitude of the DAC.

However, I am still puzzled about the above measurement, as DAC was running at full amplitude and the -3.5 dB is a result of the analyzer scale being different. I would day that FullAmpl/InputAmpl should be a negative number, like (0dBFS/-3.5dBFS), unless I'm on a completely different wavelength. :)

It's not a ratio of dBs -- dB is already a ratio. For example, if full scale amplitude is 4v and actual measured value was 2.67v. The ratio 4v/2.67v expressed in dB is 20*log(4/2.6) = +3.5dB.

In terms of ENOB, if the measured signal amplitude is less than full scale, noise floor remains at the same, but the signal is lower. Therefore, full-scale ENOB is going to be measured as a smaller number than the circuit is capable of and needs to be adjusted up.
 
  • Like
Reactions: trl
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,626
Likes
10,202
Location
North-East
Multitone version 1.0.7 is now available with some enhancements and bug fixes:

Changes in 1.0.7

  • Added: Click or right-click on the FFT Window name in the main display brings up a choice of available windows; no need to go to the Setup screen to change these
  • Added: Multitone maximum frequency setting is now also used for square an triangle wave generator, allowing for the test signal harmonics to exceed the Nyquist rate
  • Changed: Settings window frequency rates and dither bits can now be entered directly by typing them in, enabling the use of new values that are not in the drop-down list
  • Fixed: the Waveform plot legend had the Recorded and the Test signals swapped

Changes in 1.0.6

  • Added an additional pass of crest factor optimization to Multitone generator
  • Crest factor is now displayed for all test waveforms
 

MC_RME

Addicted to Fun and Learning
Technical Expert
Audio Company
Joined
May 15, 2019
Messages
854
Likes
3,564
Absolutely amazing. Congrats, Paul. Your work is going to set a new standard in multitone generation and analysis!
 
Last edited:

trl

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

trl

Major Contributor
King of Mods
Joined
Feb 28, 2018
Messages
1,967
Likes
2,523
Location
Iasi, RO
Multitone version 1.0.7 is now available with some enhancements and bug fixes:

D90MQA+Cosmos_REW_.png

D90MQA on Cosmos ADC measured with REW and external 1 kHz via foobar
SINAD: 119.9 dB


D90MQA+Cosmos_ARTA.png

D90MQA on Cosmos ADC measured with ARTA and external 1 kHz via foobar
SINAD: 119.2 dB


D90MQA+Cosmos_Multitone_2.png

D90MQA on Cosmos ADC measured with ARTA and built-in 1 kHz signal
SINAD: 104.8 dB


D90MQA+Cosmos_Multitone_1.png

Couldn't use WASAPI, perhaps something related to my laptop. However, using ASIO for Cosmos did the trick

D90MQA+Cosmos_Multitone_3(32_AP).png

Using higher FFT takes some time, wondering if Multitone can be optimized for more cores, please?
Thank you so much in your effort and time in deveopping this great application

Now the innevitable Q&A :) :
- How can I get multitones startng from 0 dBFS, to show the total dynamic of the device while playing multisines?
- Can you please let us run the show via external audio source? Basically, I'd like to run at least the 1 kHz test with my WAV file test.

Thanks again,
Raul.
 
OP
pkane

pkane

Master Contributor
Forum Donor
Joined
Aug 18, 2017
Messages
5,626
Likes
10,202
Location
North-East
View attachment 168614
D90MQA on Cosmos ADC measured with REW and external 1 kHz via foobar
SINAD: 119.9 dB


View attachment 168615
D90MQA on Cosmos ADC measured with ARTA and external 1 kHz via foobar
SINAD: 119.2 dB


View attachment 168617
D90MQA on Cosmos ADC measured with ARTA and built-in 1 kHz signal
SINAD: 104.8 dB


View attachment 168616
Couldn't use WASAPI, perhaps something related to my laptop. However, using ASIO for Cosmos did the trick

View attachment 168618
Using higher FFT takes some time, wondering if Multitone can be optimized for more cores, please?
Thank you so much in your effort and time in deveopping this great application

Now the innevitable Q&A :) :
- How can I get multitones startng from 0 dBFS, to show the total dynamic of the device while playing multisines?
- Can you please let us run the show via external audio source? Basically, I'd like to run at least the 1 kHz test with my WAV file test.

Thanks again,
Raul.

Hi Raul,

Thanks for testing!

- How can I get multitones startng from 0 dBFS, to show the total dynamic of the device while playing multisines?

I'm not sure if you mean Dynamic Range or something else? I didn't implement noise measurement with multitone signals yet, but this is something I plan on adding. This will allow DR and ENOB to be computed from any of the test signals.

- Can you please let us run the show via external audio source? Basically, I'd like to run at least the 1 kHz test with my WAV file test.
That's doable with a single sine test signal. Multitone or other complex signals are harder, as it would require finding all the peaks. It's also doable, but accuracy of such calculations might be affected.

  • The poor TD+N (104.8dBFS) value in your screen capture above appears to be noise-dominated, with low THD. I'd check why noise was so high in that test.

  • Larger FFT does take a bit longer, but usually this should happen only the first time you use that size FFT on your computer. Multitone spends some time the first time to optimize for new FFT size, but should re-use the results from that optimization the next time. The FFT itself does use all available cores (or as many as it finds are useful for that specific FFT size).

  • The WASAPI error that you posted is usually due some other device controlling WASAPI driver at the same time in an exclusive mode or with a different sample/bit rate than what Multitone is set to record.

Let me know if you had other questions from the screen shots you posted.
 
  • Like
Reactions: trl
Top Bottom