-
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
- Thread starter pkane
- Start date
OP
- Thread Starter
- #202
All custom. Other than the base FFT library, all the rest is my own implementation.
Nice work .... and lots of it!
Flexasio works fine in REW.
In addition I wanted to ask about differences in noise results between various configs.
There are 3 results below. One in Multitone with asio4all where noise/snr is being oenalized by being measured up the 1k fundamental. The other 2 are with REW using asio4all and flexasio.
I’d like to use flexasio since it provides better noise floor measurements on the d10b close to Amir’s but as I said I can’t with Multitone.
Any thoughts?
OP
- Thread Starter
- #205
Don’t know why flexasio would crash it — I’ve been able to use it but with different drivers.
The differences in noise measurements are likely due to the averaging difference between Multitone and REW: REW averages in frequency domain, while Multitone averages in time domain. Try not averaging with both and see if the noise results are more similar.
OP
- Thread Starter
- #207
Probably will take me a while, as I don't normally use flexasio. I installed it to test with Multitone and then removed it
EDIT: Just checked, it seems to be working, and I can bring up ASIO settings. Here's how it's configured on my Windows VM:
The white going up the sides of the fundamental are likely jitter or some other kind of distortion. You can zoom in around that area by using the scroll-wheel of your mouse to see it in greater detail, for example:
Last edited:
I tried wasapi as the backend fir flexasio duplicating your settings. It does indeed work and bringing up flexasio Controls dowpes not crash Multitone. So its the wdm-ks selection in flex that crashes Multitone. However the results are terrible compared to asio4all. Flexasio is much better than asio4all in REW though.
Last edited:
OP
- Thread Starter
- #210
I’ll take a look, but can’t promise anything as I have no insight into flexasio or wdm-ks. But I do find it strange that you’re having a much worse result with WASAPI. Did you try flipping the exclusive/non-exclusive option on in Multitone? Make sure you have the right setting in Windows audio device configuration, too, enhancements turned off and the right sampling rate and 24-bits audio selected for both, record and playback. Adjust system volume to maximum settings too.
I did try all the things you’ve suggested. This reference points out why wdm-ks would be the best choice and shared wasapi much less so.
github.com
FlexASIO/BACKENDS.md at master · dechamps/FlexASIO
A flexible universal ASIO driver that uses the PortAudio sound I/O library. Supports WASAPI (shared and exclusive), KS, DirectSound and MME. - dechamps/FlexASIO
github.com
OP
- Thread Starter
- #212
I did try all the things you’ve suggested. This reference points out why wdm-ks would be the best choice and shared wasapi much less so.
FlexASIO/BACKENDS.md at master · dechamps/FlexASIO
A flexible universal ASIO driver that uses the PortAudio sound I/O library. Supports WASAPI (shared and exclusive), KS, DirectSound and MME. - dechamps/FlexASIO
Sorry if I misunderstood, but did you try WASAPI directly, without FlexAsio?
OP
- Thread Starter
- #214
Can't reproduce the issues, but then I can't really get FlexASIO to work consistently.
Do you have Comtrue Cosmos ADC driver? If you do, you can use it directly and not need to use ASIO4ALL, FlexASIO, or other wrappers. Just select the correct ASIO driver and set the correct sampling rate in settings:
Last edited:
OP
- Thread Starter
- #215
Fresh off the press! Multitone version 1.0.14 is now available.
Changes in 1.0.14
- Added: J-Test 24 and J-Test 16 test signals for jitter measurements
- Added: Option to change frequency axis display between log and linear
- Added: Option to select how many harmonic components to show on the display
- Added: Automatic numbering of harmonics on the frequency plot
- Changed: IMD calculations improved accuracy and switched to using the more common IMD calculations (SMPTE and CCIF)
- Changed: Level/IMD sweep plot X-axis will now show test signal amplitude (previously showed measured signal amplitude)
- Fixed: a rare 'index out of bounds' error
That's great, @pkane. And I hate you in the same time now that I have to re-launch my runs for my next review.
Is the new J-Test the same as sq:250/12kHz 1:16777216 you sent me?
By the way, is there any chance that you could add the possibility to save results in a decicated file, or say, a way to do customisable screenshots? I know using a battery powered laptop dedicated for measurements (works much better for unbalanced captures using the Cosmos), but screenshots really lack resolution (1080p) against ones from my main PC (1440p). An option to lock zoom in X/Y axis would be cool, also.
Anyway, thanks for your awesome work !
OP
- Thread Starter
- #217
That's great, @pkane. And I hate you in the same time now that I have to re-launch my runs for my next review.
Is the new J-Test the same as sq:250/12kHz 1:16777216 you sent me?
By the way, is there any chance that you could add the possibility to save results in a decicated file, or say, a way to do customisable screenshots? I know using a battery powered laptop dedicated for measurements (works much better for unbalanced captures using the Cosmos), but screenshots really lack resolution (1080p) against ones from my main PC (1440p). An option to lock zoom in X/Y axis would be cool, also.
Anyway, thanks for your awesome work !
Sorry about that!
The built-in J-Test signal is generated using two square waves (by flipping bits in integer samples). This is how the original J-Test was defined.
The J-Test command I sent you should generate exactly the same signal, but is constructed slightly differently. At 48k, there is a low-level square wave at 250Hz just like in the original J-Test, but it's generated by computation involving multiple odd-harmonic sine-waves instead of simple bit manipulation. The 12kHz signal is also a single sine wave in that command, but in the real J-Test, it is a square wave that appears as a single sine-wave in the audio band. The effect and the resulting signal should be the same and the differences (if any) should be well below the lowest bit. I wanted to have the exact J-Test signal as part of this test, generated in the way as it was originally defined by Julian Dunn.
Let me see what I can do about the screen captures. I've run into that same problem that you describe with a measurement laptop and lower resolution display.
OP
- Thread Starter
- #218
A new version of Multitone (1.0.15) is available that adds a button to copy the current plot to clipboard. By default, this will use the current screen resolution, but you can change that by setting the "Clipboard Image Scale" option in settings. Scale of 2 would double the resolution, 3, will triple it, etc. You can also enter a fractional multiplier, if you want a very specific image size (for example, 0.5 will reduce the image to 1/2 the size). You can change the scale without redoing the measurement -- just change the scale factor and click on the camera button again. Not all the text elements on the plot scale correctly, so as you increase resolution, some of the text will become smaller...
Examples:
Scale = 1:
Scale = 3:
You made my day twice.
Unfortunately, the button doesn't work for me. Unless screenshots are supposed to go in some place. I didn't find where.
OP
- Thread Starter
- #220
The image is copied into Windows clipboard. After you press the camera button, you can go to any another application where you can paste an image, including into composing a new message on ASR, and just press Ctrl-V to paste the image.You made my day twice.
Unfortunately, the button doesn't work for me. Unless screenshots are supposed to go in some place. I didn't find where.
Similar threads
- Poll
