Comparing drivers with waterfall, not so easy to understand speakers difference. Sonogram is very good. What program can export flat sonogram?

I find this sonogram from Dibirama site very useful. Very informative and understandable how the driver sounds.
2023-03-13_18-57-33.png

What program can export flat sonogram like this?
 
Last edited:
Those images are a waterfall seen from the top. Somewhat similar to a sonogram. There is another version of charts looking like those that illustrate disperion vrs frequency.

I would really like to see software providers do this chart, and impedance, as a rotatable 3D graph. And a switch for periods vrs time. Switching to periods means that the Q and magnitude of any ringing becomes directly comparable and one can quickly see how a driver rings by whther or not the lines are parallel.

dave
 
Clio program works without interface? If I use my microphone? I don't know why but this sonogram looks much more understandable for me ower other waterfall methods. Maybe because I'm new bee to DIY audio.
Nope: Clio, by Audiomatica, it's a hardware+software measuring device.
Cumulative spectral decay (CSD) can be visualized through different plots: the waterfall or the color map (the one you like more). It's main use is for anechoic loudspeaker measurements.
Energy time frequency (ETF) can be visualized as well with waterfall or color map plots. It is mainly used for room acoustic measurements.
Then there is the wavelet analysis........
 
  • Like
Reactions: inetboy
Nope: Clio, by Audiomatica, it's a hardware+software measuring device.
Cumulative spectral decay (CSD) can be visualized through different plots: the waterfall or the color map (the one you like more). It's main use is for anechoic loudspeaker measurements.
Energy time frequency (ETF) can be visualized as well with waterfall or color map plots. It is mainly used for room acoustic measurements.
Then there is the wavelet analysis........
But really possible to hear resonances?
Like on graphs before on the right side for example.
The length of 700 hz is 1.42857143 ms. This means this resonant hits all the time at -30 db? But realy posible to hear -30db? Overal look like at -10db level all the time is some little resonance with a higher frequency. 10 000 Hz is 0.1 ms. Hmm... this is interesting... Sorry, these are just newbie thoughts.
 
But really possible to hear resonances?
Like on graphs before on the right side for example.

The length of 700 hz is 1.42857143 ms. This means this resonant hits all the time at -30 db? But realy posible to hear -30db? Overal look like at -10db level all the time is some little resonance with a higher frequency. 10 000 Hz is 0.1 ms. Hmm... this is interesting... Sorry, these are just newbie thoughts.
Yes it's possible to hear a 4 ms decay, depending on frequency of the resonance: check Fletcher&Munson curves.

Look at the sonogram on the right: at 1KHz, it's telling you that the sound takes 4 ms to decay completely or that it takes an attenuation of 50dB to disappear (assuming the 0dB response as the starting level).
 
  • Like
Reactions: inetboy
Look at the sonogram on the right: at 1KHz, it's telling you that the sound takes 4 ms to decay completely or that it takes an attenuation of 50dB to disappear (assuming the 0dB response as the starting level).
Very useful, understandable from my opinion.
I found "fattest" red line from them.
This is R2904/700000 fro SS.
Opera Снимок_2023-03-14_163155_www.dibirama.altervista.org.png

The higher prise, more resonances?
This is T29CF002 from Seas
Opera Снимок_2023-03-14_163644_www.dibirama.altervista.org.png
 
... Look at the sonogram on the right: at 1KHz, it's telling you that the sound takes 4 ms to decay completely or that it takes an attenuation of 50dB to disappear (assuming the 0dB response as the starting level).
I am mistaken? I thought the ending of the color legend at -50dB is a display/graphic choice (IIRC similarly, you can change the height of the 'lake surface' in the REW waterfall to hide or uncover ridges out to the resolution of the measuring device). I thought the audibility threshold varies with the frequency (and person), so it isn't as simple as looking at the black boundary.

Unless I have completely missed the graph, it might be better to pick a different frequency for the example, since at 1k Hz the data color/black transition, the axis label of 4.00 ms and the bottom of the color legend bar all appear at the same height on the chart, so it would be easier to confuse the color legend with being associated with the axis labels on the left.
 
I am mistaken? I thought the ending of the color legend at -50dB is a display/graphic choice (IIRC similarly, you can change the height of the 'lake surface' in the REW waterfall to hide or uncover ridges out to the resolution of the measuring device). I thought the audibility threshold varies with the frequency (and person), so it isn't as simple as looking at the black boundary.

Unless I have completely missed the graph, it might be better to pick a different frequency for the example, since at 1k Hz the data color/black transition, the axis label of 4.00 ms and the bottom of the color legend bar all appear at the same height on the chart, so it would be easier to confuse the color legend with being associated with the axis labels on the left.
I see like this: the measurement where taken at 90dB, as stated in the graph posted, which is the 0 dB in the CSD. The decay, expressed in time or dB, is what you can read in the two vertical axes of the graph.
For sure the audibility depends on frequency: that's why I suggested the F&M chart.
 
Those images are a waterfall seen from the top. Somewhat similar to a sonogram. There is another version of charts looking like those that illustrate disperion vrs frequency.

I would really like to see software providers do this chart, and impedance, as a rotatable 3D graph. And a switch for periods vrs time. Switching to periods means that the Q and magnitude of any ringing becomes directly comparable and one can quickly see how a driver rings by whther or not the lines are parallel.

dave
You might get part of the way there with this plugin to ImageJ that takes a flat colormapped image (I believe the sonogram would work) and turns it into a 3-d image. From the picture, seems to work in iOS:

https://imagej.nih.gov/ij/plugins/surface-plot-3d.html

https://imagej.net/ij/index.html