Tweet
2.0.72.1 (2021-05-27)
* Overlay offset stored to trace for information. Visible in tooltip.
* Overlay offset adjustment improved.
* Overlay offset stored to trace for information. Visible in tooltip.
* Overlay offset adjustment improved.
-
The latest build of 2.0.72.1 (2021-05-28 )
* Added dashed line support to Area charts.
* Added depth adjustment to Surface charts with Shift + mouse wheel.
* Overlay offset in tooltip refreshed while adjustment.Comment
-
2.0.72.2 (2021-05-29)
Main
* 'User's off-axis angles' in context menu of directivity chart applies to Area chart, Surface chart and Polar map (in addition to Line chart).
* 'Half space' applies to Surface chart, Polar map and Polar chart (in addition to Line chart and Area chart).Comment
-
Does someone love area and surface charts so much that I should keep them? "Speak Now or Forever Hold Your Piece".
I'm considering to replace both with single waterfall with cavalier projection - just like in MLSSA, visible @ stereophile.com. It would be fixed without rotation, zoom and pan, but has some frequency, dB and deg scales.
Comment
-
Now you have possibility to test which directivity charts are useful and which can be removed.
2.0.73.0 (2021-05-30)
Main
* Evaluation of Waterfall directivity chart.Comment
-
I personally don't get any better insight into the data on a 3D plot vs 2D. for me the 2D plot is easier to analyze, but 3D looks nice in a brochure
Comment
-
Kind of foreseable, I am very much in fond of this sort of representation, thanks for this evaluative implementation. This is a great addition to polar maps/sonograms and the spinorama display, which both are very helpful in their repective ways. Do I understand it correctly: Line ending to angle axis is already possible (as in MLSSA, slices reach down to scale base, ?)? Having a printable marker as in the MLSSA screens or as an option/checkbox in REW could make things even a bit easier: The possibility do directly point at some issue with precise data points in a discussion has use value. As this marker is a live option in the Power & DI window already, maybe setting a temporary marker which shows up on exported image is also possible. If that is too complicated, finer steps in frequency could help, too.
Comment
-
If I could suggest some added feature to the waterfall chart. If you look at the line chart or area chart you can move your cursor on it and it will highlight the individual angles and show frequency and amplitude in a tooltip. It would be great to have this functionality carry over to the waterfall. Realistically when I look at the different chart types I think the fully rotatable surface chart would be more functional if it only included axis scaling information and the same functionality as above line and area charts. This would be in my opinion a modern implementation of the MLSSA type of plot, and you would simply add a "MLSSA" view preset to set the camera angle to match the above MLSSA plot examples above.Comment
-
VCAD build 2.0.73.0 (2021-05-31) with waterfall lines ending down at the angle axis is available. Revision not changed so download and install manually.
---
I promised few comments about Stereophile's measurements. They visualize overall balance with listening window average (or RMS), and directivity with two normalized waterfalls. That is one approach, but not the best one imo. Normalized waterfall hides sound balance to all angles while ears and brain don't normalize by on axis (0 deg) so that's basically damaging information. Normalized directivity graphs are valuable for evaluating individual raw radiators/drivers, but Stereophile does not do that.
Off-axis balance would be included without normalization, and logic of directivity would be still readable from gap distribution of traces. Spinorama and unnormalized off-axis responses together would be better and more analytical approach for both designing and evaluation of complete speaker.
Directivity of speakers Stereophile has measured during many years is usually bad or very bad. I don't see much sense to use bad as a "reference". Designing better directivity should not be difficult with methods and tools described in VCAD documentation assuming that designer is willing to make more than just boxes with sharp edges.
Anyway, sound quality is more complex than just directivity and THD. Speaker with perfect spinorama and directivity plots is not necessarily winner or even tolerable. Sound could be balanced at far field, but that's not all. Exaggerated hype with scoring for spinorama and THD looks almost ridiculous on ASR forum. Best scoring products could be lame, boring and scattered, but genuine fanboy cannot hear anymore if measurements look almost perfect.
---
Waterfall chart could have some cursor system. More coding is needed because existing value tracing for line, area, polar map and polar chart does not work with axonometric plot having also shadow/hidden traces.Comment
-
I agree that free rotation is a neat feature to hone in on a problem area. But the perspective in MLSSA charts also seems to be chosen very well and should not be omitted. Above normalized axis, off-axis trends which can be expected to reach listening position without reflection are emphasized in visual contour, below in the foreground, higher off-axis angles streched out. The difference of whom can easily be understood. I also agree it could be helpful to see dB levels, i.e. to find out where -6 dB is reached.Comment
-
All 2D directivity graphs have cursor for value tracing. Polar map also interpolates within simulated directions. All points in all off-axis traces are covered for sure so cursor is not mandatory in 3D and axonometric graphs. It's just perfectionism and extra work for me.Comment
-
Normalizing to 0-axis can be limiting, as one would generally like to design for average listening window and accept a deviation on-axis. Nomalizing to 0-axis could give the impression of issues which were actually an accepted trade-off in listening window angles. There is still the option to not normalize, the easily available directivity information this sort of graph offers is still there:
At least for me, I find it easier to understand i. e. the performance weakness of the waveguide used (in the case of this graphs the differences begin at ~40 degrees off-axis), as compared to in a sonogram alone. In sonogram, I need to chase areas with the marker or set different levels to see the problem unfold into it's depth-axis. Also, a full spinorama is much more powerfull than this sort of graph, sure. It is good to start with spinorama. They relate to another as in general performance vs. detailed area of interest. What I see from spinorama sort of graph is the difference between early reflextions and listening window, but I do not see as precise the directivity issues and why I hit limits.
That's fair enough, the tool given to us by you is very powerful already and the analysis can be done with the other graphs, if coding is too much work here.Comment
-
Hi Kimmo,
VituixCAD is an amazing tool that I use now for all my Diy audio projects. I can see that it is almost improved every day !
I can only see one thing missing to become the only tool that you ever need is being able to make measurements...
Are they any plan to implement such tool in it ?
Anyway thanks very much for bringing such a tool for us audio amateurs
Jean ClaudeComment
-
Hi Kimmo,
I've RTFM and have been using VituixCad2 and love it. I think I submitted a donation because it helped me a lot whilst in 2 weeks quarantine learning about speaker design. Of course I wasn't able to go outside hotel to take measurements, but using the built in sample measurements for the 2 and 3 way, and using it to learn XO design- and learning powerful features of the app eg. Enclosure, Diffraction, SPL trace etc was great.
But one thing I wasn't sure about- is there a feature or process whereby I can automate/predict how a published FR (eg. SEAS standard 12L cabinet baffle) will be transformed when I insert the driver into a different baffle of specific size/shape?Comment
-
Provided your data is 2pi / infinite baffle like many datasheets, simply bring the response into the diffraction tool.
If the provided data is measured on a baffle, and you know the dimensions and driver location of the baffle, you can simulate the diffraction of the measured baffle using the diffraction tool, then "subtract" it from the measured FR using the calculator (IIRC the correct operation is "divide" not subtract) to arrive at an infinite baffle response, then do the same as above, bring the response back to the diffraction tool ans simulate on your new baffle with the understanding of accumulated error in each calculated process, so the result may not match reality with a high degree of accuracy.Comment
-
I have been trying to simulate a rear firing woofer like the Genelec discussed before to see what effect that has compared to the side mounted drivers. I want to make sure I am simulating it properly and not fooling myself.
Should I be using the the Rotation parameter and setting it to some value along with a z depth or invert the polarity along with the z depth or something else?Comment
-
Rotate the back driver(s) by 180 deg, choose the Z depth and the polarity inverted.Comment
-
You need quasi full space data in two planes for the drivers. Minimum response set is 0-180 deg in horizontal plane which can be mirrored to negative angles and vertical plane.
Then you locate and rotate each driver instance in crossover network how they will be in reality. For example:
- Front woofer X,Y,Z=0,0,0 mm, R=0 deg, T=0 deg
- Rear woofer X,Y,Z=0,0,350 mm, R=180 deg, T=0 deg if box depth is 350 mm and both are at the same Y elevation on opposite sides.
- Rear woofer X,Y,Z=0,-200,350 mm, R=180 deg, T=0 deg if box depth is 350 mm and rear woofer is 200 mm lower than front.
Rear woofer is delayed with low pass IIR filter and polarity inverted if target is cardioidish. Dipole needs just inversion but not delay or low pass.
Genelec style approach needs extra IIR high pass for closed woofer to match order and phase with vented woofer in order to extend cardioid range down to lowest octave. Symmetrical front and rear is easiest and simplest to get any pattern. Longer distance between front and back (due to rear woofer) also moves "ideal" frequency range down. SPL capacity increases but possibilities to cover lower mid-range decreases.Comment
-
-
-
-
Latest build of 2.0.73.1 (2021-06-03) uploaded with updated user manual. Waterfall drawing should be a bit faster and code cleaner without development drafts. Color of highligted pointer depends on fill color of selected off-axis trace.
Question: is waterfall so good that area chart or surface chart or both can be hidden or deleted for good?Comment
-
I would say yes. I certainly wouldn't miss them. I can see everything on the waterfall I used to see switching between polar maps and line charts with occasional glance at polar chart to understand lobes. I bet lobes would show up very well on waterfall. As I raise XO on my current design to see lobes, its true - they do show up dramatically although parts can be obscured behind other bumps in the response.Comment
-
One more small tweak added: 'Eighth space ±45 deg' to context menu of directivity chart. No need to play with power calculation settings or user angles in Options or remove frequency responses of driver to show off-axis within ±45 deg. Limiting visible sector also maintains correct results in Power & DI.
Comment
-
Well, the area chart never did show anything that you couldn't get from the line chart, that said I don't see any reason to remove it if there are users that prefer it. The waterfall in its current state is more useful than the 3D surface chart, however as I said before the ideal function would be to have the waterfall be fully 3D like the surface chart, but not completely necessary and I understand that it could involve a lot of work for something that would be neat but doesn't really add any new functionality.
VituixCAD has not been flagged for update for the 1/8 space addition yet. While on this topic, does it make sense to change the vertical scale of the polar map to +/- 90 degrees or 45 degrees based on the selection. It is currently a full +/- 180 degrees regardless of the selection of half or eighth space.Comment
-
'Half space' and 'Eighth space' names are not the most logical because simulated space is selected with Power & DI calculation settings in Options. Also Polar map should follow angle limits selected with context menu. Context menu will probably have 'Show ±90 deg' and 'Show ±45 deg', and Polar map will follow those constraints in addition to Half/Corner selection in Options. Exception is User's off-axis angles (not yet decided what to do with it).Last edited by kimmosto; 05 June 2021, 01:51 Saturday.Comment
-
2.0.73.3 (2021-06-05)
Main
* Added depth adjustment for directivity Waterfall with left mouse button down + mouse move up/down (or Ctrl + mouse wheel).
* Area chart removed. Waterfall is flat area chart when depth is adjusted to zero.
* 'Half space' in context menu of directivity chart renamed to 'Show ±90 deg'.
* 'Eighth space' in context menu of directivity chart renamed to 'Show ±45 deg'.
* Angle scale (Y2) limits of Polar map follow 'Half space' and 'Eighth space' settings (in addition to Half/Corner selection in Options).Comment
-
Have not found the time to say "thank you" for this update, very nice. Would it be possible to also enter the depth-ratio as a number? What is used in MLSSA?Comment
-
Comment
-
2.0.74.1 (2021-06-12). The latest build a minute ago.
* Added Logo Show checkbox to Options. Check to show logo in charts on display.
2.0.74.0 (2021-06-11)
* Added Logo open button, clear button, filename text box, position combo box and opacity text box to Options window. Logo is printed to chart area on Copy and Export commands (except svg-files). Default VituixCAD logo is printed if filename is empty and position is other than None.Comment
-
JohnPM just added this feature in the latest version of REW.
Comment
-
Working with VituixCAD a bit today, I have a couple suggestions. I have a habit of closing the Diffraction window without saving, it would be great if the baffle configuration was not wiped out when the window is closed, so when I open it back up it is just where I left it.
Second suggestion is to include a tool tip or some information under the IR to FR section of the help file for expected file naming convention, "file_hor_+deg000.pir" format needs a reminder sometimes so when I start measuring I can keep to the correct naming conventions. I found the section in the help file "angle parsing from filename" but it is not referenced in the IR to FR section of the help file.Comment
-
I'm really interested in your comment here and curious what other factors you think are most important or responsible for a speaker with high spinorama score sounding lame.
From this I might guess your answer to my next question but I'll ask anyway. Would it be a bad idea to add optimization of the DI curve as an option? For those using DSP crossover it is trivial to flatten any one of the other curves but the relationship between them, of which the DI is one measure, seems like the hardest thing to get right.
Many thanks for all your efforts. I find it incredible that such software and all the continual improvements is freely available.
Phil
Sent from my SM-N950F using TapatalkComment
-
Spinorama tells me only if tonality is right but i don't care at all about scoring system. It is flawed and it is known fact. People that look at score are usually the ones that don't understand the caveats of it and just try to decide what loudspeaker to buy without listening. In a way, that's ok. You maybe don't get the best sounding loudspeaker of the bunch but you don't end up with some underengineered crap either.
In your opinion, which factors play main roles in a good sounding loudspeakers - good sounding to you?Comment
-
Thanks Zvu. I hesitate to reply because in all honesty my experience of different speakers is too limited and is none at all in controlled conditions.
I agree with you about about the scoring system but it seems to me that the full spinorama and polar map does tell you a lot about a speaker. I'm genuinely curious what else there is other than tonality, THD and directivity, so long as power handling is adequate. Lots of other things like resonances, break up etc matter but aren't they likely to show up in the charts?
As to other factors: I have experimented with linear phase crossovers but find any benefit extremely subtle if there at all. My current speakers are an experiment in cardioid mids with a repurposed Kef coax hung open-baffle over a vertical 5 in mid woofer inspired by Linkwitz LXmini. An upwards firing Dayton 8inch sub driver just behind provides the lows. They sound good to me but I am biased! Does the unusual directivity make a difference? I don't know, it was certainly fun to play with it but I don't have a control to A/B. Likewise for the point source coax - does it sound better because it is point source? I don't know but it does make the crossover easier with no concerns over lobes.
As an aside I drove myself nuts playing with eq on headphones- different target curves, using in ear mics and switching profiles on EqualizerAPO. Two things I discovered 1. The ear quickly adjusts to a profile such that almost any change initially sounds wrong and 2. The most effective tweak is always to turn the volume up a couple of dB!
Sent from my SM-N950F using TapatalkComment
-
Hello Kimmo,
would you find it worthwhile to add a separate "Adjust rotation" dialogue to the waterfall plot? I recently stumbled upon the following kind of graph (from audioholics) ...
... and I liked the isometric rotation of the graph, admittedly, also the shading style.
RegardsComment
-
I will agree that shading by amplitude makes a lot more sense than shading by angle. It would make sense for the waterfall to follow the polar map for shading perhaps.Comment
-
I can add "Save baffle project?" question to form closing. It will ask every time without trying to detect possible changes.
That is too late. File naming requirements should be clear before starting measurement sequences. Requirements are already explained in:
- Measurements with ARTA
- Measurements with REW
- Measurements with SoundEasy
- User manual, sections:
* Checklist for designing a loudspeaker, Measurements
* Drivers tab, Frequency responses
* Options, Angle parsing from filename.
Totally 6 places in 4 documents must be enough.Comment
-
Dynamics/compression, timing, total directivity and radiator type / features of wavefront.
DI alone is not possible to optimize. Either axial (on-axis or LW) or sound power response is also needed to get some sense and quality to result. So in practice weighted combination of axial and power is the same as weighted combination of DI and axial or power.Comment
-
Roughly right. Problem is that everyone can and will make own subjective interpretation what is good relative level and shape of each response in spinorama because there is no "ideal" or "optimum". Just coarse guidelines specified by investigators with their experience (limited or not), studies, used products, listening environment and setup.
Hopefully authors on ASR would also have a bit better understanding.
Room acoustics and listening setup are usually more important than features of the speakers assuming that speakers are good. I'm keen to minimize problems in many features so naming of main roles is not easy or pleasant. Few basic items: smooth/flat response to reference axis, smooth power response with proper tilt, short excess GD down to bass range, compression spectrum compatible with music, total directivity compatible with room acoustics and desired listening setup/distance, low coloration/resonances at MF-HF, low non-linear distortion. In addition there are some features which are difficult to explain with measurements. For example horn with Ti compression driver. Problem for me is probably some toxic combination of compression spectrum, mechanical features of driver and shape and material of horn. Result is probably unacceptable for me no matter how perfect spinorama looks. So finally quality of drivers and other components evaluated subjectively by ears and brain could be more important than close to perfect spinorama.Comment
-
Whole idea of adding waterfall was to maintain correct/original tilt of frequency responses. Surface chart is for free 3D rotation (without axes).
Polar map has coloring by magnitude. Waterfall has correct/original tilt of frequency responses so color can be associated for angle to be compatible with other directivity chart(s) with frequency response traces. For example Line chart has legend in copy/export which needs color by angle.Comment
-
-
Perhaps I’m doing wrongly.
In the Diffraction tool I initialise my New Baffle using the IEC dimensions for measuring drive units 8 inches or smaller-1350mm ×1650mm (WxH)
Then I set up my driver unit in this case a 5” midwoofer at the off centred location.
However when setting virtual mic distance; the minimum is 2200 mm.
SB Acoustics datasheets use a measurement distance of 31.6cm = 316mm of their IEC baffleComment
Comment