Nov 19

Extra feedback musicians

After finishing this project we got some extra feedback from Artvark, which we like to share with you. They’ve had the change to try the mouthpieces at home and have a good look at them. The mouthpiece for the alto saxophone, Meijer 8, was tested another time by two other musicians of Artvark. After Peter Broekhuizen already tested these mouthpieces,  Rolf Delfos and Bart Wirtz gave it a try.

Meijer 8 (copy)
Rolf: surprisingly nice in the lower tone section. A good resistance. The sound of the higher tone section is round, but has less definition and suppleness.
Bart: full sound with a good resistance. Round sound but a bit undefined.

Meijer 8 #1
Rolf: less resistance, and the sound is smaller and thinner. (No go for Rolf)
Bart: more projection, smaller sounds which makes it hard to create a big volume. Nevertheless a tight sound.

Meijer 8 #2
They both didn’t like this version in the first place so they didn’t experiment on it further.

Meijer 8 #1.1
This modification is a ramp with a slit in the middle. Peter didn’t tested this one, because he was more focused on the baritone mouthpieces.
Rolf: this one is a combination of the copy and #1.
Bart: too much tight sound. In the higher note section, there is not enough resistance.


Oct 31

Project review

Looking back at the project, it was a really good project.
We have printed lots of new mouthpieces and now have a lot of feedback from the musicians.
It is now important to sort all the feedback out, so that in the future it is possible to use this information to progress.

Sorting out the steps and facing the problems that occur to get from scan to a new printed mouthpiece was a very good experience. It was educational in a way to learn to solve problems. Now we have a clearer image of what reverse engineering is about.
Many hours of solidworks shows that it is not that easy to reverse engineer a mouthpiece. It still involves a lot of human work and decisions.

There were some problems with the 3Dprinter’s nozzle during the printing. Unfortunately, a lot of pieces came out with dings in the surface. This influenced the quality of the mouthpieces. In the future it would be nice to try different printers (Form1). Also something has to be done on the material. All the testmusicians didn’t like the material’s taste.

It was also a good experience to work with real musicians and create real mouthpieces. Testing with musicians was also very useful to gain insight into the value of the mouthpieces for the users.

We had a really blasting ending of the project with the live performance of artvark. It was really nice to speak with interested people and explain what have we worked on over the past 4 weeks.

Oct 30

Science fair

Yesterday we had the science fair! This was a nice event where every group showed their results of their project. So did we! We showed the mouthpieces and a slideshow of the project. The best part was that three off the guys from Artvark Saxophone Quartet came to play our 3D-printed mouthpieces. Here you can see a movie of the two songs they played. Play it in 1080p HD!

YouTube Preview Image

Oct 28

Future expectations

During the project, we experienced that converting the mesh-files into a 3Dmodel
was the most time-consuming activity of the process. When in the future a
program is written that can easily covert the meshfile into a 3Dmodel, we see
some opportunities for shops selling wind instruments. Musicians could go to the
shop, scan their own mouthpiece and instantly adapt the mouthpiece to their
wishes. Thereafter, the mouthpiece can be printed directly.
Hereby it is needed to match terms from musicians that express a certain sound
or experience with a soundspectrum. This spectrum again can be matched with
certain geometric changes in the mouthpiece. For example, if a musician would
want a rich sound, the shop can make the right geometric adjustments to the
This way, musicians who used to search endlessly for their ideal mouthpiece, can
have it directly.

Oct 28

Overview modifications

Here is an overview of the modifications we did on the mouthpieces. We will show this with section views of the drake mouthpiece.

First we made a copy of the mouthpiece (first picture). After that we modified this copy with a small ramp (second picture) and a small gap (third picture) in the inside of the mouthpiece. We placed the modification at the place where the speed of the airflow is the biggest.

The test with Peter made clear that the modifications played pretty well. De alto and soprano mouthpieces played really well with the ramp and the Drake mouthpiece used on the baritone played well with the gap inside.

However, the Drake mouthpiece wasn’t as good as the original, the alto and soprano were. That’s why we made some variations to the modification on the Drake. We changed the place and the scale of the gap as you can see in the pictures below.



Oct 24

Peter Broekhuizen test!

On tuesday we tested our prints with Peter Broekhuizen. Pim made a clip:

YouTube Preview Image

Peter gave us a lot of good feedback. For every mouthpiece the top 3 is given. Modification number 1 is the one with the little ramp on the inside. Modification number 2 is the one with a litlle rectangle removed from the inside.

Soprano (Selmer)

#1 (might be better than the original)

The first modification was the nicest sopraan mouthpiece. There are some notes missing in the middle range though. The original copy had a very flat sound, had higher notes and a lot of volume. The second modification made it really heavy to play.

Baritone (Drake)


The second modification was Peters favorite. In the lower note range the tones could be higher though. The printed mouthpieces made that the sound stayed in front of the saxophone instead of surrounding the musician. The first modification was untrustable and seemed to fade out. Wasn’t good.

Alt (meijer 8)


The first modification was not bad at all. The sound had some brilliance. The higher notes were not too good though. The copy had a really different character than the original.

Baritone (JDX9)


All the JDX9 prints are better than the original mouthpiece! The copy has a clearer sound and has got more volume.  The sound is better in all of the note registers. The sound also seems to be surrounding the musician really well. #1 plays a bit easier, but then has a emptyer sound. #2 has an even smaller sound and the sound is further away from the musician.

In the coming days, we will prepare for the science fair on tuesday and make some modifications on the mouthpieces with Peter’s feedback. Feel free to come and visit us at the science fair!

Oct 21

New batch

So a new batch of mouthpieces came out of the 3D-printer! We did some small changes in the inside geometry of the mouthpieces. Unfortunately there was a small problem with one of the nozzles of the 3D-printer. Due to this small problem some of the mouthpieces have a little dent on the outside.

Tomorrow Peter Broekhuizen from the saxophone quartet Artvark is going to test the mouthpieces! We will film the test session and compare the different sounds of the mouthpieces. Also we will ask Peter to give feedback.

Stay tuned for a video of the test!

For a small expression of Artvark check this link :


YouTube Preview Image


Oct 18

Summary dicom to STL

Here is a small summary of what we had to do to make an STL file from the Dicomm files we got from the CT-scan.


We first used Avizo software to make a mesh from the dicom files. After importing the files, a few steps have to be made. The first step is to look at the contrast of material and air. This can be done by looking at the ortho slice. The contrast can be changed bij moving the arrows across the “hills”. The contrast is right when there is as less scatter as possible. Then an iso-surface has to be made using this contrast. From this iso-surface it is possible to extract the surface. The extracted surface can be first be simplified. The number of surfaces of the mesh can be reduced so the file isn’t that big. After the simplification a VRML file can be made.

This VRML file can be opened in the Meshlab software.

Because the mouthpieces were too big to scan in one time, we had to combine the different meshes of the mouthpiece. We did this in Meshlab. How to combine these different meshes is explained in this link: . Now the meshes are combined.


To make the out coming file smaller we did a simplification and reconstruction, using Filters Remeshing and Quadric Edfe Collapse Decimation. The size was also changed to 50000 triangles. The last step is to put the layers together using Filters Mesh layer à Flatten Visible layers. Now your model is ready to save as STL.

This STL file can be opened in SolidWorks. When we import this in an assembly we can create a new part and trace the mesh to a nice solid model.

Oct 15


Lizzy’s roommate tested the first mouthpiece!

Check it out!

YouTube Preview Image

Oct 15

The next steps

At the moment we have the mouthpieces ready for the try-outs. In the figure the next steps are visible.