May

Hi!

As I told you, I will be at SAMPTA09. I will certainly be in Marseille from the 16th to the 23rd of may.

Should I try to come already on Friday, 15th? Could we meet in the late afternoon? Can we have a discussion then? (We probably can also meet at least once during the conference, which is at Luminy.)

If yes, do you have suggestions where I could stay the night from friday to saturday? (I have rented a flat for one week, but they only rent it from saturday to saturday.)

I have heard there are guest rooms at the CNRS, could I use one of these for one day?

Best wishes,
Peter

Hi Peter,

That's a good news to know about your venue in Marseille

As far as I'm concerned it's ok to meet on the 15th of May. We can start a discussion at the LMA late in the afternoon and then pursue it in another place I don't know about the others' availability, let's wait for their answers...

Yes there're some guest rooms at the CNRS but I don't know the procedure. I shall ask the right person tomorrow.

Cheers,

Thibaud.

P.S. : I'll check for concert occasions during that period

Hi,

It is OK for me too to meet on Friday, 15th. But I won't be free in the evening, I am going to a concert in Istres. You could join me but I am not sure you will like it, it is Grace, kind of "world music", I would say "cool" music....

I let Thibaud to take informations for having a room at the CNRS.

Sabine

Hi Peter,

I called the women in charge for the room's reservation at the CNRS. There was only one room available for the 15th of May so I took an option for it. The price is about 30€. I don't know which services it comprises except the bed
Once you confirm me that's ok for you I shall ask more info and start the procedure. We have to fill some papers (as always... ) and I think you can pay on the day you arrive.

So, tell me!

Have a nice week-end,

Thibaud.

Yes, book it, please!

I arrieve at 12.20 at the airport, and will head directly to the CNRS.

That's perfect!

I'll confirm the room's reservation and inform you about the procedure.

Cheers,

Thibaud.

Here I start a short list of the important things to discuss on the 15th of May:

• Papers: Detailed plan of part I and preparation of part II.
  
• Test of the CASP model (if we could contact T. Dau before the meeting we could think about what to do then).
  
• Modeling step: Implementation with wavelets, model calibration and validation...

Everybody is obviously free to add topics!
I'm pretty sure the present list may be extended

Thibaud.

Dear all,

Here is a summary of the meeting that happened on the 15th of May 2009 at the LMA:

Participants:

• Peter
  
• Richard
  
• Sabine
  
• Sølvi
  
• Thibaud

Topics:

1. Papers (psychoacoustical experiments)
2.Test of the CASP model
3.Modeling approaches

1. Publications

Bernhard almost completed the multiple-masker experiments paper. He's currently working on the Introduction and will submit all the team members a preprint version as soon as possible. To enhance the collaboration between LMA & ARI, the idea is to submit both papers (single- and multiple-masker experiments) at the same time. At the beginning of summer would be great.
Concerning the single-masker experiments paper, Thibaud is currently writing the experimental content of part I (based on the plan agreed in this section). Then, the Introduction and General Discussion parts will result from a collaborative work of all the team members. Important aspects to think about: How to Introduce the project and its context, make the collaboration and the pluridisciplinary work clear, main conclusions to draw from the experiments (apart from the modeling application)... Depending on the results provided by the CASP model, part II shall simply consist of a “letter to the editor” mentioning that we tested the model availabilities in predicting TF masking data.
Finally, to clarify each member's implication in the hard work surrounding publications, an issue must be found to the “list of authors problem”. Since not all the members could assist to this meeting, this discussion was not extended. Thus, a meeting has to be organized quite soon (by phone or visio-conference) with all the members.

2. The CASP model

It was decided to send a bug-report to T. Dau in order to get a more advanced version of the code.
UPDATE: I've discussed and exchanged mails with Peter Søndergaard (a post-doc working within the Dau's team). He's in charge to re-write the CASP model code to as to provide a Matlab toolbox. He told me that the code I had was the official one. However, he could certify the false result predictions and did not find a reason to this. He recently sent me a new code. The latter seemed to be an older version compared to the previous one. Anyway, I didn't succeed to run it, there's a fatal error somewhere in the code. I warned Peter S. about that.
[Complaining mode ON] I propose that we forget this topic for the moment. Let's wait for a complete & stable version that I can run easily without spending hours trying to know how it works... [Complaining mode OFF]

3. Modeling, implementing and testing the model

The modeling step is currently under investigation. A wavelet approach has been chosen to analyze, represent, transform and re-synthesize the input sound signals. The psychoacoustical data obtained at different levels have to be interpolated and convolved with the original Time-Scale (TS) representation. Another approach shall consist in “applying” the interpolated masking data locally in the TS domain.
Remark: Since temporal masking has not been measured for different masker levels, some typical results from the literature have to be considered and integrated in the model.

It is quite clear that we won't have time (within the PhD project schedule) to build a completely efficient model, nor to fully evaluate it. The choice was made to implement a version that can process simple sounds (even if a long computation time is required) and to use such simple cases as a quick model evaluation/calibration. This should be sufficient for the thesis.

Thibaud.

Good job, Thibaud,
this looks very reasonable. Just one remark that it would be nice to clarify - I didn't really follow the letter to the editor-stuff, i.e.
"Depending on the results provided by the CASP model, part II shall simply consist of a “letter to the editor” mentioning that we tested the model availabilities in predicting TF masking data.",

You might have discussed this after I left the meeting... A letter to the editor is like a short article - is this what you plan to do in addition to the other articles? From what I understood, the CASP model is not yet available for running, and therefore you don't really know if you'll ever be able to obtain any results with this model... Well, maybe I skipped something, but it would be nice if you could explain what you mean by this.

cheers!
Sølvi

I have the same question than Solvi concerning the letter to the editor.

Concerning the authors of the papers, I think we should make a phone meeting (or visio conference) now and close this question. What about doddle to choose the date ?

Cheers

Sabine

To clarify the "letter to the editor" stuff:

At the moment we're not sure if we will get a stable version of the CASP model "in time", i.e., within my PhD. If this happens, I'll run the model with our stimuli and compare to our results. Then, instead of writing a full paper, we thought about writing a short letter to the editor as Part II in which we will mention the comparison between the actual results and the CASP model predictions.
If we don't get a stable version before I finish my PhD, this job will be part of the project's future work. Anyway, due to the Acoustics'08 reactions, we'll be forced to mention the CASP model within part I of the paper.

I hope this is clearer now. If not please tell me!

sabine wrote:

I think we should make a phone meeting (or visio conference) now and close this question. What about doodle to choose the date?

I agree I still don't have the "doodle reflex" but this is a nice tool. To find a date we will have to know when Peter shall be back to work. Do you have any idea?

Cheers,

Thibaud.

Could you post all comments concerning the papers in the correct topic please? That becomes messy here and I'm completely lost

I found how to merge the 2 topics about publications in a single one, pleeeeease do not post about publications elsewehere