Found out that in order to pan or balance the most significant bit is the sign.
Took another shot at trying to resolve the hum which seemed worse now. It just occurred to me that this might be because two channels are at max volume now, instead of just one. I will have to remember to check this tomorrow
After many failed attempts to fix the problem, as well as the pan and balance function not being able to completely isolate the channels, I think its best to get two more midi shields and just have one for each finger. The sound is pretty clear when only one channel is playing. Not sure why this issue is happening, I even tied to isolate it by using the arduino that isn’t hooked up to anything.
The headphones port makes the same noise when not hooked up to the amplifiers. It seems to have something to do with the ground not being isolated but everything I try either makes it so the midi commands won’t execute or theres not enough power for anything.
The reset pin has to be set high or the device is off.
I can eliminate the some of the noise by putting the channels on low volume til they are needed the main amplifier already does this though so its more a problem of the two channels.
Changed all the buffer sizes to accommodate more data.
Was able to get multiple structs to pass and import properly
Added a check command function to the arduino side. This checks for a new commands and executes them. Commands are in hex so that the first number can be the finger number. example 0x11 would send the change volume command to the index finger.
Deployed Dong’s code. It is definitely working better
Demo’d to Tangie
Found my bug with Eclipse. The [Event] within the <Metadata></Metadata> tags was killing code complete. This link to stackoverflow was helpful. Ended up posting the bug on Adobe’s Jira, since I couldn’t find it anywhere and it’s easily reproducible.
Got single search working. Monday we’ll do multiple selection search
Figured out what was causing the index and thumb to short out when pushed at the same time. One of the amp lines had the polarity of the capacitors reversed.
There are some sound quality issues. I will try tinning the leads tomorrow to see if it just a weak connection
Figured out part of the quality issue is some of the fingers had their input lines reversed.
Another thing to look into tomorrow is if the quality is bad simply because it keeps getting noteOn commands. The pervious better sounding sketch made sure it only did this on initiation.
Worked on building up the communication stuff. Realized that I was only able to send one struct at a time, so I added a loadBuffer method so that I can send all at once. Before the buffer was in the main now its part of the commgr class. The DataDictionary import method seems like it should work as is because it should continue parsing the buffer til it hits a null.
Built the interface. A little strange but it seems FLUID does not have a standard label that isn’t attached to anything. I didn’t really need it but still found it strange.
Currently the PC side is a sandbox so I’m building it out to have a more defined purpose. Now that I have things a bit more together I found that I really don’t need to pass as much data as I thought. All it needs is sensor data to the pc and command + data going to the arduino.
Still needs a bit more done before I can link the data with the interface.
Change Budget Center to something like “Identifier”, since that doesn’t seem to be a unique identifier. Maybe fine all the items that do change within a column and use them to select which line item to selct? Other common information can be displayed below the comboboxes.
The current navigation gets put in a dropdown called “month”
The navigation at the bottom should step through the items that need to be filled out for the selected month. There should be an indicator that shows progress through the list (maybe a progress bar and a “2 of 10” label)
Selected columns unselect after coming back from a Financial Data Entry dialog
When changing a budget center, load times take upwards of two minutes!
Should SingletonDataDictionary.getItemData() return a null if data has not been published? Changed this so that it is the case
Default columns should not include Exec Summary.
“Columns” button should be changed to preferences
Status should be the left-most column.
Added Generic Scrollers to the table and column choices. Need to make sure that the ordering from the column list to the way that the columns are rendered is maintained.
Changed SingletonDataDictionary.getItemData() so that it returns a null if no item has been set yet. Checked in and deployed.
Had some trouble getting FLTK to build. Usually I am able to track down binaries for things like this, but in this case, there is a project file in the ide folder which allows Visual Studio to build it easily. A note for the next time I run across source code with no project files is that I should be able to run command line compiles using the .make file.
Still lots of files and no .lib/.exe files in sight… I thought maybe it just wasn’t set up right but when i look at the project property files it properly tells it to build exe/lib files. For some reason all I getting is the object files. I was able to run demo from the IDE but still don’t see the .exe file anywhere.
Finally found where everything built to. Its in the test and lib folders of the root directory.
Lots of cool demos, looks like you can do a lot with it. One of the cooler ones involves fractals. I didn’t know how useful fractals were in making blocky graphical objects more detailed. The code for this example will help me when I try to figure out how to add a gl window within GUI.
The callbacks were a bit confusing to me. The one for buttons accepts two parameters. The first is the function name you wish to call and the second is void pointer to the data. The second parameter is optional and is limited to only passing one void pointer.
The videos also cleared up .make files for me. I tend to use VS for the most part, so the make files are built into the project. I am not sure make files can be imported into VS, but it would save manually setting up project properties every time.
Should be ready to build the UI
Things needed for the UI: Sensor data, current volume(optional), instrument selection(slider/numeric entry), note selection(slider/numeric entry), disable button, menu bar with a save/load function to save/load the current settings
I think these controls will run across the bottom. The majority of the window will be for the openGL simulation.
The DataDictionary will be used to hold/sync the data on both sides.
Change the color of the text box to show if the force is neutral negative or positive from baseline
Something that might be useful for analysis is if we could graph the increase in volume/pressure.
The load takes *way* to long on project status in the Widget.
The default login needs to be removed
The new fonts appear to mess up the financial update section
Had to change the html fore some reason, otherwise I got a blank white page. I was able to put together one that works using the old PPM html.
Brought in debug versions of Flashplayer
Working on developing a new AppEngine that uses the MySql data store functionality
Applied for a datastore account
While waiting for this, I was thinking that it might work to import the spreadsheet into the appEngine, use an in-memory DB to do all the actions, and then write out the database as a table again. Looking at JavaDB(Oracle) and HQSLDB. There are some issues with this approach – big tables/spreadsheets would be bad. There is the mapping between column id and column headers that needs to be worked out. Also, I’m not sure if the 3rd-party jar file issue that Mike warned me about will be an issue, but this may be a solution to that.