Folks,

I have been working for several months in my spare time on a tool that would provide me with an Empeg with a VGA sized screen, hosted on a PC. I had previously tweaked a VNC server and viewer to do this but was not satisfied with the results (it did work).

I have finally arrived at a point where I am satisfied with the stuff I have come up with. I call it "Empeg Streamer and Controller (EMS&P)" (I am not good at coming up with names). You can see details and download it from http://strobos.cee.vt.edu/EMSC .

Neither my son, girlfriend, or I can make it crash or behave in any way other than the way it should. So it is to the point where testing by others would be greatly appreciated. And if you like it you can continue to use it as it is free.

If you run Windows XP and have the curiosity, please download this thing, take it for a spin, and tell me how it behaves. The application can be used to either control your Empeg, or to stream music from it with the same user interface - one designed to display on a VGA size screen and controlled by short and long presses of only four buttons. Network and CPU utilization on the PC and the Empeg are minimal when controlling the Empeg. When streaming, you do of course use CPU to decode the music, and bandwith to stream the fids.

Of course, the advantage to VNC is that it works on basically anything.

This looks vaguely familiar. Deja vu?

- trs

Edit: Ahhh, I'm not crazy. *phew!*

Sure, VNC works anywhere and practically puts your Empeg on the screen, which is great when your problem is having access to the Empeg. But when you desire a different user interface designed for a bigger screen its a different story.

By the way, thank you very much for providing access to the Empeg cross compilation toolchain in your server. After struggling for weeks to set it up in my machine, unsuceesfully, I really apreciate it.

very nice!

What are the chances of comming up with a vitual command console? It would be very cool to have this set up to remotley control the empeg using a touchscreen monitor on the dash and empeg in the trunk or glove compartment.

Thank you. I have no clue as to how touchscreens work, so I couldnt tell. But I assume they are closer to mice than to keyboards in how the provide input to the application, so it would probably require some work (this thing will allow you to navigate menus with the mouse but it is really designed for press-button use). For now it is necessary to stick to a (steering wheel) ir remote.

This looks great. Just ordered a IRDA receiver to use with my SWI-X. Can't wait to try it out.

No! You are not crazy.

This EMS&P is what I know use instead of the VNC. In addition to a change in aesthetics (I now use margins ) There are many other differences, all of which stem from where the "state" of things is maintained. In VNC, the client is just a viewer of what is happening in the server machine. The client sends keystrokes and mousemoves for the server to interpret and do stuff with. EMS&P uses the Empeg as a music database and as a sound card of sorts (or WinAmp if you are streaming). All state is maintained by EMS&P. It is EMS&P that maintains the running order; enqueing, inserting, replacing, searching, etc. The menu interaction is based on information contained in EMS&P, so it can present nice menus with extra large characters that are not scaled bitmaps of what the Empeg would show. (For some reason I can't capture the screens when showing the menus, would appreciate any help on how to do that...)

EMS&P actually has running order manipulation features that are not present in the Empeg, such as "No more songs like next one...". What is completely different is the search facility, which can do so by searching for matches inside strings, not just at the beginning. Unfortunately, the screen shots of the searching windows are at the end of the web page, so you may have not gotten to see them.

Chances are that if you like the Empeg, you'll like this thing as your desktop music player. At least if you like the Empeg for the reasons I like it, which relate to how I can dynamically choose the music I listen to. If you're into visuals and those kinds of things, then EMS&P is not there yet...

Edit: the VNC mentioned here (http://strobos.cee.vt.edu/EmpegVNC) is one where the state is transfered from the server to the client as information, rather than screen copies. But of course, this could only be done for some things, not the menu interaction.

Just curious, are you going to have a version of this that doesn't require the empeg. (not that I would ever think of getting rid of my empeg )

It wouldn't be too dificult.... I personally don't need one since I've got three of the beloved Empegs. But if there is a demand for it...

By the way, I don't know if you have tried it already and fooled around with the search screens, but I initially posted an executable that starts out with the letter "A" already typed in the search string. There was only one download before I realized this and posted a corrected executable that starts out with an empty search string. Perhaps it was your download?

the app is awesome. im enjoying it so far
a couple things
in control mode, if i bounce really fast from song to song for an extended amount of time it tends to restart my empeg (freeze it, then restart)
also, is it supposed to skip a track when i hit left or right?
if im playing, say, track 39 and hit right it jumps to track 41
if i then hit left it skips back to 39
no matter how much i try i can not get it to hit 40 (or any other even number)
the only way that it hits an even number is if i let it play through a song, but then i cant get any odds
if i hit left, it will never restart the song. just skip back two songs

Thanks for trying EMSC out!

The first thing you observed about rebooting the player when you bounce real fast between songs for extended periods happens because the Empeg player app freezes when overwhelmed with a back log of serial port commands to play. I put in a check in the software to detect this, which forces the Empeg to quit and restart and allows things to continue. I struggled quite a bit with this and that's the best solution I've been able to come up with so far. I don't know if you were doing this as part of a very much appreciated testing, or whether you just like to do that. If you just need to do this from time to time, a way to avoid it, and to switch songs way faster, is to Pause first. Then do next song, prev song, in fast succession (you can just leave the button pressed if it is paused); when you found the song your looking for unpause...

Regarding the right and left key behaviour, the intended behavior (and what I get in my empegs) is that pressing right moves you to the next song, whether it is even or not. Pressing left moves you to the start of the current song if you are past the first four seconds, otherwise it moves you to the previous song. I would like to investigate and correct this behaviour you are observing. Does this also happen in Streaming mode? Does it happen in control mode if you pause before pressing right or left, and then unpause after pressing right or left (n times)? How distant is the Empeg from the computer controlling it (e.g., is the travel time of network packets large)? How fast is your PC?

Thanks again for trying it out and helping me with the testing.

I forgot to ask if you were using the keyboard, or an IR Remote control. If so, does it happen with either?

How distant is the Empeg from the computer controlling it (e.g., is the travel time of network packets large)?

Surely the packets travel at the speed of light, making the length of his network lines insignificant. No?

--Dan.

Sure, I guess I should have used some other term instead of "distance". It is not the same to interact with an Empeg connected directly to your computer, than it is to interact with an Empeg in your office, from home, connected over a dial-up line, during hours of peak traffic. I really mean the effective speed of the link between the Empeg and the controlling PC.

The first thing you observed about rebooting the player when you bounce real fast between songs for extended periods happens because the Empeg player app freezes when overwhelmed with a back log of serial port commands to play. I put in a check in the software to detect this, which forces the Empeg to quit and restart and allows things to continue. I struggled quite a bit with this and that's the best solution I've been able to come up with so far.

If you just need to do this from time to time, a way to avoid it, and to switch songs way faster, is to Pause first.

I'm sure you thought of this, but why not have the empeg pause when it detects this instead of restarting. I'm guessing it's because you can't get the pause command to be put at the head of the queue, but it's worth pointing out if you just missed it for some reason.

Thank you for the comment. When the empeg player app freezes, it will not do a thing. I can't pause, resume, or send a cmd to play anything. In fact, 80% of the time what appears as a freeze is an exit of the player application (if you have serial terminal connected you would see the shell prompt). At that point I send an "exit" to the serial port, which restarts the player exactly where it left off.

When this freeze happens is unpredictable, and something that will very rarely happen if you are just using the application (as opposed to testing it and trying to make it fail). At least in my machines. For example, right now in order to make it happen, I had to press the right button 217 times over a 28 second time frame. I would like to know if it happens in other machines under what would be considered "normal" use.

I thought about doing a Pause behind the scenes prior to a next or previous track command, and a Resume once I detected that no previous or next track button has been pressed for, say, 3/4th sec. But then that meant that I would at least have that lag when changing songs. (The application doesn't inform the Empeg of change in songs while it is paused, only of the then current song when it resumes).

Wow.. I'm putting together a car-pc in the next couple of weeks, and this looks like just the ticket to tie my empeg into my system. I'll try it out and let you know what I think.

I'm guess that the empegLink binary doesn't work with 3.0 a7.

empegLink: connection has been accepted
could not open /empeg/var/database to determine date
empegLink: waiting for a connection

V3's database file is called database3

Got the program running (XP home edition) but can't get certain things to function. I can get the player into search mode and get the menus up. I can also select menu items. I can't unpause my player to get it started, also cannot cursor to letters in search mode. I'll read your webpage more carefully to see if I'm doing things correctly.

Okay, when I select playlists, no lists come up. If I enter search mode all the letters are darkened. I don't think I'm actuallu interacting with the Empeg. The only thing I'm wondering about so far is: could fidsort somehow interfering with the program. I did install FIDSIFT.SH, don't know if that would be a proble or not. Don't know enough to be sure, but though I'd throw it out there as the only thing I can think of so far.

Thank you for letting me know. I have never worked with anything but 2.0 final. So I have no clue as to what is going on. There is a lot about the Empeg side of things (EmpegLink) that relies on Fids, and certain information layout in empeg_state, hda3 and empeg_notify. EmpegLink doesn't actually look at the content of the database file except to look at its date and use that to determine whether EMSC needs to rebuild its database.

Could you please copy Database3 to Database (or just create any file called Database), so that EmpegLink can find it and get a date, and let me know if the other things work by doing this under 3.0? If it does work I'll have EmpegLink look for database3 if it doesn't find database.

If the layout of dev/empeg_state, dev/hda3, and /proc/empeg_notify has changed... then I'll need to work a lot to make this compatible with software version 3.0...

Thank you for both observations. When you first ran EMSC, did you see a flying list of all your playlists in the middle of the screeen, and a count-up of all your tunes in the title bar? This is a process that takes 2 or 3 minutes with my 5,000 songs. Also, you can check the size of the file named "Empeg.your.ip.address.PlayerDB" in the directory where you installed EMSC. If it is zero length, then EMSC is not being able to recreate your playlists from the content of the FIDS in /drive0/fids and /drive1/fids.

Based on what you are describing, all of the above is happening. The reason all the letters in the search are dimmed is because there are no character choices available that would refine your search further, since EMSC doesnt know about any of your songs. There is probably a single match in the search window that reads "all 0 songs in 0 matches" when you start out.

I don't know what fidsift.sh is. EMSC expects your FIDS to be either in /drive0/fids/ or /drive1/fids. I believe that if you use the new organization of fids with more subdirectories, HIJACK will take care of things and make applications think that they are available from one of those two locations.

One more thing, when you are in the main menu, does the "up" key allow you to navigate up in the menus? Cause the Pause UnPause not working is probably just that EMSC doesnt know about any of your songs.

Don't give up on this. Well make it work.

OK. I think you were right regarding the Fidsift.sh thing. I searched for it and found a post by Mark where he said the automatic translation worked only through /empeg/fids0/ and /empeg/fids1/, and not through /drive0/fids and /drive1/fids which is what EmpegLink was using. I recompiled using /empeg/fids?/ and it works the same for me (i dont use fidsift.sh). But it will probably now work for you as well. Please check it out. The new EmpegLink is attached. I also updated the distribution files. Hope this change doesnt break anything....

Edit: Important - Before running EMSC again after installing the new EmpegLink, make sure you delete the zero length "Empeg.your.empeg.ip.addr.PlayerDB" file that was created in the EMSC directory, or that you run EMSC with the "-r" option.

Ok, i just copied my database3 to database. The EMS&C program loaded this time, loaded up exactly 2400 of my 7000 tracks, then locked up.

r7 : 0200006e r6 : 00000000 r5 : 00000000 r4 : 0200006e
r3 : 00000000 r2 : b7340200 r1 : 00010000 r0 : 0200006e
Flags: NzCv IRQs on FIQs on Mode USER_32 Segment user
Control: C0CB517D Table: C0CB517D DAC: 00000015
Function entered at [<4006ac24>] from [<4006adf4>]
Function entered at [<4006add8>] from [<4006b060>]
r10 = 400FF128 r8 = 00000000 r7 = 0200006E r6 = 00000000
r5 = 00000000 r4 = 0200006E
Function entered at [<4006afe8>] from [<4006456c>]
r4 = BFFFF8C8
Function entered at [<40060d78>] from [<40064e1c>]
r10 = 400FF128 r9 = 02001AC4 r8 = 02003104 r7 = 0200083C
r6 = 00000001 r5 = 00000004 r4 = 00000000
Function entered at [<40064e08>] from [<40064e3c>]
Function entered at [<40064e24>] from [<020015a8>]
Function entered at [<020013ac>] from [<020024c0>]
r4 = 00000000
Function entered at [<02002084>] from [<02001eac>]
r6 = 4000C2A8 r5 = 00000001 r4 = BFFFFE74
Function entered at [<02001ac4>] from [<40037040>]
Function entered at [<40036ee4>] from [<02000ac8>]
r10 = 4001D858 r9 = 00000000 r8 = 00000000 r7 = 00000000
r6 = 02000AA4 r5 = 00000000 r4 = 4001E5EC

I also had a couple of these before it crashed, but hyperterm won't let me scroll up at all without scrambling all the text

Could not open Fid datafile, probably bad FID:y Erik Stahlman (erik@vt.edu)
/drive1/fids/40c1

Could not open Fid datafile, probably bad FID:, ATA DISK drive
/drive1/fids/6a31x000-0x007,0x038

Thanks Jules, I'll try that out and let you know. I have some other thoughts but I'll way and try your suggestions first.

Jules, it is working better. The only thing now is that it only loaded 3 songs out my entire database. When I do select one of those songs all the functioons do work however. It's getting close ! My alternative solution is to remove fidsift and see what happens, but if you'd like to straighten this out with fidsift than I'm more than willing. If so, just let me know what you want to try next. (FYI: I reloaded your 2 disk verson of empegVNC, just for kicks, and it still doesn't work. I have a feeling that this is a related problem).

Rtundo,

Thanks for trying. I just realized that I accessed the FIDS in two locations of my code, but only fixed on of those. So after the fix I was able to access the tune fids but not the playlist fids. Here is yet another EmpegLink file that will hopefully completely solve the problem.

Your reasoning with regards to VNC is correct. The code that looks for fids is the same as in EMS&C, althought I did not make the last modifications to that one. Will correct and upload later today, will let you know. But hopefully EMS&C will work this time with the newest EmpegLink, attached.

Please note that the EmpegLink file attached to the post I am replying to, only half solved the problem. (could not edit that post) The EmpegLink file attached to the previous message (dated and timed just a few minutes before this one) hopefully does the trick for those running siftfid.sh.

Oliver,

I wonder whether the 3.0 software uses a fidsift.sh scheme by default. Could you please give it a try with the newest EmpegLink I have posted (on a reply to Rtundo in this thread). Please make sure to erase the "empeg.xxx.yyy.zzz.ttt.PlayerDB" file that was previously created before running with the newest EmpegLink Thanks.

P.S. I might try the 3.0 software later on

Great Program! One thing I did notice... I could not get the app to play files on the Empeg. The interface acted as if it were playing, except the counter was not counting and no sound. I could scroll through songs, playlists, but the unit was permanently paused. The streaming did work however. What I found was that my player was set to the "Aux" input instead of the player. You may want to automatically switch back to player mode when the app boots. Just a suggestion...

JeffreyB

JeffreyB, Thank you for the comment and the suggestion! I have done as you suggested. When EMSC starts up in control mode it will switch the source to Player in case it was in aux or tuner. The attached EmpegLink will do this, and it has been incorporated into the distribution files.

Jules,

I'm running 2.0 final, with Mark Lord's fidsift , so all my fids are in subdirectories of /drive0/fids and consequently EMSC can’t find my fids.

Jules, looks like that did the job. Great program, can't wait to install it in the car. I've been playing around with it and think its very intuitive. I noticed one thing, after I select a few letters in search if I decide to stop the search and exit, I can't find a way to go back out of search (you have to finish the selection). All in all I'm very impressed. I'll give more input as I use it more. Thanks.

Joe,

Rtundo had that same problem and we were able to update the Empeg side of things to work with fidsift. You probably downloaded EMSC before we fixed that. The version you can download now will look for the fids in /empeg/fids0/ and /empeg/fids1, and even if they are in subdirectories of /drive?/fids they should be found. Thanks for trying it out.

Rtundo,

Well, thanks to your testing we took care of coexistence with fidsift.sh.

To get out of a search, use a longpress of the UP button (if the focus is on the group of letters). If you are in the bottom pane (where the matches are listed and you can select one), the LEFT button will put the focus on one of the letters and then you can longpress UP to exit.

Edit: The original post incorrectly indicated that pressing the RIGHT button will put the focus on the top part of the search window, when in fact it would execute the selected match.

Ok, i just updated my copy of empegLink on my player, deleted the local database on my pc, and tried to sync the database again. Exact same issue.

I got a bunch of these errors in hyperterm, what is really weird is i only have a 1 drive empeg, so everything should be /drive0 i would think?
Could not open Fid datafile, probably bad FID:
/drive1/fids/50c1
Could not open Fid datafile, probably bad FID:
/drive1/fids/5dd1
Could not open Fid datafile, probably bad FID:
/drive1/fids/5de1
Could not open Fid datafile, probably bad FID:
/drive1/fids/5df1
Could not open Fid datafile, probably bad FID:
/drive1/fids/69e1
Could not open Fid datafile, probably bad FID:
/drive1/fids/69f1
Could not open Fid datafile, probably bad FID:
/drive1/fids/6a01
Could not open Fid datafile, probably bad FID:
/drive1/fids/6a11
Could not open Fid datafile, probably bad FID:
/drive1/fids/6a21
Could not open Fid datafile, probably bad FID:
/drive1/fids/6a31
Could not open Fid datafile, probably bad FID:
/drive1/fids/7701

With the final error on track 2400,
r7 : 0200006e r6 : 00000000 r5 : 00000000 r4 : 0200006e
r3 : 00000000 r2 : b7340200 r1 : 00010000 r0 : 0200006e
Flags: NzCv IRQs on FIQs on Mode USER_32 Segment user
Control: C0CB517D Table: C0CB517D DAC: 00000015
Function entered at [<4006ac24>] from [<4006adf4>]
Function entered at [<4006add8>] from [<4006b060>]
r10 = 400FF128 r8 = 00000000 r7 = 0200006E r6 = 00000000
r5 = 00000000 r4 = 0200006E
Function entered at [<4006afe8>] from [<4006456c>]
r4 = BFFFF8C8
Function entered at [<40060d78>] from [<40064e1c>]
r10 = 400FF128 r9 = 02001AC4 r8 = 02003104 r7 = 0200083C
r6 = 00000001 r5 = 00000004 r4 = 00000000
Function entered at [<40064e08>] from [<40064e3c>]
Function entered at [<40064e24>] from [<020015a8>]
Function entered at [<020013ac>] from [<020024c0>]
r4 = 00000000
Function entered at [<02002084>] from [<02001eac>]
r6 = 4000C2A8 r5 = 00000001 r4 = BFFFFE74
Function entered at [<02001ac4>] from [<40037040>]
Function entered at [<40036ee4>] from [<02000ac8>]
r10 = 4001D858 r9 = 00000000 r8 = 00000000 r7 = 00000000
r6 = 02000AA4 r5 = 00000000 r4 = 4001E5EC

That did it, thanks. Two requests to make. If you someday would add "mark track" as a menu function that would be great. Also, a pause sign of some sort would be nice. Your program rocks. The people on this BBS never cease to amaze me.

Oliver, I doesn't seem like you updated the EmpegLink.

Your messages like:

Could not open Fid datafile, probably bad FID:
/drive1/fids/50c1

are from the old EmpegLink. The new EmpegLink I mentioned would have said that it couldn't open Fid datafile " /empeg/fids1/50c1" and not the one with "/drive1/fids/50c1".

You see bad attempts at opening FIDS in drive1 because opening the same FIDS in drive0 failed. EmpegLink simply tries drive0 (actually fids0) and if it is not found there, it tries drive1 (fids1).

Before updating EmpegLink, you need to make sure that the EmpegLink that is currently on your player is unloaded, so it that it can be replaced. To Unload it, you execute EMSC from you computer with the "-t" option (this is in the EMSC web page). For example: "EMSC your.empeg.ip.address -t" from the Windows command line, or through a shortcut. This unloads EmpegLink from your Empeg's memory, and allows you to be able to replace it with the new one. If you replaced EmpegLink by running the provided LoadEmpegLink.cmd, you would have seen some FAILED responses when FTP tried to PUT the new EmpegLink.

All that said, I have read many of the posts regarding the version 3.0 alphas. And many users experience all sorts of problems, even just using the official software (e.g., Emplode). So, although I guess that replacing the EmpegLink may solve the problem, I don't know that it actually will, and if the problems are not solved, I can't tell whether they are due to the instability of Version 3.0 alpha, or to a redesign of empeg_state and hda3 with the new v 3.0 alpha.

I have three empegs, but I am hesitant to install v3 alpha due to the issues about it that have been posted...

Do you have fidsift.sh in addition to V 3.0 alpha x?

Rtundo, great that everything is working and that you like EMSC. The search method gets a little "getting used to", but it should be quite effective once you get the hang of it.

Regarding your requests, I've been lazy about the Pause indicator (which I have experienced the need for as well), but I'll try to put it in soon. I never mark tracks, but it is obviously a useful thing and I'll try to add that feature as well. Thank you for the suggestions.

Awesome

Two new features are included in the EMSC package I have just uploaded to the website (http://strobos.cee.vt.edu/EMSC).

1) Visual indication of Pause/Play status.
2) Easier upgrade of empegLink. The new LoadEmpegLink.cmd will edit config.ini for you (new installations) eliminating a re-sync via emplode. The new LoadEmpegLink.cmd will also automatically unload a currently loaded empegLink (upgrade installations) before trying to install the updated version.

I've received a lot of feedback and corrected some issues. I'm still waiting for further feedback from a coule of people before changing status from "alpha" to "released" (no apparent need for betas). The latest delivery (1:50 am EST 4/20/04) now includes Scaled Window Content when in Full Screen. So if you have a 1024x768 display (common for a car), the full-screen mode will look nice and scaled rather than occupying only the top left 640x480 with info.

Featurewise, the first release of EMSC will consist of what is now available and described at http://strobos.cee.vt.edu/EMSC . I will really like to correct any errors that may still remain, but do not want to get into issues related to version 3.0 alpha of the car player software. So please let me know of any issues that have not already been addressed with the latest available version.

The following are the target features for the next version:

1) User-configurable named screen layouts, which can be activated dynamically. Right now there is only one layout to choose from, and it is permanently shown.

2) Support for the equivalent of Flags & Filters, except that these will be dynamically changeable from EMS&C.. "Marked" will work by simply creating a "Marked" Flag, and setting this flag on/off for tracks when you are listening to them.

3) Ability to refine the current running order by using criteria similar to those of searches (in addition to Empeg-Like flags and filters).

4) Ability to see detailed track info (VBR rate, codec, etc)

5) Maybe: supoort for mouse operation for compatibility with touch-screen displays. (I don't have a touchscreen so motivation is low.....)

This whole project should in time (not too long) result in something that will make any Empeger using a car computer very very happy.... Suggestions for the next version in addition to those listed above are very welcome....

excellent!! just downloaded and installed. it crashed towards the end of the cataloguing process but seems fine now. Being a shallow and superficial person I am very much looking forwards to custom skins ;-)

I've had no problems with it so far. I'm waiting for for an IR interface I ordered before putting it in my car. Using it with my steering wheel controls will be the ultimate test for me. Oh, by the way, thanks for the pause indicator update

What IR interface are you getting? With what remote will you use it?

The step i missed the first time was the -t option to upload empegLink.

After removing and upgrading to the latest release of EMSC. Then connecting with the -p switch. It got past 2400 tracks this time, didn't give me any can't find fid errors. But then at 2487 it bombed again.

empegLink(11): memory violation at pc=0x4006ac44, lr=0x4006adf4 (bad address=0xb
r6 = 00000001 r5 = 00000004 r4 = 00000000
r7 : 0200006e r6 : 00000000 r5 : 00000000 r4 : 0200006e
r3 : 00000000 r2 : b7440200 r1 : 00010000 r0 : 0200006e
Flags: NzCv IRQs on FIQs on Mode USER_32 Segment user
Control: C0CB517D Table: C0CB517D DAC: 00000015
Function entered at [<4006ac24>] from [<4006adf4>]
Function entered at [<4006add8>] from [<4006b060>]
r10 = 400FF128 r8 = 00000000 r7 = 0200006E r6 = 00000000
r5 = 00000000 r4 = 0200006E
Function entered at [<4006afe8>] from [<4006456c>]
r4 = BFFFF8C8
Function entered at [<40060d78>] from [<40064e1c>]
r10 = 400FF128 r9 = 02001AC4 r8 = 02003110 r7 = 0200083C
r6 = 00000001 r5 = 00000004 r4 = 00000000
Function entered at [<40064e08>] from [<40064e3c>]
Function entered at [<40064e24>] from [<020015a8>]
Function entered at [<020013ac>] from [<020024c0>]
r4 = 00000000
Function entered at [<02002084>] from [<02001eac>]
r6 = 4000C2A8 r5 = 00000001 r4 = BFFFFE74
Function entered at [<02001ac4>] from [<40037040>]
Function entered at [<40036ee4>] from [<02000ac8>]
r10 = 4001D858 r9 = 00000000 r8 = 00000000 r7 = 00000000
r6 = 02000AA4 r5 = 00000000 r4 = 4001E5

any other ideas?

I have no clue. Suggestions? User player version 2.0 final developer with Hijack 380. That works flawlessly.

The condition of Player version 3.0 alpha is such that it would make it futile to go beyond what we already have tried. It may very well be that empegLink has nothing to do with the behaviour you are observing. If even the Emplode versions designed for Player 2.0 bomb when using Player software 3.0, then I wouldn't expect more from EMSC-empegLink.

The IR receiver is from zapway (link on winlirc site). The transmitter will be PAC's SWI-X (described in faq section I think). With your program it should be easy to navigate using four steering wheel buttons leaving 2 buttons for other things.

Great app Jules,

A request; an option when terminating program to put empeg in standby instead of pause.

And if you decide to let more of the board try it, (non-XP users) a Java version.

Thanks,

Thank you Joe.

Your suggestion is a good one (about stand-by). Will try to see if I can do that programatically. Regarding non-XP version, it is something I would like to do but unfortunately do not have the knowledge to implement. I wonder if WINE for Linux and the equivalent for Macs will run EMSC.

Jules,

I've yet to Linux myself, however it is my understanding that Wine emulates 98's environment.

Thanks,

Apparently the new phatbox/keg interfaces with video screens and is capable of displaying album cover graphics. Not really a "request", just a "thought" to put in the back of your mind maybe for some future rainy day .

Jules, do you know of anyway to setup a Winlirc mapped button to reload EMSC in case the quit command is selected during car use and no keyboard is available? I guess the next best thing would be to have a computer reset button somewhere up front.

Hey, sounds like you got this thing up and running in your car!

Obviously, EMSC can only interpret stuff from WinLirc if it is running..., so you need additional sofware independent from EMSC, such as Girder and some others listed in the WinLirc site) that allow you to perform certain actions (start program x, shutdown the machine, etc.) when you press certain buttons. You can use those no problem in coexistence with EMSC. Perhaps Quit, when chosen from the menu instead of pressing the q in the keyboard, should ask for confirmation...

It will be in the car tomorrow. Can't wait. Have it hooked to my TV in the meantime. I'll take a look at one of those programs to set up basic Windows macros. Thanks. Oh yea, I think confirmation before exiting would be a great idea.

It looks like IRAssistant can be deactivated when EMSC launches allowing the same buttons to be used for both programs.

I have it installed in my Saab. Still ironing out problems with my SWI-X remote adapter though. One bug I've run across:
When the program first comes up, the pause indicator is on but the song is playing. Hit the up button once and the play indicator comes up, hit the up button again and the player pauses and the song pauses and things stay in synch.

One future request would be to have a way to lock the volume at a level in the EMSC program. Other than that, things are working great.

Re Pause/Playing sync, does this always happen, or just sometimes? Do you use that empeg with anything other than EMSC (this shouldn't matter, but knowing it may help find the problem)?

It would be cool to see a picture of your install with EMSC running full-screen.

I'l check out when the conditon occurs. I believe it only occurs when the program is loaded during computer boot up. I'll be better able to confirm this when I finish the setup (get the steering wheel remotes working) which should be early this week.

I'm not using any other programs at the moment. I plan on using IRAssistant to control windows commands (inactivates itself when EMSC loads) but it isn't loaded yet.

I'll try to get pictures up soon if I can coax my my crappy digital camera to download photos to my computer.

The problem does not occur all the time but I cannot link it to any event.

One other thing I noticed (thought I was crazy for awhile): When starting EMSC, it starts playing the same song everytime. I think this is because I shut down the car without quitting EMSC. In turn the power supply shuts down windows then the computer. Is it possible to have EMSC save the new song position during a windows shutdown sequence (If this is indeed the problem).

By the way, so far I'm very impressed

OK, I just verified that this happens. I have set mine to go into hibernation rather than shutdown so I never observed this behaviour. I'll look into it and tell you when I post an updated version. Of course when you exit EMSC and then bring it back up it should start off exactly where it left off, right?

Exactly, if I exit program first it starts up where it left off, no problems there.

The version of EMS&C now at http://strobos.cee.vt.edu/EMSC will preserve state when the system shuts down (even if you don't quit the application before). This version will also put the Empeg in standby when it exits.

Thanks Jules, I'll try it out ASAP.

rtundo:

Following your observations regarding use of EMSC in the car, and the issues with restarting at the same song and the mismatch between the play/pause indicator and what was really happening, I decided to create the same environment and test. I observed the same things, which allowed me to fix those problems. They were due to the Empeg being put on standby automatically by its power management when the car is turned off. EMSC would shutdown 20 seconds after that (assuming you have an Opus PS), and so the state of the Empeg was in mismatch with the state of EMSC. Creating havoc. I had previously suggested you hibernate instead of shutdown and restart, and that would have actually increased the problem.

The EMSC you can now download should work well in the car. It should work if you set your computer to hibernate (35 seconds) or standby (12 seconds), in addition to shutdown restart (70 seconds).

It should be happier testing for you with this new version than with the previous one I offered 2 or 3 days ago

Jules, thanks for the upgrade, I'll try to download it tonight. Other then the problems I described already I've had no other problems with the program. Me like . The only annoying thing (remote or winlirc related) is that the steering wheel buttons respond slowly or sometimes not at all. Do you know of any settings that can be tweaked that would allow the remote buttons to be more sensitive to a quick touch?

Also, do you know how to setup the computer to hibernate automatically when the ignition power is turned off? It would be nice to use hibernate instead of shutdown.

5) Maybe: supoort for mouse operation for compatibility with touch-screen displays. (I don't have a touchscreen so motivation is low.....)

Instead of touch screen support maybe button codes display of raw untranslated data coming from the serial/usb ports might be another option. For example: someone hooks up touch screen to serial port, touches certain area of screen, reads untranslated value and maps it to some translated button such as "select playlist" etc. In otherwords, something simlar to hijacks mapping functions. This way any control device using serial/USB could be set up relatively easily.
I know you have a large to-do list and only offer this as something to think about.

Might be your remote. I have somewhat slow response time from my Pioneer SR-90 (780 milliseconds for complete reliability), However that remote is notorious for being extra difficult to decode. I set it to 380 and have about 85% reliability (sometimes a long press gets interpreted as two short ones). With the Rio Remote from my Empeg I set it to 250 and get 100% reliability. 250 millisecs is pretty fast, not quite as a keyboard, but theorectically you could press and get proper response for 4 presses per second. These timings are the ones that you set up in the RemoteMappings.ini file. I would try to set up the Rio Remote in Wirlirc and then map it to EMSC and see if the response is acceptable to you. If it is, then you know that that the problem lies with the remote.

To make the computer hibernate, or standby, instead of shutdown:
1) right click on empty space on the screen
2) select properties
3) screen saver tab
4) select the "power" button
5) select the "advanced" tab
6) where it says "when i press the power button in my computer", select hibernate or standby.

I think standby is OK unless youre not going to use the car for several days. My computer is playing music within 12 seconds (including the 5 second delay that the opus PS adds) of turning the ignition.

Jules,
I've put EMSC thru its paces and it performs great. The only 2 issues I have are:

1. Occasionally still starts with pause sign but song is playing.

2. Slow button response from steering wheel controls (probably the settings in Winlirc and not EMSC).

No other problems to report

What exactly is the difference between hibernate and a normal shutdown? Just curious.

Hibernate dumps the content of RAM to the hard disk and then turns the computer off. When you start up again it will load it into memory and you will continue from where you left off. Very generalised description but that's basically it.

When you restart your computer after a normal shutdown you only load the operating system and the programs in your startup folder are started fresh.

When you restart from hibernation, all the programs you had running are still running exactly as you left them. So, for example, if you are writing a document in word and you are typing on paragraph 3 of page 28, then you hibernate; when you restart word will be open with the cursor on paragraph 3 of page 28, even if you didn't save the document.

Jules,

I attempted to run ESMC this evening without luck. It's seems to be crashing at about the 2500 track mark when building the initial database. I am running v2.0 final and hijack v389. I added the config file changes and the EmpegLink file using the CMD file that you supplied. I attempted the load 4 times with similar results. The player has 2 drives (60 & 80gb) with about 1.5 GB free. Is it possible that its crashing on a bad track? If so is there a way to tell which track or the last track read before the crash.

Thanks.

Sorry to hear this is happening. I thought I had taken care of most problems. Are you using the developer version of 2.0 final? If so, can you please connect your terminal to the empeg before running EMSC, and then provide me with the terminal output? Hopefully it will provide hints of what the problem might be.

I am still running on Hijack 380 and never tested with a subsequent version. I doubt that it has anything to do with the problem, but checking with 380 may also make a difference.

Thanks.

Here's the output...

empegLink: waiting for a connection

empegLink: connection has been accepted

Playlist FID 100 has more than 133 direct entries, only the first 133 will be us
ed



Policy and Priority were 0 and 0



Policy and Priority changed to 2 and 50

Hmm. This ouput takes place at the start of the process, before the counter of tunes starts to go up. Correct? Nothing else is produced once the countup is going?

As you saw from the output, EMSC can only deal with Playlists that have less than 133 direct entries, the 134th and on are ignored. At first hand I don't see how this can foul up things other than those entries not being accesible from EMSC (need to check the code when I get home). Certainly shouldn't make the thing crash. But one thing to try, just for testing purposes, would be to break up that top level playlist into parts, such that you dont have any playlist with more than 133 items (which is probably a good idea anyway for the sake of manageability of the music)..

This output happens first:

empegLink: waiting for a connection

empegLink: connection has been accepted

Playlist FID 100 has more than 133 direct entries, only the first 133 will be us
ed

Then the application locks up and the ouput is this:


Policy and Priority were 0 and 0

Policy and Priority changed to 2 and 50

If I understand the playlist issue properly, the output is saying that there are more than 133 playlists in the root playlist? The playlists that I have are organized by Artist/Albums. The only exception is a various artists playlist of nested albums for soundtracks and compilations. Any idea how I can find the offending playlists?

FID 100 is the root playlist. Apparently, it contains more than 133 entries ( a mix of playlists and tunes). You can check by connecting to your empeg using FTP and looking at the file /drive0/fids/101 or /drive1/fids/101. The length field in that text file tells you how many entries it has. Do you have more than 133 different artists? Given the size of your drives that is a possibility. I have more than 133 artists, but I have them subdivided by genre... you can also verify that with Emplode.

You didnt verify to me that the countup of tunes in EMSC is still at zero when you see the terminal output you posted, and then starts going to circa 2500.

Sorry, the first output block in hyperterminal happens before count up. Count up begins and the application crashes around 2500 tunes and the second output block happens after the app crashes.

Yes, I have quite a few more that 133. The 101 file shows

length=2424
title=All Music
type=playlist

Thanks for the help on this. If the answer is to re-organize, I'll do so and get back to you before I test again.

To be able to use EMSC you will need to reorganize, which will probably be a good idea. Looking for stuff among 2424 entries at the top level should be difficult. However, I am not sure that that is the reason for the crash. After "Policy and Priority changed to 2 and 50", isn't there a blank line and then a line with the left square braket "[" ?

Also, when you say the app crashes, what are the symptoms? Does the EMSC title bar change to read "controlling Empeg xxx.yyy.zzz.ttt? Can you grab the window by the title bar and move the window around? Does the name of a playlist appear in the middle of the window? As the number of tunes is counted up, do you see a flying list of your playlists in the middle of the window? It might be that EMSC is not really crashed, what happens if you press the "Down" cursor key in your keyboard?

My apologies. It didn't understand that it was finished. When I closed the application using the close button it apparently was still busy. XP asked to close it anyway and it looked like it crashed. I'll reorganize my playlists and let you know how it goes.

I'm not sure why my player is reporting 2424 entries under the playlists top level. In emplode, it only shows 605 in the next level under the "Playlists" folder.

Thanks for the help.

Well, the 2424 is the size of the playlist data, and each entry is 4 bytes long, so it really means 2424/4=606 entries and not 2424 (sorry), which matches Emplode (606 and not 605?).

Still, do you get a menu when it is finished and you press the down key? This should work, although you will only be able to see the first 133 items.

Yes, I get a menu when I get a menu when I press the down arrow key.

I've been playing with ESMC and I like it alot. Nicely done!

(606 and not 605?)

There'll be a hidden one (FID 110, Unattached Items) in there that emplode isn't displaying, most likely.

Great! Thanks for hanging on and giving it a shot. I guess at some point I'll need to remove the 133 item per playlist limitation.

rtundo,

I have implemented the interaction with WinLirc in a different manner which provides much quicker response than the prior. You should donwnload the latest EMSC and see if that makes your remote more responsive.

I guess a little explanation on what the parameter for "maximum reliable repeat rate (MRRR)" does in the current implementation, will help you tune that value to provide the best response.

IR Remotes send signals only when the buttons are pressed. There is no signal that tells you when the button has been released. The remotes, however, continue to send repeat signals (which can be recognized as such) at a certain rate. Due to the layers imposed by WinLirc, and the complexity of some remotes, these signals reach the application at somewhat irregular intervals. For example, my Pioneer SR-90 is particulary difficult to decode, and WinLirc will not decode a repeat signal every now and then, sometimes several in a row. So, while the interval between repeat signals for this control should be around 180 millisecs, the delay is sometimes 360, 540, or exceptionally even 720 millisecs.

The fact that we want to distinguish between short, long, and sustained presses complicates things when we have this irregularity. (180 is very responsive, buty 540 is sluggish).

Here is what the new implementation in EMSC does:

Receives the first signal from the remote and does nothing. We dont know if it is a short or long press yet. If a repeat signal is not received within MRRR seconds, then EMSC knows it was a short press and takes action. So MRRR determines the lag in time before EMSC can be sure that a button was pressed "short".

If the repeat signal arrives before MRRR msecs have passed, then EMSC knows that it is not a short press right there and then, and takes appropriate action immediately. So even if I you set MRRR to 740, if the repeat comes after 180 millisecs, the action associated with a long press will take place right then without having to wait for MRRR millisecs. If the button is one of those associated with a sustained action, like scrolling a menu, changing volume, or fastforwarding; then every repeat will take the appropriate action as the signals are received.

When more than MRRR millisecs have elapsed after the last repeat, then EMSC assumes that working with that button has ended. However, sometimes a delay larger than MRRR msecs will take place between repeats. EMSC just ignores the repeats that follow because it doesnt want to take a short or long action that wasnt intended. What you will see when this happens, if you are scrolling a menu list for example, is that scrolling stops after a few scrolls no matter how long you keep the button pressed.

So the value of MRRR you choose has impacts in how long you have to wait before a short press can be determined as such. You want the smallest value possible. However, when the actual repeat generated by the remote and decoded by WinLirc is longer than MRRR, then the sustained action you want to carry will stop. In this case, you want MRRR to be larger than the maximum possible repeat rate.

You should set an MRRR that gives you reasonable response time when you want to execute short press actions, but that at the same time doesn't stop sustained actions too frequently (e.g., press down to scroll a menu, actually go down 3 or 4 items, and then no more scrolling until you release the button and hold it down again can be very annoying.)

If the remote is easy to decode and consistent, like the Rio Car remote, then the repeats are pretty regular and shortly spaced. This allows you to use a rather low value of MRRR, obtaining quick response and no interruptions in sustained presses. Check with your Rio Car remote to see this, then compare with the thing you really want to use.

Im pretty sure the new EMSC with the behavior described above and a properly tuned value of MRRR will get you a better experience than what you have been experiencing.

Let me know how it goes.

-Jules

Jules, thanks for the update and the expanation of setting the rates. I'll give it a shot this week, and I'll let you know what happens. Other than button press problems, I've enjoyed your program a lot (No problems have popped up). Also, just curious how (if) things are progressing on moving the song database to the computer.

hey jules cool program just wanted to know if the computer you have running xp in your car are you running all your audio from the computer or from the empeg and if its from the empeg can your run movies files from the computer and see them on a screen and hear it through the car speakers. Basically is it possible to run both empeg and computer audio on only the empegs audio connections

You can use an RCA switch (controlled by one of the units remote turn on wire) to have two sources hooked up to a single amp. Whenever the one designated as master is active it is switched through to the amp.

good reply but (to Jules) if you have a computer hooked up to the empeg in your car (maybe) and u wanted to watch a video how would you hook up the audio (question to jules)