I'm starting this thread because a conversation has developed about MIDICar Player and some of the comments are falling on an innocent device, yes, innocent. Although it has its shortcomings, which I too notice, most of the faults, unfortunately, lie with MIDICar Player.

If anyone is using this program then please feel free to direct all problems in this thread or, what would be even more advisable, in the repository on GitHub https://github.com/GSoftwareDevelopment/MIDICar-Player.

patjomki schrieb: The software is running fine with my Midi Car but there are some midi files that are not played perfectly.

Here is an example where some channels are missing.

THUNDER.zip

For misplayed MIDs, please very much describe what is happening - slowed playback, bad sounds, etc. What version of MCP is being used - I always have the latest, I would say developer version with me, which I release as working 🙂

If it's an issue of slowed playback then older versions i.e. <1.6 may have problems with this, as there was a bug with the speed calculation. In the thread about MIDICar (recent posts) I uploaded the developer version in a disk image. It's unofficial, but that's the one I'm working on. Just in case, I'll upload it here, but I don't promise that everything will work in it as in older versions 🙂 you know how it is 😀

PS. I'm just playing this track and it seems to sound very good. Check the dev version.

EDIT: Attached file is an 'unbundled' version of MIDICar Player, i.e. it does not contain a built-in driver. This version should be run with the switch '/D:MIDICar.DRV /AD580' where 'D580' is the address set on the MIDICar jumpers.

Let me remind that the 'bundle' version can be prepared independently by combining two files (without special combinations) First is to be the driver file, then the MCP program file (not the 'bundle' - that is, as it is here in the attachment)

Under Linux, I use:

cat driver.drv mcp.exe >> mcpdrv.exe'

Under windows:

COPY /B driver.drv + mcp.exe mcpdrv.exe

patjomki schrieb: Feature request. Would it be possible to show the played time in minutes and seconds instead of a counter in seconds (and milliseconds) in the lower right corner?

You're not the first to request this, and yet there's a problem with it. You'll have to believe me that I think all the time about how I could solve such a time indicator

Everything is trivial when one song tempo is kept. Set, for example, at the beginning and nothing changes. Then you can calculate the number of ticks - because what is shown in the lower right corner are so-called ticks (according to MIDI specifications) - per second and display the time in minutes and seconds relatively nicely. However, when the tempo changes, the period that falls per second also changes. When that period changes, so does the whole count of minutes and seconds

A great deal of computation is done on the zero side - this is for speed reasons. It's getting harder and harder to chunk 4 bytes next to each other to add, for example, a temporary value that could be used when the rate changes. And as you can see (4 bytes) are 32-bit operations. Atari doesn't like them 😉 because they are slow. Compromises have to be made, and these are difficult.

Important note, related to MID files!.

You can find a whole lot of MID files on the internet - that much is known - but, some of them can be problematic to listen to.

The most common problem encountered is wrong track length. This is not a bug that causes harm in listening, but it does take time. This is caused by end events in MID tracks. These have very large DeltaTime values, resulting in a very long track length overall. For the moment, the way the length is calculated is very simple, and for the time being, it will remain so.

2

Another problem I'm aware of is track stratification. This is where the listening experience can suffer. The cause is very similar to the above, but much harder to detect, as the high Delta event is inside the track(s). Very rare, but nevertheless can be encountered.

3

Errors related to playback speed. Auditory impressions - extended in time 🙂 The cause is the simplified MID playback engine. The little Atari is an eight-bit system, and MIDs are very fond of 24- and 32-bit words. Some tempo settings, give an extreme value (very high, or very low) for the main Timer. Since this Timer is based on the POKEY, there will be POKEY imperfections involved, and the POKEY's non-linearity is its biggest drawback. Here, as much as I could manage, I tried to eliminate this non-linearity, but I didn't want to go beyond certain limits, hence.... it may happen that a MID played on an Atari will sound slower or faster than on a PC. It's worth looking for an alternative MID - there are often such - or trying to adjust the tempo with the '<' and '>' keys. WARNING! It may happen that this will crash the player!

4

Complete disaster. The file is impossible to listen to. It often happens that the MID file is simply corrupted (for the MIDICar Player). I won't help here.

Conclusion: MIDICar Player supports a maximum of 16 MIDI channels. It does not support Ports (multiple MIDI devices) Probably never will. Three things to keep in mind: - I have been learning MIDI for a year now. Just thanks to MCP. - The Atari (small Atari) has a 1.7MHz clock and an average of 4 cycles per command. This computer has what it takes to play MIDs, especially the so-called MID #1 type, where the tracks are divided into independent streams. - I'm not the best programmer in the world 🙂 and I do all the programming myself, as part of the Fun

MIDICar Player dev-140923.

.

Novelties: - track time minutes:seconds view (request from @patjomki) It wasn't easy, but it worked out. For now, only the time from the beginning of the song is available. The 'T' key toggles time presentation modes: Hex MIDI Ticks, Dec MIDI Ticks, Min:Sec (as default) , Off; - when you load a song, next to its name (in the status bar) its size in KB appears (request from @Pancio 🙂 )

.

What's not mentioned

• Playlists By creating a text file with the structure shown below, you can create your own playlist. File structure:

• the file name of the playlist file must have the extension 'PLS'

• one entry is one line ending in RETURN (the last entry does not have to have an end of line character)
• No whitespace characters at the beginning of the line - otherwise, the entry will be skipped.
• Line no longer than 80 characters!!!

• each line consists of a comma-separated ',' in turn:

• the file name including the extension, and if the file is in another directory/drive, the full path must be given

• a long name, consisting only of printing characters:Small, capital letters; digits; punctuation marks (ASCII #32-#127) No invers characters! No longer than 32 characters.
• There may be blank lines (lone RETURN) - such an entry is disregarded.

• Support for long track names for the disk directory

• the name of the file containing the long name information is 'DESCRIPT.ION' This file must be placed in the root directory of the floppy disk, or in a specific directory on the disk.

• the file structure is identical to that of the playlist

• there are two syntaxes inserted in the part containing the filename with the extension:

• any filename starting with a '>' will be visible in the selector as a directory (the appropriate icon and the behaviour of the RETURN key)

• if a lone asterisk '*' is placed in the last line of the file, this will allow missing entries to be read from the disk directory. This may increase the time it takes to load the directory list, as entries are selected for occurrence (eliminating duplicates)
• The option to load long names can be temporarily disabled by pressing the CONTROL key when accepting a new location (change of drive) or exiting or entering a directory.

The contents of the playlist files '.PLS' and 'DESCRIPT.ION' are not checked for file exists! Entries will always appear, but misspelled entries will simply not open. - Simple view* This can be achieved by pressing the 'I' key. The selector is then retracted and the music handling closes in the current directory or playlist. Passing a parameter, e.g. (SDX system)

-MCP MISSON.MID

will start the program in just such a Simple view. - Equalizer control (DreamBlaster S2, there may be limited functionality in others) Use the '[' and ']' keys to adjust the Master Volume. The 'E' key will allow you to toggle the equalizer range, and use the '[' and ']' keys to change the settings of a range. - Remote Control (with MIDI ReMote only) Currently, remote control capability is available ONLY for MIDICar controllers. SIO based controllers are deprived of this functionality (temporarily 🙂 )

.

A word from me This 'feature' is very useful for creating personalised MIDI music floppies, or in the case of the SDX system, where you can name directories with clear names and arrange tracks. The order of entries in the 'DESCRIPT.ION' file is also preserved in the selector (the same goes for Playlists) Unfortunately, you have to make your own long name editor. I make the 'DESCRIPT.ION' file under the ED editor, completely by hand, using the SDX system's 'DIRS' function for large directories, but this still requires a lot of work.

It seems to me that it's worth doing such files on the fly, as it gets messy over time 🙂 If anyone has ideas related to long names then write - we'll consider the pros and cons 🙂

Sorry you had to wait 🙂 I already saw yesterday that you had a problem, but it was too late to think anything. I also have another project on tap and now I found a moment.

Quite interesting that on the 600XL it works and on the 800XL there is a problem - well.... we'll try to do something about it.

First the issue of error 97 - it's an internal player error, not an I/O error but.... don't worry 🙂 It's nothing terrible, except maybe that it doesn't work hihi 😉 The player won't load a file that's already loaded and that's how it manifests itself, unfortunately. I'm curious why it doesn't want to play it(?) Have you tried, after loading the file, running it with the 'X' key? It's a keyboard shortcut from Play 🙂 If that doesn't help, please send the MID file, because maybe there's something wrong with it. Also, please state what the configurations of both computers are, how you're running, from what - the more info, the better it will represent the runtime environment. You can send a private message if you don't want to disclose this information, and openly send the MID file. 🙂

I'll try to work something out, however it may take a while.

O... I saw Rapidus and I was reminded 🙂

A friend once tried to fire up MCP with Rapidus installed. He didn't succeed. However, the problem may lie on the MCP side.

Slightly deviating from the topic, there is a patch for IRQ in MCP - it's about a bug in 6502/6502C processor, where when NMI and IRQ interrupts collide, the former is skipped. The patch aims to eliminate this bug programmatically, however.... The processor type is not checked (yet) which results in installing the patch in processors that do not require it (65C02 and 65816)

I compiled a version without the patch It is not marked in any way in the program (except today's date).If it works, great. You can do the tests with Rapidus enabled in that case - it ran on the emulator for me, but.... emulator vs real hardware is different 😛

Patiomki schrieb: (...) the 600XL the led on MidiCar is blinking and you can actually hear the song while on my 800XL the led isn't blinking and the speakers stay mute.

On both Atari, do you have MIDICar set the same? Are you referring to the address that the $D580 is running on?

If that address is different on either device then it won't walk on those settings that are in the MCP.CMD file. Open the file in the ED editor (included on the floppy) Load it with CTRL+L typing MCP.CMD and change the /AD580 address to the one set on the MIDICar PCB jumpers.

Still a power supply issue, i.e. either the MIDICar is not powered correctly, or .... (worst case scenario) as a result of a badly done power supply, the MIDICar has broken down (let's hope not)

The other option I can't see why it won't play.

This will not be a simple process.

Try changing the address of the MIDICar. Maybe in the $D180 space it will be ok? Try in high addresses, i.e. $Dx80-$DxE0 - the lower ones are usually occupied by other devices (they were created earlier, so they have the right to have first addresses 🙂 )

Don't forget to change the MCP.CMD file ! (see post above) Or if you use SDX then in the MCP.BAT file.

Anyway, check only on the non-working computer (800XL) If it runs on it, check if it runs on the 600XL too. If it doesn't, look for an address where it runs on both computers. You should find one where it will be OK.

At this point, only address conflicts come to mind.

Does the same thing occur in the DEV version of MCP at your site?

You can download the DEV version from this thread. There are two. One with the IRQ patch, the other without. The patch is applicable for 6502/6502C processors, which have a bug related to IRQ and NMI interrupts. Please see if the memory detection problem also occurs in this version, as the emulator indicates the correct amount of memory.

One more thing.

MCP will not work with the AXLON memory extension!

Montezuma schrieb:

PeBe schrieb: Does the same thing occur in the DEV version of MCP at your site?

The DEV version shows even less memory available (64KB)

Yes, it will show less when from the DEV version, the program requires extended memory, i.e. on 64KB it won't run - Sorry, progress has its victims too.

Montezuma schrieb:

PeBe schrieb: MCP will not work with the AXLON memory extension

It has nothing to do with AXLON

I suppose it has nothing to do with AXLON. It's just an indication that MCP doesn't work with this type of memory extension in any release.

Montezuma schrieb: Memory detection issue is related to U1MB settings ( http://atariki.krap.pl/index.php/Ultimate1MB ). As I mentioned before - if SDX, PBI and HDD is disabled in U1MB settings, and SDX is enabled in SIDE2 cart, MIDICar player properly detects available memory. Otherwise not

Detection of extended memory is based on two ways: - for computers without SDX by the way presented in Atariki ( http://atariki.krap.pl/index.php/Obs%C5%82uga_standardowego_rozszerzenia_pami%C4%99ci_RAM ) - and for SDX systems, where memory is detected by SDX routines.

So maybe I'll ask it in a different way. Under SDX, with all optisions enabled, as shown by the MEM /X command. Are there any free banks? If not, the problem could be the RAM-DYSK being loaded during SDX startup. On the other hand, if there are free banks, then the error is actually somewhere in the MCP program loader.

Montezuma schrieb: Do you have U1MB ? Perhaps Pancio could help you with testing (he has an ATARI with U1MB)

No. I do not own a U1MB. I can talk to @Pancio in terms of testing.

Mathy schrieb: Hello guys

Montezuma schrieb: Here is a solution for other users who are not familiar with SDX (like me).

.

It really sucks that MIDICar player is "SpartaDos only". This makes it a lot less interesting for me.

Sincerely

Mathy

.

Hey, but that's not fair. After all, I've posted here floppy disk images where MCP boots from under classic DOS (XDOS). How about instead of complaining that "it sucks because it's only for SDX", post how you'd like MCP to boot? How you'd like it to work?

Outline your vision. SDX is different, but it's like any other system. I'm not forcing anyone to use it, but I'm also not going to give in to those who don't want to use it - because they're blocking progress, and SDX is progressive.

So much for the topic.