mAirList startup time/OnTheFly cache building

My techie colleagues at Leith FM are concerned at the length of time it takes mAirList to start up compared to products like (for example) PC-DJ. On a PC containing about 20,000 tracks in ‘on the fly’ directories and their subdirectories, it can take several minutes to go from click the startup icon to having mAirList ready to play tracks. They are (rightly) concerned that this is much too long a startup time for a live, on-air playout system; and we are likely to have (at least) 50,000 tracks in our real library.

PC-DJ does the equivalent of building/updating an OnTheFly cache at startup, BUT it does this as a ‘background’ task so that it starts up very, very quickly while it continues to build its list of tracks/artists/other info for anything up to about 10 minutes. Would it be possible to change mAirList initialisation so that the GUI etc. has ‘priority’ over the build of the OnTheFly cache(s)?

I do realise that it is possible to set an option to not build an OnTheFly cache at mAirList startup, but this would mean any recently added tracks would not appear in the Search (Catch-22!).

Ideally mAirList would be available for use in less than 60 seconds from clicking its icon, but it would still be possible to (for example) quickly add some tracks 15 minutes before a show and have those appear in the Search/OnTheFly cache, even if mAirList has been up and running for several hours or days.

Again, I do know that a ‘proper’ mAirList database is planned, but will likely not be available for some time yet.

PS: My people are also of the opinion that we can only use the Production version (2.0.x) of mAirList rather than the current Development release. I understand what they mean, BUT … !


With Cached=true, mAirList reads the cache file, scans the directory for new files, and adds them. This should be fairly quick, even with 20.000 files. With Cached=false, mAirList will import all files again into the virtual database. This should take significally longer than the rescan.

Just be aware that the OTF databases are not designed for that number of songs. It’s rather a task for a real database.

Unfortunately, the database interface does not support concurrency or multi-threading at the moment.

Regarding the v2.0 vs. v2.1 discussion, you know that the v2.1 versions are very stable, despite the “development” status.

I hope we’ll have a stable v3.0 early next year.

Our people at Phoenix FM have been asking about a search facility (I admit that was on the only good thing about Jazler!), so I read this topic and had a play with the OnTheFly database and for our library of 8000 or so tracks, it’ll be fine - it seems to load fairly quickly on my PC, so I’ll go live with it at the station tomorrow.

As for versions - we’re running v.40 and had no problems.

Torben: yes, all agreed and understood, but I have a techie whose avowed view is that we should only be using the “stable” or production release (and for any software other than mAirList, I would agree ;)) >SIGH<. Hence, the sooner we have a new “production” mAirList version incorporating all the great features added since 2.0, the better.

Charlie: Yes, I agree that the DB is about the only good thing about Jazler (their support, both pre-sales and post-sales, is, er, legendary ;); and is one of the main reasons I was able to persuade our lot NOT to buy Jazler … !).

I will attempt to persuade them that using mAirList in combination with the (paid-for) MediaMonkey Gold version will be the way to go. They can then use MM to do track/year/genre searches etc. and then drag the found track(s) into a mAirList playlist (I’ve tried this, it DOES work!). The reason being that no-one on our station uses pre-compiled playlists, and we do need the ability to find tracks quickly, to be able to satisfy listener requests and presenter whims. Sure, it’s a compromise, but unless anyone has a better idea, I can’t think of a superior Very Low Cost solution (MM Gold licence costs $19!) until mAirListDB is a reality.

So my idea is to run mAirList and MM Gold on the Playout PC, so if someone wantsto find a 1970s track, or anything with ‘jane’ in the title, they use MM Gold. On the flip side, we use mAirList so that we still have truly professional playout software, and we can still ‘push’ ad. break playlists on to the studio playout PC’s mAirList from the production office.


OK, I set up the Phoenix FM machine today to load 4 OnTheFly databases on Startup, plus the main audio tree of folders - all saved in standard.mlt. Total files for the databases were around 5000 tracks. First loading took about 2.5mins … Subsequent loads (with no file changes) took around 12-15sec.

We’re really missing the ability to search via Years at present, so I think Cad’s suggestion may be worth a look - altho people are getting a bit confused at the station as Jazler had all kinds helpful search features - but sadly, it’s only main plus point.

As an aside, when I was at Big L in 2005, we used it on a touch-screen and it was so-so, but it was just really clunky and not very slick - you couldn’t set-up segues and PFL with Jingles over song intros etc. It was just “click and go” ;(

Are the years stored as attributes? It would be fairly easy to include the attributes into the search.

They’re not as yet - but pretty much every track has an ID3 Year tag. I haven’t looked into adding ID3s as Attributes, but I think that’s something we can do, right ?

Yes. And remeber that the ID3 tags are never imported again once you have a MMD file or OTF cache file etc.

They can then use MM to do track/year/genre searches etc. and then drag the found track(s) into a mAirList playlist (I've tried this, it DOES work!).
Wow! It didn't occur to me to do this but I have tried it and as you say it works a treat.

It really shows how much more powerful mAirlist will be when it has this enhanced search/database functionality!


Sounds like I have missed a trick here. ???

Are you saying that an ID3V2 tag field like Year can be included as a custom Attribute in mAirList? If so, how is that done exactly? Is it as simple as going into Config. and adding an Attribute named Year? I think that adding a facility to search by custom Attribute would be a great idea; there seems little reason to have custom Attributes if you can’t search by them!

Even so, my issue is that mAirList cannot scan for new files while it is running. If we are running mAirList 24/7, and add new MP3s to our network library drive/directories, mAirList will never ‘see’ these added files, so the only way to use them is to either a) open a directory tree on the drive/folder, or b) use external software like MediaMonkey (or even Explorer! :D) and drag/drop the files from there on to a Player or Playlist. The lack of drive ‘monitoring’ is the real issue.

We can’t afford to stop/restart mAirList every time we add tracks, since as I said earlier, presenters will often come in and add a bunch of files 30 mins. or so before they go on air. We need a solution which is constantly monitoring the drive(s) for new files and updating itself, which is why MediaMonkey immediately came to mind. I would of course like to know if anyone can suggest something they think is better than MediaMonkey for ‘categorising’ and searching for music tracks, and which constantly monitors drives for changes.


Just go to the Attributes tab in the Properties Dialog and enter as many name-value pairs you like. Press the down array key in order to add more lines to the table. In the config, you can define a list of attributes names mAirList will propose for any newly created item, so you don’t have to type the name manually each time. For each attribute, you can also specify a list of values (semicolon-delimited) mAirList should offer as a drop down list.

Searching these attributes in the OTF database is not possible yet, but I can enable it of you like.

You can do a rescan of the OTF database as follows: Open a database browser (the one with the tree) and right-click the database icon. You will find a “disconnect” option. Right-click the database again, and select “connect”. The OTF database will reconnect, and a rescan will take place.

If you are concerned about the time needed for the rescan, I propose that you use two distinct OTF databases/folders. A big one for your basic archive, and a small one for files added by the presenter in the short-term.

Of course, “real” databases just like the upcoming mAirListDB do not have this problem. You can just add and delete files with the management software, and mAirList will instantly display them.

I think this may have already been asked but is it possible to have an OnTheFly database point to multiple folders and search from them all from the same search button?

Kind regards Tony

Just add multiple OTF databases to the config. They will appear as distinct databases in the database browser. In the database search browser, you can search all of them at the same time.

Cheers Torben.

Tried this and it adds drop down databases to search through.

Is it possible to search through all folders without using a drop down search of this nature?

An artist such as Cliff Richard may be in several different folders if sorted by decades. As an example you may just want to list all his singles and select two for a then and now feature.

kind regards Tony

Take a closer look at the top-most item of that drop down box: “all databases”.

It searches all databases at the same time, gathering the results in a single list.

Wow, I had to include a large folder (about 30GB worth of live sessions, album tracks etc) today and it took an hour to cache, reduced to about 30s on subsequent loads. The “remove splash screen” is really helpful because we loaded a 2nd copy of mAirList whilst the first was caching in the background.

Everybody is really happy now that they can just type some letters in to find a particular artist/title :wink:

Hi Torben, yes sorry. I was using version 2.0. Really must sort out the shortcuts.

Thanks as always.

Kind regards tony