new program for music lovers - testing needed please!

[snyp]

hiya,

Originally Posted by Grimbog

It would be awesome if this could play some of the rarer formats... especially if it can do like Deliplayer and play .cust files somehow, although I can imagine this isn't an easy thing to program.

custom modules do present a problem..
it wouldn't be too difficult to play them.. but any kind of understanding or translation of the data wouldn't be possible, because the structure of the data would be undefined; this means no possible support in the sequencer view, instrument mixer, module info.. etc

a custom module cannot be defined as a format, because it's really just a wrapper for an undefined format.. but the data enclosed within the wrapper is still a format, and i think it's important for historical preservation that the format itself, independent of the wrapper, must be properly defined and recognised.
because of this, i don't aim to provide any support for custom modules themselves, but i do want to support the formats contained within them

hrm.. unless..

do you think i should write additional play-only support for custom modules?
.. this would sadly mean special exception to eMod's overall vision.. eMod's full range tools should be adaptable to all types

Originally Posted by Grimbog

(would be great if rewire was available for synchronisation with cubase etc)

great idea!

i don't intend to support any new-gen formats.. but rewire would be superb.. would give you the flexibility to use eMod with the new-gen apps if you wish

i know very little about midi's inner workings, and i would indeed like your help with this in the future, thanks

added this to the todo

Originally Posted by Grimbog

Encountered 1 cash when using this so far.. but I'm unable to replicate it. It occurred when I loaded up a mod whilst one was already playing.

oh no!! first crash report..

it's very important to catch these elusive bugs now before they cause more serious problems if more layers are added on top of them. rotten roots = dead tree

please grab the debug info from the windows report if you can, whenever you get a crash

Hi Snyp.

Nice effort. All the best for your project.
Keep it good work.It's pleasure to be a part of your forum. I love to try also what you are discussing here. Thanks!

[snyp]

New release 16th November 2010. Check it out

[Cammy]

This program looks really cool, I can't wait for the Aros/AmigaOS version.

[snyp]

Originally Posted by Cammy

This program looks really cool, I can't wait for the Aros/AmigaOS version.

Thanks!

I don't know anything about the API for AmigaOS, hopefuly someone will port it when I release the source

[BuZz]

Just saw some mention of my search stuff earlier. I am happy to give external (sql) access to my search databases if its useful to you.

[snyp]

Originally Posted by BuZz

Just saw some mention of my search stuff earlier. I am happy to give external (sql) access to my search databases if its useful to you.

Excellent, I would really appreciate that BuZz.
Actually the 'browser' will probably be the next thing to work on. WIP right now is some functions of the 'converter'.

I will throw a PM your way when I get an SQL framework set up

regards

[snyp]

Originally Posted by 8bitbubsy

However, the pattern offset calculation is rather slow... Hencing the "Seek" box when you jump to another pattern. What's up with that? Should be instant...

Hopefuly the work I have done since the first alpha has addressed this.
On my Pentium 4 (overclocked to just over 4Ghz) it now takes around 600ms to seek to a 10-minute offset.
Within the 600ms, a complete run of all emulated 68000 code is run to preserve integrity

Originally Posted by Asle

"those three.mod" has 2 extra bytes. "classic.fc" has quite a few extra bytes too

Hopefuly the 'File Medic' can help diagnose issues like this. In some circumstances it can offer an automatic repair of corrupt data.
I am still working on this area of the application

[BuZz]

Originally Posted by snyp

Excellent, I would really appreciate that BuZz.
Actually the 'browser' will probably be the next thing to work on. WIP right now is some functions of the 'converter'.

I will throw a PM your way when I get an SQL framework set up

regards

sure thing. If SQL on the client side would be annoying, we can setup server side stuff with xml output perhaps. i have that in place for modland actually in a basic form, which is used my my xbmc modland plugin - where you send a search string and you get back tunes related to that but the streaming urls, so they can be streamed to the xbox1 or htpc running xbmc. for modland we also store instrument data for some formats, and in hvsc we have a bunch of other data. mostly relational depending on if we need to index from it, otherwise its data packed in json format.

since we already use md5s for reference actually people could already use th streaming interface again their mod collection.. for example

http://modstream.exotica.org.uk/mods...2313b8b98bfcd3

would stream termination day from 4-mat in ogg format or

http://modstream.exotica.org.uk/mods...98bfcd3&fm=mp3

in mp3..

so anyone with a mod collection could actually do an md5sum and try and see if we have it and stream if it matched..

there is actually

http://modstream.exotica.org.uk/mods...8bfcd3&fm=flac

also... currently non public due to bandwidth.. oops. well it was ;-) the service isnt used as much as i thought so what the hell

[trackah123]

Looks good, actually maybe it could be a good converter in the future to convert all kinds of old formats to new tracker formats like MadTracker or Renoise files. Would be great to make Amiga-remixes!

[Loedown]

Originally Posted by trackah123

Looks good, actually maybe it could be a good converter in the future to convert all kinds of old formats to new tracker formats like MadTracker or Renoise files. Would be great to make Amiga-remixes!

I think it would be even better to convert things to MID0 or MID1 format.

Samples are a bit harder but most software nowadays will import nearly anything, cue the Aphex Twin track.

[trackah123]

True, for example Renoise loads XM and MOD.
But i tried a MED file from Addams Family game, and that was a bit harder.

First i had to convert all hybrids to samples in Octamed v4 itself, then used a tool from Aminet called Med2XM and then i could import the XM into Renoise.

[Loedown]

Originally Posted by trackah123

True, for example Renoise loads XM and MOD.
But i tried a MED file from Addams Family game, and that was a bit harder.

First i had to convert all hybrids to samples in Octamed v4 itself, then used a tool from Aminet called Med2XM and then i could import the XM into Renoise.

So what's the point, make something that converts to a current standard and keep the source code open for future developments, MID 0/1 is an old standard but it's still current.

[snyp]

Originally Posted by BuZz

If SQL on the client side would be annoying, we can setup server side stuff with xml output perhaps. i have that in place for modland actually in a basic form, which is used my my xbmc modland plugin - where you send a search string and you get back tunes related to that but the streaming urls, so they can be streamed to the xbox1 or htpc running xbmc. for modland we also store instrument data for some formats, and in hvsc we have a bunch of other data. mostly relational depending on if we need to index from it, otherwise its data packed in json format.

Well, if people will be using eMod to browse and fetch data from your domain, bandwidth becomes an issue, especially if we are thinking about mp3 (or flac!). I would like to see access to the module files themselves, with SQL queries to search and sort the title/artist/game/demo etc.

You have given me other ideas too, maybe I should include a seperate workspace to connect to radio streams

Originally Posted by Loedown

So what's the point, make something that converts to a current standard and keep the source code open for future developments, MID 0/1 is an old standard but it's still current.

The thesis behind conversion between two formats is to convert the source into a 'language', and then convert that to the destination format. Because of the different types of effects provided by different formats, the conversion language needs to be as generic as I can make it, and unfortunately in some cases there will be accuracy issues in the destination output; synth sounds converted to pitched samples is one example, arpeggio effects is another example.

Point of interest, when I begin work on the editor/sequencer workspace, I intend to develop a 'super-format', which will closely emulate the synth sounds of several formats in one piece of music. Example, playing Delta Music sounds with Future Composer sounds with Sonic Arranger sounds.

[BuZz]

Originally Posted by snyp

Well, if people will be using eMod to browse and fetch data from your domain, bandwidth becomes an issue, especially if we are thinking about mp3 (or flac!). I would like to see access to the module files themselves, with SQL queries to search and sort the title/artist/game/demo etc.

the bandwidth for the data wouldnt be much for the database. the streams would use a fair amount if it was busy, but usage is quite low at the moment. the streaming backend was not really relevant for you I guess but I wanted to mention it . for the database stuff for modland for example you can pull out module name, author, instruments, collections (which could be a playlist, or the name of a game the module is from) and a few other things. this is just generated from the modland filesystem structure. for hvsc, we use the filesystem + the meta data in the sids, + the stil + remix.kwed.org references to link to remixes. I am happy to share the scripts for hvsc/modland etc if you are interested (made in perl).

[snyp]

Originally Posted by BuZz

for hvsc, we use the filesystem + the meta data in the sids, + the stil + remix.kwed.org references to link to remixes. I am happy to share the scripts for hvsc/modland etc if you are interested (made in perl).

I would definitely be interested in the HVSC scripts Buzz, thanks
However I'm not planning on the work for SID (and C64 hardware emulation) support for many months yet.

Because HVSC is as important as it is, I will most likely begin work on the 'browser' workspace, the SID format, and c64 emulation at the same time. Can i ask you to retrieve them from your backups in a few months?

Originally Posted by BuZz

for the database stuff for modland for example you can pull out module name, author, instruments, collections (which could be a playlist, or the name of a game the module is from) and a few other things. this is just generated from the modland filesystem structure.

Generating lists and metadata from the filesystem is pretty much what I'd planned with most online collections..

This reminds me of something you can clarify for me BuZz, a few months ago I downloaded a chunk of the filesystem on Unexotica to get an idea of how to organise your collection, and I spotted .txt files such as "composer.txt" and "archivename.txt".. How are they generated? Do I need to parse them, or will our SQL interface bypass anything useful in these files?

[BuZz]

Originally Posted by snyp

Can i ask you to retrieve them from your backups in a few months?

sure. are you like

I'll sort you out with a copy. you might need to remind me - you can contact me on buzz [at] exotica.org.uk - I need to just remove one thing from the script which is the private reference/access to the remix.kwed.org site to pull out their data. you can contact them and ask for access if needed.

Originally Posted by snyp

Generating lists and metadata from the filesystem is pretty much what I'd planned with most online collections..

This reminds me of something you can clarify for me BuZz, a few months ago I downloaded a chunk of the filesystem on Unexotica to get an idea of how to organise your collection, and I spotted .txt files such as "composer.txt" and "archivename.txt".. How are they generated? Do I need to parse them, or will our SQL interface bypass anything useful in these files?

they are manually generated by contributors, but then the data is gathered up by a perl "bot" which then edits the various related wiki pages.

http://www.exotica.org.uk/wiki/UnExoticA/Structure

however I may well make a sql database for this stuff too to have a similar search frontend as for modland.

[snyp]

Originally Posted by BuZz

private reference/access to the remix.kwed.org site to pull out their data. you can contact them and ask for access if needed.

I can't decide.. Something to include as something more than a web link..

Once I heard a remix of Saboteur II and I loved it so much I emailed the remix author to discuss it, maybe he is here on the forum, Marcel Donne?

Remixes shed a light and cast a shadow..


Then again, special interest is special interest

[8bitbubsy]

What .MOD replayer do you use? It even plays black_queen.mod correctly - impressive!
I guess it's not one you coded yourself since you use the UAE emulator in your program?

[snyp]

Originally Posted by 8bitbubsy

What .MOD replayer do you use? It even plays black_queen.mod correctly - impressive!
I guess it's not one you coded yourself since you use the UAE emulator in your program?

I use protracker's source replayer with some modifications. UAE is used to translate/execute the motorola opcodes, then the Paula/DMA state is captured.

Re: black_queen.mod,
I looked this up on youtube and I can see what you mean
As you can see, playing this module in eMod doesn't come with the visual 'tricks' which comes with playing it in Protracker.. The pattern data you see is produced by a function in the plugin DLL. Each format has it's own version of this 'translator' function which basically runs through the module and produces a table of the data for each pattern step.
The red lines in the pattern view denote a pattern break, and yellow lines denote a loop.

Following on from your post I tested 'ode 2 protracker' as well, apparently notorious for it's trickery..