JammerLive: Changing Styles

If I try to change song styles while a song style is playing, the one that is playing shuts off. I was wondering if I could avoid this by locking the playing style, but no.

On the other hand, you can have up to about 28 band styles per song style. These can be mains, intros, endings, fills, etc, without limitation. You can mix band styles across song styles before starting autoaccompaniment. Also, you can make new song styles with any combination of band styles. So there is great flexibility as long as you know what band styles (variations) you want before you start autoaccompaniment.

However, this is certainly a let down from my Yamaha PSR-8000 which allows me to switch styles on the fly.

Clif

Clif,

I understand your concern, but frankly, I am not surprised. You are comparing recent top of the line "professional" keyboard (which probably cost around $2500) to a $60 program, which is mostly intended for home/practice use. Yamaha has, I am sure, a large team of engineers and almost 20 years of experience designing and making this types of auto-accompaniment systems, whereas JammerLive is a fairly new program, likely written one fresh-out-of -college kid.

I am convinced that JammerLive has more flexibility and probably quicker response to users' comments - so by all means do express your wishes to Soundtrek, but software companies live by issuing continuous updates, whreas the hardware makers ship you a fairly complete and full-featured instrument, but then not supporting it with hardware updates beyond some obvious bug fixes.

Best Regards,
Alex

Hi Alex

I am not really "concerned". I am quite happy with JammerLive. I think it compares quite favorably with dedicated hardware. I am not a gigging musician, but I feel that a combination such a Yamaha S80 (with added harmonizer), and a Sony CS1x minilaptop would work well live. Maybe I am wrong, but I assume that it is not usually necessary to change styles mid-song during a gig. Changing styles without restarting is more useful for experimentation. I consider JammerLive already a competitor to high-end arrangers, and its competitiveness will only get stronger with time.

Clif

Hi Uncle Dave

I trust you, but I would like some insight into your opinion. I can accept you personal preference for having styles built into the keyboard, but do not understand your conclusion that others might not find JammerLive suitable for live performances (assuming that the lack of available styles is overcome). Is it because you do not know ahead of time what you are going to play? If you play only from a preset list, it seems you could set up a style for each song and not have to change while playing, right?

Clif

Uncle Dave Pt 2

I think it is a mistake to try to combine new programs, e.g., GM500 and JammerLive with the expectation they will work together. I expect down the line, they will work well together. Right now I am slowly learning JammerLive by itself. The next step will be to use it like a QY-70 either to expand the style selection of my PSR-8000 or to provide arranger capabilities to another keyboard (such as a Yamaha S80). In either case, the sounds will be provided by the keyboard. There will be no need for a workstation, only a 2-3 lb laptop will be required, e.g., Sony CS-1x or whatever (the one with the built-in video camera).

If I did not have a good arranger keyboard, I would get one. But given that I have one, I do not want to feel I must upgrade everytime a new model is introduced. If JammerLive can be used as a way of extending the style prowess of the PSR-8000, it will serve me well.

Clif

Hi Frank

The two midi ports you need are "In" and "Out". I am not sure why you label them "A" and "B". JammerLive would translate chords to accompaniment patterns. Thus function requires that the chords be sent to the computer and the accompaniment be sent to the keyboard (with built-in sounds). The non-accompaniment sounds do not require midi as they rely on local control by the keyboard.

One attraction here is the freedom to select any synth as a controller. Now you can have a Korg Triton arranger, a Kurzweil 2600 arranger, or a Roland XK88 arranger. I might pick the S80 because I could add vocal harmony to it.

However, I agree that not being able to select styles from the controller is a serious limitation. Soundtrek should treat sounds as patches and allow them to be picked with bank and path numbers (to allow up to 16k styles).

Clif

Hi Frank

Thanks for helping me think this through. Yes, if you use an external module for both accompaniment and lead, then a merge is required somewhere (with the exception you noted). I would prefer not merging at the computer. Note that merging is only necessary when one module is used for both accompaniment and for lead. If the funtions are split between modules (or between a module and the keyboard) no merging is required. Where merging is required, this can be performed in some cases by the keyboard itself (e.g., my Roland A-50) and in some cases by the module (I will try to get some models that permit this).

Given your warning, I would limit the computer to only one function, and that is converting chordal inputs to autoaccompaniment. I guess it is best to let this new field mature some before trying to combine programs and functions, especially from different manufacturers.

Clif

There seems to be some confusion on using multiple styles in a song. You can set up keystroke and MIDI triggers for any band
style / performance event that is loaded, and you can load multiple song styles per song. So, when jamming, you can call up any of the loaded band styles by issuing the proper trigger. JAMMER Live is set up to auto configure triggers for 2 loaded song styles (each with an intro, grooves, ending etc...). Depending upon the instrumentation in the styles you could conceivably load and use more. The limiting factor is the number of MIDI channels used. At any rate, loading a new song style in JAMMER Live is quick.

Nevertheless and unlike ROM based samples, the GM500 sounds do require pre-loading in gigasampler. This isn't the case with a hardware sound modules.

Of course you can always (pre)load once the complete GM500 library. Joe Hardy (on the gigasampler forum) did just that:

"I made a Performance which includes all 128 instruments plus the drum kit. On a 450 MHz machine it takes 4.0 minutes to load the Performance. (It also requires 196 Megs of RAM as a minimum.) The advantage of loading everything is that GS will respond to all 128 program change messages. Among other benefits, this will speed up play between songs. GigaStudio is expected to drastically reduce loading times which will make pre-loading even more attractive. Also, the new Effects in Studio should make it easy to produce killer arrangements."

BTW, the "Conexant Player" is "built in" to Gigasampler version 1.6 - it is not a stand alone program as many of us had thought

From the Gigasampler Forum:

Regarding the Sonic Implants 24M GM Soundfont set, like Frank, I am very unimpressed. As I said in another post, the only two instruments that I like are the t-bone and the soprano sax. By comparison, I much prefer the 8 Meg Emu set supplied with the Soundblaster Live.
It is a shame that Jammer Live does not seem to want to work with GS. Even using the Hubi drivers, MIDI communication just will not happen. I have no trouble at all with Cakewalk - AND - all program changes work faultlessly (with the full GM set pre-loaded). When a new sequence is loaded (from within Cakewalk, not by dragging it onto GS), it is ready to play instantly. If Cakewalk can work, it seems logical that Jammer could be made to work as well if Soundtrek were motivated to do so.

While Frank is understandably disappointed that the GS/Conexant/JLive does not seem at this point to be a viable replacement for a Yamaha PSR-9000, I cannot agree that the Conexant GM kit suffers by comparison with the 4 meg Roland set or any soundcard in my acquaintance. I have been putzing around with GM sets for years and I can tell you I am DELIGHTED with the Conexant set!

Joe Hardy

Hi Folks,

I guess it's time that I surface and say a few words about JAMMER Live in this forum. I stay pretty busy developing software and styles so I don't get over here too often.

My name is David Castles. I am one of the lead developers of JAMMER Live. I guess a little background is in order since most of you don't know me.

I started developing MIDI software about 10 years ago with the birth of JAMMER Professional and have been developing music composition and MIDI software full time for the last 10 years. In the 7 years before that I designed software for digital recording telecommunications hardware and ROM based microprocessor operating systems for intelligent computer terminals. Before that... I was a professional musician working in Atlanta GA and on the road for about 5 years with various bands. I am a writer, a vocalist, and a recording artist, and I play guitar, keyboards, and bass. The other members of the development team are also seasoned professionals with both musical and computer expertise.

Having said that, let's talk about JAMMER Live and what it can do for you. JAMMER Live was developed as a professional tool for both writers and performing artists. Don't let the price tag fool you. We just wanted everyone to be able to afford this software. JAMMER Live was developed with an open architecture in order to give you a lot of flexibility in setting up whatever types of jams you can imagine. The number of styles you can preload or patches you can use in a jam or in a style is limited only by your hardware.


----- About Changing Styles on the Fly -----

With hardware accompaniment, the manufacturer knows what patches are being used in each style and can access those patches directly. In the world of software, a program must send MIDI messages to the various MIDI devices to access patches which can result in several problems if you try to do it on the fly.

Problem 1 - MIDI Choke
If you are in the middle of a jam and try to load a style that accesses 8 new patches, the software must initialize each of these new MIDI channels to the settings stored in the style for each musician. This could result in sending out 8 patch messages followed by 8 level messages, 8 pan messages. 8 reverb messages, 8 chorus messages, and even some expression controllers. And after that, would come all of the notes that play on beat 1 in the style. Because MIDI is sent serially and each message takes a given amount of time to be transmitted, sending this many messages out all at once can cause a noticeable delay or gulp in the music known as MIDI choke.

JAMMER Live avoids this problem by knowing what channels and patches are to be used before the jam begins and initializing them before starting the jam. During the jam, some controllers are still sent out to control levels, reverb, chorus, and pan as grooves change and fills occur, but JAMMER Live keeps track of the state of each controller on each channel and only sends out controllers when that state changes minimizing the number of actual controller messages required to be sent when a groove changes or a fill occurs.

Problem 2 - Insufficient Channels
With JAMMER Live, you can access up to 512 MIDI channels in a jam session (32 ports x 16 channels per port). You are only limited by the number of MIDI ports and MIDI devices that you have hooked up and configured for use. If you have a single port card and a single MIDI output device, then you are limited to using 16 channels during a jam (not a big limitation unless you're doing symphonies). However, if you tried to load styles on the fly and exceeded your system's available channels, you would be unable to load the style properly during the middle of the jam, something you do not want to experience in a live performance. Therefore JAMMER Live calculates and initializes all of the MIDI channels needed before the jam starts to insure that there will be no unwanted surprizes during the jam.


----- Loading Styles Before the Jam -----

Before you start a jam, you can load as many styles as you like. After loading a style, you can lock down the performance events so that they will not be unloaded when loading the next style. You can even lock down selected grooves from various styles and build a custom jamming landscape. After loading your desired grooves, fills, intros endings etc, you can customize the triggers for each of these in the Performance Event Editor. This open architecture allows you to create highly customized and robust jam sessions. This setup can then be stored as a song or a new style. The song format is recommended for a performance you want to repeat in the future since it saves more information than a style file. Note: A style file will only save the grooves, intros, fills etc that you have performance events define for.


----- Sending Patches and Controllers -----

JAMMER Live has been updated to process incoming patch messages and controller on the user's channel. The updated version will be available for download in a few weeks. Check the SoundTrek web site (www.soundtrek.com) in a few weeks for availability.


----- About Playing Different Riffs Over Different Chord Types -----

Instead of using different riffs for different chord types, JAMMER Live adapts riffs to fit the chord they are playing over by adjusting notes on the fly to accomodate the current chord. This is a large part of what I do as a musician when playing a musical theme or phrase over different chords. Of course with "wildcard notes", JAMMER decides what scales would be good to use over the given chord and composes notes on the fly using the selected scale.

We discussed using different riffs for different chord types in the early development of JAMMER Live but decided on using the "fit to chord" approach for several reasons.

One problem you can run into with the other approach occurs when you play two or more chord changes during a bar or during a single groove riff. Aligning the new chord specific riff mid bar to where the chord changed can lead to awkward timing splices and awkward groove alignment especially if the chord changed on an offbeat eighth note boundary. JAMMER Live avoids this problem by not starting a new riff mid bar but continuing on with the current riff and simply adapting the riff notes to the new chord type. While aligning riffs to chord boundaries within a measure can produce some interesting results, it seems more like something a computer would do rather than what a human would play.

Another issue we dealt with was leadins. JAMMER Live riffs support leadins or parts that play ahead of where the main riff starts leading you into the riff. Here you run into a problem if you are trying to play chord specific riffs and encounter a chord change at the bar line after you've already started the leadin based on the chord in the previous measure. Once again JAMMER Live avoids this by adapting the riff notes to the new chord type and playing the leadin into the riff seamlessly over the chord change.

One last issue with chord specific riffs is that using it requires you to create a lot more riffs. If we created a piano groove in JAMMER Live that contained 5 riffs to give it variety, doing the same for major, minor, m7, maj7, and dominant 7, would mean creating 25 total riffs instead of 5.


----- Please Stay in Touch -----

I appreciate the open discussions you are having in this forum and will be glad to help out in any way I can. I don't always have the time to check out all the boards around so, if you have a specific technical question for SoundTrek, please email us at support@soundtrek.com and we will be glad to help out the best we can. We are always interested in knowing how you use JAMMER Live and what other features you would like to see us implement in the future.

Thank you for your business and support!

Keep jammin,

Dave Castles
SoundTrek

Dear David Castles

I want to thank you wholeheartedly for your contributions to music lovers and to this forum. I have been waiting many years for a program like JAMMER Live. Often I would read advertisements for “autoaccompaniment” programs, hoping they would be what I wanted. However, the critical real-time feature was missing. Judging by the responses in this forusm JAMMER Live meets a long-felt need. I believe the only two-page thread in this forum addresses JAMMER Live.

With any important new software introduction, there is a learning curve, both on the parts of the users and the developers. I think those with a mature respect for complex software will take time to learn the capabilities of JAMMER Live before trying to integrated it with other types of software. I think a little patience would have avoided some of the “sour grapes” attitudes.

I think the complaints about having to preload are largely unfounded. Of course it would be better not to have to preload. However, hardware arrangers have styles preloaded into ROM. Inherent in any RAM based system is the need to transfer stuff to RAM before using it. I do not see the need to preload as limiting creativity. All musical instruments have limitations. Acoustic instruments have limited sound palettes and no arranger functions. The styles in hardware arrangers are limited in complexity. JAMMER Live offers the most flexible format available for real-time accompaniment styles.

Those concerned with creativity forget that just because you have JAMMER Live does not mean that you have to use it all the time. Styles generally are canned to some extent. What JAMMER Live does do is to allow a musician to choose any keyboard and turn it into an arranger. You do not have to pick a 61-key PSR-9000 or Solton X1, and give up two of five octaves for chording. You can use the full capabilities of your favorite 76-key Triton or 88-key Kurzweil, or two-register Roland Hammond imitation, and add (or not add) auto-accompaniment at will.

We need to thank you for making JAMMER Live available at an irresistibly low price. Of course, a computer is needed to run it, but its cost can be amortized over other applications. When it is time to upgrade, the hardware arranger will have to be replaced, while a software update will suffice for JAMMER Live. Over the short and long run the economics are compelling in favor of JAMMER Live.

The most promising aspect of JAMMER Live is your willingness to contribute to user discussions of your product. I found the explanation of the decisions that were made regarding JAMMER Live very illuminating. I think with this sort of engagement, the future of JAMMER Live is very bright.

Here’s to the further development of JAMMER Live, including “mega-versions” with expanded style sets.

Clif

I just got home after playing in front of 150 people. Although every show is different, I must concur with Uncle Dave - in a live performance spontaneiety (is this how you spell it?) is the key. Being able to switch between styles and sounds is a key ability for a musician in my situation. No matter how hard you work on your repertoire and try to arrange the music in sequence, each gig will inevitably demand them to be played in different order. And with the crowd of dancing people, I do not have the luxury to stop the music to preset the next song - I have to do it without stopping the beat. Of course, I use G1000, and it's usability leaves much to be desired (I am sure many of you remember my not so complementary posts before), but I am still able to switch the accompaniment style, or the solo voices with a few (very careful) button clicks.

If I were playing at home, this would not have mattered, but on stage it does.

A note to David Castles - I welcome your participation in this forum, and I think this says much about your company's commitment to your customers. I am sure that you will be able to learn of many requirements which the arranger keyboard players consider essential in their instruments. For example, my Roland keyboard will not switch the styles in mid-bar (only at the beginning of the next bar), but it will allow me to change chords to my heart's content within a bar. For many songs, this is essential. The other capability it allows is differentiation between the accompaniments playing in major, minor, and 7th chords. Granted, I have the ability to make all three sound the same (while still responding to the chord changes in the recognition region), but I also have the ability to make all three sound different. Is it useful? - for some songs yes, very. I guess this is not essential, as I understand that Yamaha does not do this on their keyboard (someone with a Yamaha keyboard, please correct me if I am wrong).

I do not have the Jammer Live at present, so I can not comment on the specific features of the software.

I am glad to know that you and your colleagues at SoundTrek are also experienced musicians. Having played with an orchestra in my younger days, I find that the skill set, the environment, the type of pressure in that situation is totally different from being on stage as a solo or a duo live gig with an arranger keyboard. Although I will admit that my instrument technique was much better in those younger days, then it did not require the ability to follow the crowd and to adjust the program to the changing demands and moods of the party.

A full-featured easy-to-use instrument allows me the ability to do that, and the features about which the session musician does not care about, are detrimental for a guy like me.

Regards,
Alex

I just got in from "the gig" and was browsing through the forum. Uncle Dave's comments got me to thinking about tonight's job.
On a four-hour job, I probably played 50 songs, and at least 40 were requests. Most of the requests were songs I had no intention of doing ahead of time. Tomorrow night will be more of the same, but the requests will be for different songs. I don't see how the "Jammer" system will ever be suitable for professional live use if it has to be prepared ahead of time.
And I totally agree with the problems that the Karaoke Mentality is presenting. I wish all the Karaoke units could be rounded up and sent to Japan. People are running around doing Karaoke for $40 a night, and I don't want anybody to confuse me with them.
Incidentally, wouldn't it be great if somebody in charge at Yamaha, Roland, Solton, Technics, Korg, GEM, etc. would take the time to visit our discussion like David Castles did? I never get the impression that they listen to musicians or care about what we want or need.
While I'm rambling, I had a job in Texas last weekend. I told them I couldn't come back until gasoline prices went down, or I'd have to raise the price by $60. That's what it takes to fill the Suburban with gasoline at today's prices.
I have vented sufficiently to get a good night's sleep now.
Don

Frank,

I got this back from Soundtrek. Maybe it is of help to you.
-----------------

>It is a shame that Jammer Live does not seem to want to work with GS.

JAMMER Live works with GS, I'm using it with a Roland Sound Canvas which is
the definition (and originator) of GS. I've yet to ascertain exactly what
isn't working. This is why I haven't contacted Nemesys yet - I want to have
more information so that the discussion doesn't go like this:

Me: "JAMMER Live doesn't seem to be working with GigaSampler".

Nemesys: "How so?".

Me: "I don't know".

I know that the current version of Live does not pass through patch changes played live on a MIDI controller, but the upcoming update will address that issue. I'm wondering is this the issue Frank is talking about, or is he not able to get any patch changes to work at all? If that's the case he needs to make sure the bank select method is set properly (based on what GigaSampler wants) and of course to get the actual bank and patch numbers correct (depends on how the samples are loaded up).

With some sound modules, if the wrong bank select method is used or the wrong bank number(s) / patch number(s) are used, the result is silence. An easy way to check this out: First, clear the GM soundfont bank (bank 0) using the Creative Labs soundfont utility in AudioHQ. The reason I have you
clear bank 0 is that the Sound Blaster live "falls back" to bank 0 - i.e.
when an unused bank / patch is sent or when an improper bank select method is used the SB Live card defaults to playing the oundfont that is in bank 0. Not all sound modules do this. Go to Cakewalk and start a new song. Go to a track and enter a few notes in the piano roll. Now, load a sound font into bank 1. Next, set the track's MIDI channel to channel 1, assign it to bank 1 and choose a patch (the soundfont patch name for the chosen bank / patch should show in the track view). Make sure that the track's bank select method is set to controller 0 only (double click on the patch or bank field to set this). Press play and you should hear the MIDI notes.

Now, rewind and change the track to bank 2 and play again. What do you hear? Silence! Change back to bank 1. Sound again. Now change the track's bank select method to normal instead of controller 0 only and what do you get? Silence. This should illustrate what I'm talking about - given the fact that this experiment can be done with Cakewalk and a Sound Blaster Live card shows that it's likely a setup issue and not directly a problem with either JAMMER Live or GigaSampler.

Seeing that it works fine with Cakewalk, and if I understand GigaSampler correctly, it just looks like a standard Windows MIDI device to Windows applications, it should work fine in either app. Maybe Cakewalk has added GigaSampler specific support but I really doubt that Nemesys would require
applications to code explicitly for GigaSampler instead of using the standard Windows multimedia services, kinda defeats the purpose. Those should be the only two pieces to the puzzle - get the bank select method right and send the appropriate bank and patch numbers (based on the device
and style patch configuration) and all should work fine. If not, then at
least I know the configuration is sound and I then have enough information to approach Nemesys to figure out what's going on. Until then we need to make sure that his system is set up properly.

I really wish Frank would contact us about this. I've received a few emails from him a while back but none addressing this issue. If not for you, I wouldn't even know of the problem.

>Even using the Hubi drivers, MIDI communication just will not happen.

Are you sure that MIDI communication is not happening? Or is it a bank select method / bank / patch problem? It could certainly be a JAMMER Live problem, but without more information and communication there's no way to know and absolutely nothing we can do about it.

>I have no trouble at all with Cakewalk - AND - all program changes work faultlessly
>(with the full GM set pre-loaded). When a new sequence is loaded (from within Cakewalk, not by dragging it onto GS), it is ready to play instantly.
>If Cakewalk can work, it seems logical that Jammer could be made to work as well if Soundtrek were motivated to do so.

Since Cakewalk works, he might want to compare the bank select method used in Cakewalk to that used in JAMMER Live.

>While Frank is understandably disappointed that the GS/Conexant/JLive does not seem at this point to be a viable replacement for a Yamaha PSR-9000,

I'm not convinced of that yet! If there are features of the Yamaha that Live doesn't have, great let us know - they will be great ideas for the future. But it seems like the the complaints I read were about such things
as load times from GigaSampler but I hardly see how JAMMER Live can do anything about that. Maybe GigaSampler isn't the answer, but that's _not_ a JAMMER Live issue. This loading problem simply does not exist with a
hardware sampler / sound module. We should be able to get JAMMER Live
sending MIDI messages to GigaSampler though.

>I cannot agree that the Conexant GM kit suffers by comparison with the 4 meg
>Roland set or any soundcard in my acquaintance. I have been putzing around
>with GM sets for years and I can tell you I am DELIGHTED with the Conexant set!

Well, you know sounds are completely subjective

Richard Grant
SoundTrek

Hi Folks,

Maybe I can clear this style loading thing up.

The load time for a style in JAMMER Live is about 2 seconds.

This is a pretty insignificant delay unless you're trying to do a "dance mix" thing where you run one song into another and the music never stops. If this is your desire, then you can create one mega-dance style that has 20 grooves in it and jump between them any way you like with no delay and with all the spontaneity of a bottle rocket with the stick chopped off.

The only other situation I see that would require advanced preparation in JAMMER Live is if you wanted to play a song that used Groove A from Style 5 and Groove B from Style 9. This sounds like it would be more than you'd want to keep up with during a live performance even on a hardware arranger, and the button pushing would require a degree from the Commander Data school of live performance.

If we're not talking about one of the above two cases, then we probably have the typical gig where you end a song, the crowd starts to applaud, you say thank you as you select another style from the style window and press start. Two seconds later you're jammin again, long before the applause ever came close to dying down or anyone even thought about leaving the dance floor. Of course, every now and then, you'll want to grab a sip of your drink and toss out a one liner. And JAMMER, like any good backup band will join the toast, laugh at your jokes, and wait patiently for you to call the next song.

keep jammin,

dave castles
soundtrek

ps: JAMMER Live can change chords and/or styles on any quarter note, eighth note, triplet eighth note, sixteenth note, triplet sixteenth note... or any clock value for that matter, and the quantize for each can be set to whatever value you like as well

ppss: BTW, have any of you experimented with the interactive modes in JAMMER Live for doing custom fills at the end of a section or for doing experimental composition? It's not something that you'd use much in a ballad or a real straight ahead pop groove, but theres a lot of open ground in stuff like jazz improv, fusion, dance and electronica.

just wondering... dc