Have had a quick go with BIAB 2016 abd like the changes very much. Particularly pleased to see the big band horns style which is very good overall (Crooner Big Band 9-part Horns). However although most of the chordings worked well i tried a song where part of the chords went Bm7 followed by E7. I found that the horns were playing a G sharp in the Bm7 chord which you definitely don't want as you need that for the E7 chord following. Strange as all the other chords in the tune were fine. I was trying 'Baby It's Cold Outside' in the key of D. Otherwise sounding really good. I also tried substituting a straight Bm for my Bm7 but the horns still harmonised with G sharps which is wrong.

If PG Music want me to send the chart I can do so.

Great work though on all the improvements

Also try Bm9 and see if that omits the sixth. I haven't heard your problem on a minor seventh, but with a dominant seventh, I often find adding the ninth makes BIAB play correctly. Worth a quick try.

Then if you still hear it, I suggest you make a short recording and send it to PG Music Support support@pgmusic.com along with the number of the RealTrack(s) playing the wrong note, or even just tell them which RealTrack. This has been known to happen before, and they can fix it in a patch.

I will try that Matt. However BIAB putting a maj6th into a Bm7 chord would always be incorrect as it kills any II-V effect. Thanks for the suggestion though

Unless that sixth is voiced as a 13 in a Bm13. That functions fine as a ii in a ii V.

I agree though, you need the minor seventh to sound and I'm not suggesting otherwise.

))
Great work though on all the improvements


Thanks. For the Bm7 | E7 vs E7 Issue,
try regenerating the song, you'll hopefully
Get something different.

For soling and background tracks like this,
Band-in-a-Box uses the chords as a guide,
but not as an absolute.

This usually works well musically because
Musicians solo like that too. So in this case
BB has decided the 2-5 is close to the same
as a 5-5.

Thanks for responding Peter. Yes I did try regenerating the tracks a couple of times and the same thing occurred. I disagree with soloists treating chords 2-5 as close to 5-5 though. In the sequence I tried all the other big band chords were working well though.

Just in response to Matt's idea of trying Bm9 in place of Bm7. Just tried that and the horns still but a G sharp in my Bm7. It's such a shame as all the other chord renditions are good. I would have to fade that bar right down or maybe change the notes using Melodyne.

The big band horns in general are great though and represent a big achievement for BIAB.

))) The big band horns in general are great though and represent a big achievement for BIAB.

Thanks

Funky, when I get my new drive from PG Music, I want to try this to hear for myself what you hear. Just to be sure, what instrument plays that G#, and what RealTrack is it?

I'm really looking forward to messing with the big band horns! They sound awesome!

Can't wait to try Sinatra classics with it.

Just responding again to Matt. the RealTrack I used was CROONR1.STY I think the trumpets were producing the G sharp on my Bm7 chord.

Just to add that I tried another play through today of the song and it has three basic repeats and today the Bm7 - E7 problem didn't happen until the 3rd time through the chords (I always use one big long chord sheet and never use the chorus/tag ending).

I froze the horns for this generation of the songs and then output the wave files. Placed them into Reaper and did a bit of cut and pasting over the problem bars.

There was another funny output where the horns decided to pay a long prominent G over a chord of D which was very odd. Anyway it only happened once and I cut the bar out and inserted another.

I think the big band style horns overall are great. Probably the best thing to do is to generate a few times. Freeze and note best bits then output them each time and then stitch together the best version from all the different wave files.

We horn players just can't be trusted to adhere to the chord tones...

When you make those audio files, keep examples of what is happening, just in case. When I get my shipment, I'm check out what you are saying.

OK Matt. I'll keep the BIAB file so should be able to repeat the error. I could also send you the BIAB file and see if it occurs on your install. As I said I managed a version with it not happening so much today.

Matt, you've just changed one of my standard life examples forever.

In an effort to teach my kids that trust isn't something you can appropriate across the board just because someone has been trustworthy in one way, I've told them:

"There are people that I'd trust with my wallet, but not with a secret...

People I'd trust with my car, but not with my wife... (etc)"

now I'm going to have to add:

"And there are musicians I'd trust to play my guitar... but not to adhere to the chord tones..."


Ba-boom! (tsh!)

;-)

Interesting insight Pat.

I now appreciate those few friends I would trust with my wallet, car AND my wife a whole lot more.

Did this ever get solved?

This thread about the horns adding things is just too close to a real-life situation with my boss, Lawrence Welk! He was very much a talker/thinker like Yogi Berra. So I must add this. One day at the end of the rehearsals for our TV show, he came up to me and said . . . "Chonny!" "Is dat da same trumbet you had for dis morning's rehearsal??" I said . . "Well, Yes it is Mr. Welk"! He then said . . "I dont' tink so"! "Dis one has more notes on it than da udder one!!!" I just agreed and replied that "I'd fix it right away"!! JZ

Does CROONR1 successfully hold a chord? I'm getting a rest instead.

I just tried it and it holds a chord for me, using an ellipsis (Dm11...)
How did you specify the hold?

Also, what Chord(s) did you select that showed the issue?

Yes, there was another post about this in which it was discovered that a chord type had not been recorded. But Bm7? Can’t be.

Bm7 hold works. I'll try a few others and report back...

OK, mixed results but all the ones I tried held a chord (sort of)

I created a song with 24 bars and placed one different chord and hold in beat one of each bar.

Many chords held for 2 beats and then followed a 2 beat rest, others held for 4 beats.

These (of the ones I tried) held for 2 beats only:
Cm
Fm
Em6
Cm11
F9sus
Gm

Of course if I generated again I could get different responses.

None of the chords I tried played a rest for the entire bar.

I've always had questions about some of the things that BIAB does, whether it be Midi or RT. Pushes in odd places, notes that don't work, etc. I made a wise crack about it once and offended someone on the forum, which I didn't mean to do.

It would help me if I understood the concept that BIAB uses to make choices in Midi. Is there some programming by a human or does an algorithm do everything? If a human is involved, that person should be (and probably is) a musician.

It would also be helpful to know if RT's are edited before release so problems like this can be avoided. I'm not asking for any secrets, just a general idea.

Finally, it seems the easiest fix has been mentioned. Let the horns play the entire song, go to Real Band (or another DAW) and cut/paste to get an arrangement. That way you can set up themes and figures that appear more than once and tie the piece together.

2b

Good questions, 2b. I don’t know why these reasonable questions in this form would have offended anyone (hope I wasn’t the one!).

I don’t think it matters about man versus machine because we do know that several of the programmers are also fine musicians, Dr. Gannon included.

I can’t reveal what goes on in beta testing but I think it may be ok to say that we do not review all the new RealTracks before public release. Perhaps that’s something to reconsider. However, it boils down to distribution: there are HUGE files involved, with many releases to test over a short time, and RealTracks are large files to download.

Many of the beta testers do not relax upon the public release, though. We get the full new versions when you do, but we keep testing. We report the occasional problem with a RealTrack to the developers, and PG Music issues a reasonably timely patch.

I hope that gives you your general idea.

For MIDI styles, there are settings as to how "jazzy" the arrangement should be. You can specify whether to jazz-up or jazz-down the chords. I suspect that's where some of the "off" notes come from. That being said, for standard MIDI tracks (not SuperMIDI), the arrangement you get is based on rhythmic phrase that have been normalized to a C7 chord. The other chord extensions are based on the basic C7 chord. The engine then picks on of the available rhythmic phrases based on its weighting, or in some cases whether it is flagged as a special case (every fourth bar, a jump to a fifth, whatever), is transposed accordingly with the additional notes added to complete the chord, and that's what you get. Yes, sometimes you get some bad notes, but again, the engine picks them based on how "jazzy" your setting is.

For RealTracks-based styles, you can set the song use simpler arrangements and now natural arrangements.

All these things go into determining what gets played, and sometimes it gets it right and other times it gets it wrong. But hey, so does a live band, too. Just regenerate until it's too your liking, then lock it down if it is of great concern.

I'm sure there's a lot more that goes on behind the scenes, but that's my rudimentary understanding. The best person to answer about MIDI styles, however, is probably Bob "Notes" Norton, since he has created so many of his own styles for sale that he is intimately knowledgeable about the output produced by BIAB (whether by any inside knowledge or observational intuition).

Thanks, Matt and Jford. This helps.

I don't think it was you that I offended, Matt. Also, I was being a bit too much of a wise-a. My point was right but I presented it poorly.

Thanks again. This forum is great and you guys are very generous with your time and knowledge.

Sounds good.

John correctly mentioned some of the controls a user does have, and there is always Regenerate. In RealBand, you can even regenerate part of a track.

I’m not sure any of us, even Notes, could tell you much more because the algorithms are proprietary information for PG Music. We can only make educated guesses.

That's where SFZ RealTrack Instruments would help, as you can make any hold you like and you are not restricted to just those that have been recorded in Holds_350 \bb\RealTracks\Library\Holds & \bb\RealTracks\Direct Input\Library\Holds