Any serious musician knows the the bass part in a song is very important for the global harmony of a song and for the groove.
So, getting a reliable and precise bass part is very important.


***THE PRESENT ISSUE :
When we specify a bass note over a chord, the realization is mandatory. Up to now, the requested bass note may be played or another note may be played.
Many Bass RealTracks have a defective implementation of slash chords. It has already been reported in the forums to PGM people.
So, to attempt to get the right notes, we regenerate the whole song, because we can't - yet - regenerate only the bass RT.

But, we may get the right bass note for some chords in the song and, in the same time, there where it was ok with the bass, we now get wrong bass notes.
So, it is time consuming and the result is poor.


For example : Bass RT N°2426
C2 chord / 2 beats : the bass plays G - E, sometimes you may get the notes C - E, that is correct.
Cmaj7 chord / 2 beats : the bass plays G - E, sometimes you may get the notes C - E, that is correct
G7 / 1 beat : the bass plays F, sometimes you may get a D note or a G note, the correct one.
G7/F / 1 beat : the bass plays E or B.
See view below.

***PROPOSAL :
For each Bass RealTrack recording, it would be fine to record sufficient combinations of slash chords.
I reasonably propose to do that for each supported chords having 3 or 4 notes.
For example :
C chord : check to get the right bass notes for C/E and C/G for durations of 4/3/2/1 beats
C7 chord : check to get the right bass notes for C7/E, C7/G, C7/Bb for durations of 4/3/2/1 beats.

***ADVANTAGES :
- we could get more elaborate bass lines : useful for diatonic/chromatic bass notes
- we could get the exact bass note as requested in the arrangement
- we could get precise bass lines : we don't get wrong notes
- The final result is a better sound, a more musical and a more professional arrangement.

An issue I have with slash chords is the random octave choice, this is particularly annoying when trying to create a sequential bass run but a note in the middle of the run plays an octave up or down from the rest.

Vintage

Well articulated and very worthwhile.
I hope that this gets the attention it deserves.

John-Luke:

Is there any chance that this could be easier than we thought?
I can see where you would need to do a ton of extra recording for "slash CHORDS", but for the bass, surely BIAB could be made to just:

- identify that there is a "slash chord" in the upcoming chord cell
- ignore the first part of the chord and play ONLY what's after the slash

Even if it has to play a run or arpeggio or whatever else,
couldn't it use the note/notes for the designation AFTER the slash?

If the chord cell indicates C7/E, why couldn't BIAB just play the bass note/notes for E, and ignore the C7?

Just askin'.
LLOYD S

I'm not sure to fully understand your proposal. Could you detail or provide an example.

In the other hand, maybe the 'ton' of recording is not necessary.
We, users, don't know anything about the RealTracks and how they work under the hood.
For example, the chord C7/E (2 beats long)and the chord am7/E (2 beats long) would play the same portion of audio wave.

You're right, that is another issue with the Bass RealTracks.
We have the same problem with the MIDI bass, but there we can edit the bass part and freeze the track, and it's done.

For this issue, as BIAB knows the sequence before he plays, we could imagine a feature checking for the right choice of octave. It don't seem very difficult to implement, just to calculate the gap between two optional notes. When the value of the gap is minimal, you get the right note.

So, we would get a more professionnal sound.

Agreed about the register jumping, and it should be possible to program such a check. It may be there already (I think it is) but what does it do when no such bass notes on that chord were recorded in the desired register?

Good point.
I guess this would be more easily attainable in a MIDI track, and highlights the never ending challenges that the designers face when dealing with creating Real Tracks.

John-Luke:

You said: "I'm not sure to fully understand your proposal. Could you detail or provide an example."

When PGMusic records a musician for a RealTrack, they must typically record CHORDS or at least multiple notes at one time. So for example to record a guitar RealTrack, they would need to record him/her playing a C chord, C7 chord, Cmin chord and so on for a goodly number of different chords. And then do the process all over again for every key (unless they use their "transposition magic" for other keys - I don't know).

In any case, for them to include slash chords, they would have to record considerably more, to get (as you suggested) C/E, C/G etc. chords.

All of that pertains to guitar, and any other instruments where you typically play more that one note at a time, like piano.

However for playing the BASS, like electric bass, double bass etc., you typically only play ONE NOTE at a time. So when BIAB shows C/E in the chord sheet, why couldn't it be programmed to just play an E note there? It's just the bass instrument that needs to ALWAYS play what comes after the slash.

I play the electric bass.
When I'm following along on a chord sheet (or BIAB chord grid), and I come to a C/E chord, I typically play an E note, as that's what the composer expects to hear from the bass.

Perhaps I've misunderstood what you're trying to get BIAB to do.
Good luck!
LLOYD S

I’m not disputing that the wrong bass note was played in the very clear example given. But I’ve never experienced this and I use slash roots all the time, so the question becomes, is it an error in the programming for just the RealTrack(s) you are using? If so, PG Music can fix that and issue s patch.

Related, for anyone desiring the bass to remain on the same note, you should also explore pedal bass.

Matt,
You're right : all the bass RT don't show this issue.
I have not checked all the bass RT, but following RT have defective slash chords realization : 711/2475/2426/2344/1901 and probably others.

The purpose of the post is to ask PGM people to add a check phase of slash chords when PGM releases a new bass RT in order to get a full functional RT.
I agree all defective bass RT have to be fixed.

So, that would avoid PGM to do an extra job after the release to fix the issues.

+1000

Also,RT2957, 2608, 2590, 2702, and probably many others (including Ron Carter's) don't work with slash chords, and because of that they're nearly useless on all projects using Slash chords.

This is basic functionality, not an obscure feature nobody uses. These bass realtracks definitely need to be fixed.

I run into this problem all the time with real tracks and just live with it. I use many of the medium to up-swing jazz styles, with walking bass. Say, you’re in the key of F. Often at the end of a phrase, when the basic chord is F, the bass part will play a note in the scale, but not the F. It’s noticeable to me, but, maybe not to others.

Let's stay on this one. I suggest if you have this problem, you send a frozen song that demonstrates this to support@pgmusic.com and reference this thread.

Here's an interesting twist.

I'm using Bass RealTrack #2284 (Acoustic Bass, Bossa Ron EV 140). It plays this fine:

|Dm9| |Dm7/C|

But the bass stays on D when playing this:

|Dm9| |Dm9/C|

EDIT: I take that back. It played the first occurrence of these two chords Dm9 to Dm7/C correctly, with the bass moving to C, but all subsequent cases the bass stayed on D.

This is a test file you can download and try for yourself:

https://drive.google.com/file/d/1nBwoqQT_GMPvUmetPinTuhs0XksAveNo/view?usp=sharing

The chord progression is:
| Cm7 | Cm7/D | Cm7/Eb | Cm7/G |

There are five bass tracks, so you can try all of them, by soloing each one in the mixer, one by one.

If you try, you'll see that all of them stays in Cm for the whole 4 bars.

RealTracks in song: 2284:Bass, Acoustic, BossaRon Ev 140
RealTracks in song: 2957:Bass, Acoustic, ModernJazz8thsNeil Ev 120
RealTracks in song: 2590:Bass, Electric, RockPickedQuarters Ev 165
RealTracks in song: 2702:Bass, Synth, CloudRapSub Ev16 075
RealTracks in song: 2608:Bass, Electric, ModernDiscoFunkAlex Ev 110


PS: I'm actually testing other Bass Realtracks over that chord progression. Very few play the correct notes, most of them simply keep playing (incorrectly) over Cm for the whole 4 bars.

Confirmed, the basses on this song play only on the Cm.

I guess you need so much recorded material to cover all the possibilities.
That's where Melodyne and ARA comes in handy, just edit the midi notes in one track and the audio changes in the other.
That's how BB will need to work eventually, move the RealChart notes and the RealNotes will move.

You can also drag the midi RC into RapidComposer with the same chord track imported from BB it will then refit the midi to the notes you need, but you will need to output it through a Virtual Bass.

I disagree, to get a basic functionality, the bass should simply ignore the upper part of the chords, and play D (maybe with simple patterns using fifths and octaves) over Cm7/D.

For example, take a look here:
https://drive.google.com/file/d/1IsK_91ZdvXoSdPRQ4IYd-2Bn1hfBhjMC/view?usp=sharing

RealTracks in song: ~903:Bass, Electric, SmoothCool Sw16 100
RealTracks in song: 2765:Bass, Electric, PuertoRicoAguinaldo Ev16 120
RealTracks in song: 3111:Bass, Electric FunkSimpleAlex Ev16 110
RealTracks in song: 2620:Bass, Electric, FunkSwing16ths Sw16 085
RealTracks in song: 2607:Bass, Electric, ModernSlapFunkAlex Ev16 110
RealTracks in song: 2960:Bass, Acoustic, SmoothBalladNeil Ev16 065
RealTracks in song: 3049:Bass, Acoustic, SpanishTangoNeil Ev 120

Only #903 is playing correctly, all the other ones are simply ignoring slash chords instructions, playing Cm7 patterns over Cm7/D chords, which is plain wrong.

Thanks for pointing these out. We'll take a look at these and see what we can do.

ps It certainly used to work, so that may be a clue.

Hmmm... Was making a tune and realized that guitars do not "catch" slash chords...remembered somebody starting this thread. I assume the problem is global... and hopefully it will be resolved with next patch.

Instead of spilling my problem in detail, want to post in this thread confirming that I too experience the issue with slash chords. If tech team needs my SGU, please contact me and I will share.

Thanks.

PG Music people : Don't lose the initial spirit of my post.

Not implementing this chord Dm9/C, for example, would be 'acceptable' if we get Dm7/C

BUT

Having frequent chords as C7/E or Dm75b/Ab not working for many RealTracks is not acceptable IHMO, for a professional reputed software.

That is what I mean in my post.

Today the slash chord feature for RT is working in about only 50% of the slash chords.

I understand. In my case, I need Dm9/C to work; I had a song with this just yesterday. But I get your meaning if it comes to priorities.

The odd thing is, slash root used to work fine whatever the chord.

All the comments in the posts demonstate that it becomes urgent for PG Music to fix this request and to make sure that new coming Bass RealTracks have not this bug.

Otherwise, as long they are generating new RealTracks without checking them according to my post, they generate a potentially bugged RT, whilst they know the problem.

So a new Bass RT and potentially a new bugged RT.

Seen the oustanding list of bugs waiting for a fix, I'm not sure that is what they want.

John-Luke.
I think what Matt is trying to point out is this:

"The odd thing is, slash root used to work fine whatever the chord."

I thought so too...

Maybe Dave D (of PG Music Staff) can shine some light on this.

+1

Also seems to me that all the issues you raise can also be bundled with the one i raise here:

"Bass RealTracks: fix the unrealistic bass-line "note-jumps"

https://www.pgmusic.com/forums/ubbthreads.php?ubb=showflat&Number=513300#Post513300

This has already been asked, but

+1

Yes, the algorithm does a wonderful job already, but in some aspects it could still be improved.

So +1 also here

John-Luke, I'm all for your proposal, but I'd like to point out one segment of your comment that I believe is in error. You wrote:

"For example : Bass RT N°2426
C2 chord / 2 beats : the bass plays G - E, sometimes you may get the notes C - E, that is correct."

BiaB spells the Csus2 chord as C2, but incorrectly includes the E, the 3rd of the chord, which does not exist in either sus2 or sus4 chords. The major 2nd is substituted for the 3rd with a sus2 and the 4th is substituted for the 3rd with a sus4 chord.

This is a recent pet peeve of mine -- specifically the misspelling on the staff of the sus2 chord by including the 3rd -- and I have notified PGM that they are not following universally accepted chord theory. I hope they correct this. Instead of a C2, they are actually creating a Cadd9, which is not what one needs when one wants the suspension.

Member Noel pointed out to me that one can use another sus4 chord's inversion to arrive at the same result. E.g., a Csus2 is also a Gsus4/C. And of course making sure that BiaB behaves when this /C is indicated is, of course, essential.

Good point cooltouch !

You're right : PGM don't follow the standard of definition of some chords, thus adding some difficulty to understand each other.

In my mind, according to what I know of the PGM realization, the "C2" chord does include the E note, that is why I said that the C - E bass notes were correct.
Absolute speaking, the C2 chord do not include the E note and so the C - E bass notes are wrong.

Cooltouch is absolutely right. BIAB has three chord types that are all implemented as a 2 chord. The following is part of the contents of the file /BB/Data/pgshortc.txt where this is clear. I highlighted the 2 chords:

Note these are already defined in BB, and you can't over-ride these shortcuts
;sus2@2 add2@2 m+@m#5 dim7@dim 7aug@7+ sus4@sus 11@9sus 9sus4@9sus
;13sus4@13sus add9@2 7sus4@7sus s@7sus h@m7b5 d@dim J@MAJ7
; f@7b9 m7+@m7#5 NC@C. no chord l@7Alt u@(Blues) 7u@7(Blues)
; Lyd@lydian Lyd@lyd Lyd@4#

After the last update, all bass realtracks on the example plays correctly.

Thanks!