I played a melody in B Phrygian then played into Scaler after hitting record. It detected the notes and then suggested many scales which were incompatible with my melody. Does anyone know why that is? B Phrygian only has an F# and it is recommending scales with other sharps or flats. That doesn’t work.
Also is there a way to filter the scale recommendations if you tell it the root note?
Thanks in advance.
Welcome to Scaler ! You will find this community useful, I hope.
I’m assuming what you are trying to do is to determine a chord progression to fit against the melody you have (Scaler doesn’t play melodies).
The root note won’t necessarily help determine the scale.
Scaler is essentially about chords, and that, I think, is the primary target of the detection. Figuring out a scale from a series of single notes is not always deterministic.
For example, you could have been playing B minor pentatonic, which only has a F# in, and your melody might not have had a flat second.
Or it could have been a Melodic Minor scale B Dorian b2, but you didn’t play a G# in the melody, but it has a flat 2nd and F#.
Could have been a Harmonic Minor mode, B Phrygian major (flat second and F#) but if your tune didn’t have the third in it, Scaler couldn’t distinguish between it and B Phrygian.
And or course if you were ‘playing outside’ fusion stuff or used a chromatic passing note somewhere, that would have influenced Scaler’s choice.
So the test is how the scales it came up with map against the notes you played.
Scaler is used to construct chord progressions. Having established a progression, however, any of the performances will create melodic lines over that progression. However, this is not relevant in your case, since you already have a melody (although you could use Scaler to generate a bass line or rhythm parts).
In your case, I’d set the scale to B Phrygian in section A, and then use the harmonised chords in section B to drag into section C against your melody by ear. This is normally fairly straightforward to do as first pass, as most melodies tend to use only 4 of the harmonised chords, and in any case, the flat 2 is easy to pick out.
You can then embellish those using the ‘suggest’ feature, or jazz it up by using 7ths, 9th etc in Section B, or altering the voicings.
It seems linked somewhat to what I said here
I asked to @jjfagot, and he said the difference between my 2 scales is little
And he added:
“Despite that half-tone difference between the 5th degree of both scales, I will tell you that sometimes dissonance is pleasant and avoids boredom. If everything were tonal, without dissonances, we would not evolve.”
Hi Panda, No I’m playing the melody into it to determine the scale. To find out what scales are compatible with my melody…
If I have a F# in the melody then you can’t play a scale that has F natural for instance or if you have a G natural in the melody, then you shouldn’t suggest a scale that has a G#.
I believe this is what the scale recommendation function is supposed to do.
Sorry I couldn’t follow your example.
What I mean is that the accompaniment and the solo scales can be different, but may work well together anyway
So, are you sure that the scales suggested by Scaler don’t really work?
Yes because the suggested scales would have to include the notes in the melody.
Many of the suggested scales did not.
Well when you do capture your melody like that, and it lists a number of scales, Scaler will tell you how many of the notes you played fall within each scale. If you played 6 notes, it will show something like 6/6 or 5/6 or 4/6 next to each scale, showing the number of notes in or out of each potential scale.
Just because a note doesn’t fall within a particular scale doesn’t mean you can’t use that scale. People play music in C major and still use sharps and flats as passing tones all the time. Just as the most basic example. How important each note is that it be part of the scale you select is up to you. That’s kind of where the art comes in. 
Did Scaler not identify B Phrygian as one of the options?
You are correct, an I agree with you entirely, but that wasn’t my point. I was describing the Scaler determination process, or indeed the general process for determining what scales having S notes a subset of N notes would uniquely match with. If N<S then there may be more than one scale which match the N notes.
So if you play a melody with root note C into Scaler which has the notes C E G A and B in, what scale should Scaler pick ? Even in the diatonic set it could be Ionian or Lydian - there is no way for Scaler to tell. The fact is that there isn’t a unique and definitive answer here; Scaler cannot possibly tell what scale is involved until it has either all notes or a subset of notes which would uniquely match with one and only one scale of the thousands that there are.
A melody would not typically contain every note of the scale/mode that it is in, but only a subset (so N<S), and more than one scale might contain that subset of notes. So it’s not scaler that can’t give you a definitive answer, it’s just that there might not be one.
So in my example, you might play 6 notes of melody and Scaler could not know whether it is Ionian or Lydian. So as @RBIngraham says Scaler will tell you how many notes of you played match with the notes of each suggested chord. In the above example, Ionian and Lydian would match 100% with the data it has i.e. the notes played. There is no ‘right’ answer in this case.
As you rightly say, the root note is key here to determine the modes which share the same notes, but the key doesn’t help with parallel modes as in the above example.
Ok, I see the problem now - it lists all the scales even with partial matches, like you said telling you how many were matched e.g. 5/5, 2/5, etc.
It goes all the way down to 1/5, so basically it’s going to list every scale in existence.
So I guess my new question is: Can I view only the scales that are a 5/5 (complete match) on the list?
Also when I go from Detect mode to Scale mode in order to further filter the scale suggestions list, for example by telling it the root note, it still lists every scale under the sun, even the ones that have only a 1/5 match! What use is that? This is a major flaw no?
And to make things even worse it no longer indicates how many notes were matched for the scale e.g. “5/5”, “4/5”… on the list.
So some other questions are: like my previous question, can you view only the scales w a perfect match 5/5 in this Scales mode?
Or can you turn on an option to view how many notes were matched on the scales of this list?
When I do it the same thing here, Scaler puts all the results in order by how many notes match. So any scale that is 5/5 is at the top of the list and then it proceeds down to 4/5, 3/5, etc… Does it not do that for you?
If it does do that for you, then how hard is to scroll to the top of the list and see all the 5/5, or how ever many notes you have, options? Maybe you’ll find what you created wasn’t in the scale that you thought it was after all. Or there are better options?
Because I don’t play the piano (other than to trigger MIDI things) I don’t really use this capability, but I had a go., and I sort of get your point from a practical perspective.
I tried with various combinations and what it seemed to me from a quick test is that one way to improve the selection and order is to focus your input melody on the ‘distinctive’ notes in the scale which you suspect it might be. These might be a b2, or a b5 for example. Better still with a combination of notes which might be distinctive, so b3, 3 and b5 would knock out a lot of stuff.
So given your root note you can list out the degrees in your melody, and then look to see which might be an odd combination and enter those, plus either a 3 or b3 to tell Scaler whether its major or minor (which presumably you can tell by ear) etc.
But as I say, I’m a one note per second player.
I think I might be mistaken in how it works - I thought after you detect the scales in Detect Mode, then when you switch to the Scales Mode, it lists only the scales that were detected. However now I think Scales Mode simply lists all the existing scales regardless of what was detected, is that correct?
In Scales mode you have the filters where you can specify a root note or type of scale and it will filter the list…
In the case where if it showed only the detected scales, I wanted to view only the scales with a 100% match (to filter further by the root note).
Yes, if you use the detect method to capture chords or a melody, then all the information you need or that Scaler will provide is there as soon as you hit the stop button. Actually, it works almost instantaneously as you are “recording” if memory serves.
If you then try and set a root note in that one drop down box you’re actually leaving the detect mode I believe. I’m not sure exactly what happens then as I don’t think I’ve tried that.
But the short version is that you’ll see any 100% matches for the notes you captured in detect mode, right away and you don’t need to do anything else. It will automatically order the scales with the most likely options right at the top of the list and the further you scroll down the list the less likely they are a match for what you captured.
Actually, if you only play a few notes e.g. 4 notes, the 4/4 matched scales list can be quite long. That’s why I wanted to filter that further by specifying the root note.
I can probably find out by example whether that list stays intact when I go the Scale mode. I’ll try that later.
Thanks for your help.
Hi @Nora1 welcome to the forum and thanks for sharing your thoughts.
I have had a good read through this post and have discussed this with the rest of the team and we agree that the ability to filter your detection would be a good addition. We have added this to our roadmap now.
As it stands the scale browser and the search bar do not relate to detect mode and we can definitely see there is scope to include this functionality or something like it.
That sounds great. Thanks for listening.
Since I might still have your attention, I want to make you aware of a problem I discovered. I’m not sure if it happens at all resolutions, but it happens at the default resolution and the resolution settings shown on the last image of this post.
Basically what happens is on the detection screen, after you get your list of results - when you scroll down either by clicking below the little square on the scroll bar or carefully scrolling down one page (by clicking on the little box and dragging on the scroll bar) it either skips one item on the list or only partially displays it. Here are some screen captures:
(Had to remove image because new users can only insert one image)
The above screen shows a list of results after detection. You can see the top of the next list item which says “C Mixolydian mode”.
Then if you click below the little box once on the scroll bar, it shows this above. You can see that it scrolled down a little too much and clipped the top of “C Mixolydian”.
(Image removed)
Now if you tried scrolling down while dragging (I have a feature on my mouse to drag really slow and precise) it skips displaying “C Mixolydian” altogether on the list.
(Had to remove image but resolution is: 1400x980 and font is POPPINS.)
These were the settings where it occurred. It also happened at the default resolution and font.
——
I also have a suggestion that would make a great feature - What would be cool is if you can customize voicings of chords that span more than one octave (say 2-3 octaves). For example, with seventh chords you can have a screen where you can customize a voicing by specifying the Root, 3rd, 5th and seventh of the chord you would like to hear over say a 3 octave range. You can put as many or as little of each note (root/3rd/5th/etc…) anywhere on the map. For example: I can put the root in the lowest octave, then the 3rd on second octave then 3rd and 7th on the highest octave. Then you can play the diatonic chords with one click using this voicing map the same way like your other preset chord voicings.
Hey @Nora1 I’ll let James answer the rest but this (Extract Voicing) feature allowing to do what you want is already there:
Hi @Nora1 it does appear that the scrolling does cut a little off when clicking and dragging the bar. One thing I will say is that scrolling with a scroll wheel, trackpad or your finger on the top of an Apple magic mouse instead results in a very seamless scroll.
I will lodge this with our developers to take a look at as it does seem like it could be refined.
I’m on a PC w Windows 10 using a Kensington trackball mouse.
I have a mouse utility which slows down the mouse speed to the slowest setting so even w that, using the scroll bar it happens.
But regardless of the mouse, clicking below the square on the scroll bar, briefly once, should go down one page exactly. In this case it scrolls down too far and skipped 2 items in my case (the “C Mixolydian”, “F Mixolydian”) entirely.
Screen after scroll:
