Audio Detection sucks?

I listen to EDM and because I use FL Studio my method of audio detection is to drop an audio file on to the plugin. When I slice the reference track and try to get Scaler 2 to detect the chords and melody, it fails, even though i try to get audio that has the least amount of other instruments playing. Anyway to fix this?

Hi bigloo,

What type of audio file do you use. If you tried mp3, this will not work. You need .wav .I use Reaper and detect audio (.wav) quite frequently (even live from my guitar) and have no issues. However, I don’t think you’ll be able to detect both chords and melody. Oh, btw, do you use Scaler Audio? If not, that’s the issue.

It is a .wav file. It does detect the audio, just gives me the wrong notes and chords. I am not sure if I have Scaler Audio, but as mentioned the audio detect option is there and works when I drop the file onto it, just not correctly

P.s If you cannot detect audio and chords at the same time, how do you detect either in a reference track .wav file where the chords and melody are always playing at the same time?

In your VST list you should have “Scaler 2” and “Scaler Audio 2” as two distinct entries. So, you should use the latter.
image

My bad: I believed you wanted it to recognize and export chords as one midi and melody as a separate midi (like in “Save as MIDI”).

Scaler audio detection caller is about detecting chord sequences, and if you want to pick up ‘melody’ there are other tools. You need to use Scaler Audio and drop it on an audio track, and not vanilla Scaler which sits on a midi track, Scaler Audio should have been installed with the basic installation.
The mixer of the piece would have attempted to give the various sonic elements ‘space’ and avoid too much in the way of overlapping frequencies. Hence, one approach that helps in a full mix is to try and pull out (say) a pad with a band pass filter and that will give the apps algorithms a better change of detection.

I’ve found the detection needs editing 3 times out of 4, but once you have a key and a basic progression, it’s normally straightforward to figure out what the best fit with the music is.

Remember also the voicing will make a difference to the chord analysis - this is a reality. Finally, AFAIK from my experience with feeding in audio (mostly trance / EDM stuff) Scaler looks for changes in the input, and disregards time intervals; so if you had 4 bars of C C G F, Scaler would most often throw up 3 bars as C G F. Again, this is easy to fix.

It’s an inexact science - don’t expect to feed in John Cage’s ‘Living Room Music’ and get any chords out …

1 Like

Also… In Scaler 2 Audio, use Detection Threshold slider to set the signal level, in order to obtain more accurate detection.
image

1 Like

I found deCoda (per this Forum’s recommendation) very useful for that purpose.

Hi, thanks for the reply

I have now loaded the relevant audio clip as a channel in FL Studio, and have loaded Scaler 2 Audio as a plugin. However, I still get terrible results, and sometimes doesn’t give me anything at all

I followed up on @Bernd 's suggestion of deCoda., and that has worked fine on all examples I’ve tried to date. I’ll report in due course, and might do some comparison tests with Scaler vs xxxx.

2 Likes