Dorico Blog: Transcribing Jazz

I noticed this nice feature on Max Gámiz, the transcriber for Open Studio, was published yesterday:

While it’s a good write-up on Max, and Open Studio is a great company, it seemed like a bit of an odd choice for a blog post as Dorico doesn’t really offer any tools to specifically assist with transcription. There was even this exchange with the interviewer:

As I also use Dorico to transcribe daily, I can think of a lot more “features or workflow improvements” than that! If features to assist with transcription are under consideration for future development, here are a few things that I think would be helpful. I realize many people will think Dorico should stay in the notation lane and not bother with these, but they would all assist with transcription and eliminate the need for a bunch of third party apps.

1) Import audio. The ability to transcribe audio obviously requires an audio file. As of now in D6, Dorico can’t import audio files. Converting to a video file is an unnecessary step if you don’t need video.

2) Audio sync points. Recorded audio is not always exactly fixed tempos. Once we can import an audio track, the ability to set specific sync points in the score would be very useful. (I haven’t used the video feature much, perhaps there’s some capability with video sync now that I haven’t figured out.)

3) Piano audio. Most of the time when I’m transcribing, whenever I touch my MIDI keyboard I just want it to sound like a piano, not like whatever instrument I’m inputting. Using an external standalone piano VST is just an unnecessary step when this could be handled within Dorico. Windows users then are required to use an additional MIDI routing program as well so both the standalone VST and Dorico can receive the MIDI signal simultaneously. That’s three pieces of software when this could all be within Dorico. Finale has had this feature for as long as I can remember. (Set MIDI Thru to a fixed channel with the piano VST.)

4) Audio speed control. When transcribing, the ability to slow things down while keeping the same pitch is essential. I usually work in with Transcribe! and most often have the speed set at 70%. Occasionally for a really fast passage I’ll have to drop it down to 50%.

5) Foot pedal control. I typically leave Transcribe! active in the background and control the playback with my MIDI foot pedals. That way I can input directly into Dorico while staying focused on the music, can quickly review passages, etc. If playback of an audio file is implemented, some shortcuts to control it with MIDI pedals would be great. In Transcribe! have my right pedal set to toggle Play/Pause and the left set to Review then automatically start Play on release:

6) Pitch control. Not every recording is A440 so some fine tuning of playback pitch to match the piano VST is required. Being able to change the playback pitch by a couple of semitones when transcribing a Bb instrument is nice too. Finally, being able to change the playback by an octave is very useful when transcribing bass lines. If I wanted to transcribe this bass line …

… taking it up an octave into the “Alvin & the Chipmunks” range gets the bass out of the mud. The bass notes now pop out and are much easier to hear:

7) Stem splitting. Getting a bit far afield from notation now, but Steinberg already has Spectralayers so could integrate some very basic functionality from it. Stem splitting technology has come a long way and is pretty amazing now! I personally don’t use it to isolate instruments that much, as it sometimes loses some of the ghost notes and nuances, but just removing the drums from a mix makes things so much easier to hear as the drums affect the entire frequency spectrum. Here’s the 8 bar audio from above after applying stem splitting and muting the drums:

Anyway, if Dorico is going to do blog posts on Transcribing Jazz, those are some of the features that I feel would really help with transcription!

I guess they know who to go to for the next Jazz blog!

I’d be happy if we just got point 1,2 and 3!

Haha, happy to help out! Open Studio is a pretty great online education company though so it was nice to see they are using Dorico.

Yeah, I mean those aren’t exactly new requests LOL! People have been asking for those for years and years. It seemed odd to me to do a blog on using Dorico for transcription when you can’t even import a .wav or .mp3 file. #4-7 are obviously more transcription specific requests.

There’s certainly a balancing act between developer time, feature bloat as tools unrelated to notation are introduced, and marketable features that will sell licenses. If they are going to market Dorico as useful software for transcription, I think all those elements will help. It would be nice to eliminate a program or two from my transcription routine, but I’ve got my own setup pretty streamlined by this point anyway.

I concur to everything you wrote :folded_hands:

Let me add this : I would be really glad to read such article with you, as a long-time Dorico user, interviewed :wink:

My undergrad advanced improv class is filled to capacity this semester with 15 students. Each student is assigned a “Transcription Project” where they have to select and transcribe 5 solos by the same artist or 5 solos by different artists on the same set of chord changes. So that means I basically have to do 75 solo transcriptions myself this semester just to be able to grade theirs. Plus all the transcriptions I do to hand out in class to demonstrate different techniques. Lots of transcribing ahead of me this semester! Jazz transcription workflow with Dorico is definitely something I’ve spent a lot of time trying to get as efficient as possible!

“Author, author!”

It’s outdated now, but I used to use a program called “Slowblast!” that slowed audio down but maintained pitch. A tool like that in Dorico would indeed be extremely useful :wink: There is also the ability to do this in Band-in-a-Box-even stem splitting, I believe. You can slow audio to 1/2, 1/4, or 1/8 speed without affecting pitch.

Amazing Slow Downer is another one that is pretty common. Of course, YouTube allows it as well:

Most DAWs can do it and then let you export audio too. If we ever get the ability to import an audio file into Dorico, it would be pretty great to be able to control the playback speed directly within the program.

Indeed it would :slightly_smiling_face:

This is also in Cubase now, and I think it is based on open-source unmix AIs. If we ever have an audio import (and time sync) capability, it would make sense for this to support multiple audio tracks, and likewise, it would make sense for an un-mix process to put stems into those separate tracks for us.

I also use Transcribe with a foot pedal. I wouldn’t mind continuing to use a separate program, but it seems like the Transcribe author isn’t really interested in taking the program farther along the lines you suggest.

Transcribe has a “chord guess” feature that sometimes is useful, but could be a lot better. It would be really cool to see Dorico present several potential chord spellings at each of the big beats, allowing me to click on the one I like, to enter that chord into the Dorico score.

While we’re at it, it would be nice, when editing an existing Dorico chord, to have the option to jump into the chord wizard mode (see Cubase – I don’t recall what the feature is called), where I could try different color tones, inversions, or adjacent/related chords.

I honestly mostly use this feature in Logic now. I generally don’t need all the control Spectralayers has and just want to remove the drums quickly. Logic’s stem splitting feature is so fast and generally so good at removing the drums that I usually just do it there. I haven’t tried it in Cubase yet.

I find this feature completely useless, LOL! I don’t think it’s ever once been helpful for me so I just leave it off so it doesn’t lead me astray. The “Show Note Guesses” is also mostly useless, but occasionally has been helpful, mostly when I’ve been stuck hearing a buried harmony note in a large ensemble transcription. Certainly any “Note Guesses” need to be confirmed and verified with my own ears, but it has assisted me with sleuthing out some inner harmony parts before.

I can’t compare with Logic, but the Cubase feature sounds like what you descibe. Very fast and integrated in the normal Cubase workflow. Maybe not as precise as Spectralayers, but much faster to a result good enough for transcribing, so I hope that if we get multiple audio tracks in Dorico, the unmixing will be there the first day, tightly integrated.

I only mentioned it as a feature that SHOULD be good, and could be very helpful. I agree it doesn’t help much as implemented in Transcribe. I generally know what the chord is, or what chord I want to use – I don’t often transcribe 100% literally; I often rehamonize a little – or a lot.

But there have been times when the sound is so muddy that I have tried to get some guesses from Transcribe. The problem is that they flash by so quickly; you really have to zoom in and scrub very carefully over the waveform to see the various guesses. Sometimes that has provided some insight, but not very often. I’d like to think that a good AI program could:

  • identify the rhythmic positions where there is a distinct chord (which might include passing tones and such)
  • compile a list of different chord spellings, ranked from most likely - to less likely - to other tonalities you might want to try.

Functions like that would be very valuable to me, and if it worked well, I’d pay good money for an add-on or whatever.