Fastest lead sheets - hand or app?

[dixonge]

I don't know why, but I really thought that by now you could just play a series of chords and have them pop into a staff in the appropriate key/time signature. Then maybe click a button and have the chords show up above the notes. Somehow that seems iffy at best, downright impossible at worst, depending on the app. Does no one transcribe original tunes any more?

 

So I'm working on a project, started in on one song, found the chords embedded in a mish-mash of sound effects and ... I have to write them out by hand, I guess?

[Mighty Motif Max]

I do it “by hand”, usually Finale. Never had any luck otherwise.

[obxa]

Long time Finale user and chart writer here.   Last weekend I did a large festival gig for a local Roots-Americana  artist who was headlining. 

 Had to learn 17 songs in 2 days.  13 originals,  &  4 covers.  Some of the songs were more complex than others, and some of the covers were things I kinda  already knew, just needed a basic chart.  Also was playing accordion and melodica and had to keep track of which songs and parts used them.

 

 

 

First I downloaded some Music notation keyboards for Ipad/Iphone.  They work like emojis. One was called: "Musicians Keyboard" the other was a free app: called "Song Memo_.  Highly recommended if you need to make quick legible charts.     The keyboard that Song Memo installed was great, and can use it with it's app, or any other note taking app. It's has chord shortcuts, barlines,  you can enter relative chords  just by picking the key. Very cool.  

 

That said : like Mighty Motif Max, writing a real chart always works better for me.   Notating arrangement and signature riffs/line is essential for (lots of) unfamiliar material and by doing so,  I end up actually learning the song better too.   Unless charts are provided,  it's always been my MO and it ends up being faster in the long run.  So ended up doing that for all the originals, and one of the covers that had some odd measures and mixed time sig stuff. 

 

One  indispensable thing I use is an app called "Capo" also free.    It can loop sections. Change keys and tempo, and tune the track.   It spits out the chords along bottom of player-  not that I need that but never hurts to get me in the ballpark.    I'll slow down lines and riffs to notate.  I often also play songs faster than the performance tempo  to check the chart or do a quick run-through.  Sometimes will run an entire show at fast tempo just to get a quick rehearsal in.      There are similar apps like "Amazing Slowdowner", or Hornet's "Songkey"  that do the same things.  Forscore's built in  MP3 player does turning, transposition and speed. 

 

 

If I ever need to just do something on the fly, using notepad and one of those music emoji keyboards would work.  But for music I might end up needing again in future, real charts all the way.     BTW-  The artist  thanked me several times for doing my homework and knowing his show cold - got a bunch of future dates because of it. 

 

[ElmerJFudd]

^^^
 

everything is always which tool(s) you are most comfortable with.   it took time as a chart writer to learn how to use pencil and staff paper.  it takes time to learn Finale or Musescore, etc.   it’s fast and easy when we practice and things become fast and easy.  Thankfully, lead sheets don’t ask for much, right? Chords and melody.  

[Reezekeys]

99% of what I do is lead sheets - just melody & chord symbols, occasional grand staff stuff. I always sketch out in pencil on manuscript paper, because I usually have to do erasing and "reconfiguring" when I transcribe something - often finding out I can shorten a chart by putting a DS in, or that one section is a copy of another previous section, etc. I find this way easier with paper, pencil, and eraser. Pausing to work software at this point breaks my concentration and flow.

 

When I'm done with this process and have the "final" version of how best to make the chart, I fire up Sibelius and quickly copy my penciled, marked-up & erased-over scrawl into the software to make it nice & pretty.

[dixonge]

11 hours ago, obxa said:

One  indispensable thing I use is an app called "Capo" also free.    It can loop sections. Change keys and tempo, and tune the track.   It spits out the chords along bottom of player-  not that I need that but never hurts to get me in the ballpark.    I'll slow down lines and riffs to notate.  I often also play songs faster than the performance tempo  to check the chart or do a quick run-through.  Sometimes will run an entire show at fast tempo just to get a quick rehearsal in.   

Capo is the bomb!

 

I mean, I can do it by ear, but at least on this tune, with it's 'lush' very distracting soundbed and weird levels that app would have helped. It also caused me to notice I was switching chord references around (E-flat instead of D-sharp, etc.). Very helpful, thanks!

[Mills Dude]

On the topic of why there isn't some software that can instantly analyze what's being played and crank out a lead sheet with chord symbols, let me take stab.  I'm a programmer guy, 30+ years in the biz and a pretty good one, but as a referred to in another thread, I'm not a computer scientist.  

 

I'd say the 2 biggest reasons as to why this doesn't exist.

1. Money - Nobody with deep pockets willing to foot the bill
2. Lack of research interest for Computer Scientists

Natural language recognition has been a holy grail for computer scientists for a long time.  It's a set of difficult problems to solve and there's significant money to be had for the research required to solve those problems.  Today, language recognition is pretty good, much better than it was 10 or 20 years ago.  I can use google translate (or others) on my phone and it works wonders, even allowed me to get by in China for a few years.  Its everywhere now, "Hey Siri? ..." and it's changing the world. 

 

The big tech companies have put a lot of money and research to tackling the problem.  They have funded academic research and hire very smart computer scientists and pay them well to solve these problems.  Now, after about 20+ years of work and research, language recognition is quite good.

 

Music recognition seems like a much easier set of problems to solve and it probably is.  Even given that the primary use case is to use a midi enabled keyboard as input as opposed to attempting to analyze audio, it should be easier to accomplish and it probably is. The companies producing notation software have gotten quite good at interpreting midi input to notation but we all know its far from perfect.  Frequently requiring a long process of corrections that make most people think it would have been easier just to notate from the start, instead of trying to tame the stream of notes into something usable.

 

When I look at a written score, I know and understand the rules that say -- Oh that's an A/C# which is a very simple interpretation as opposed to trying to analyze a more complex piece.  Quantifying those rules into a set of algorithms -- well I don't know if I have the time, resources and ability to do that.   I'm sure the same people at Sibelius, Finale, etc. say the same thing.  Its a niche market and those guys are probably struggling to keep their operations in business dealing with us consumers who constantly balk at the price tag they put on their products.

 

Capo is probably the leader in this right now and I bet at the heart of that organization is a passionate coder willing to take the risk and dedicate the time and resources to making something that  works.   I hope they can be successful and continue.  The market is niche and in my view will continue to narrow.  On the positive note, none of that work will go to waste and if that company doesn't make it, someone else will pick up the mantle and move the football further down the field.  So maybe another 10+ years, we'll have software that just works the way we expect it to.

 

tl;dr -- While it seems that it should be straightforward to do, it takes a lot of money, research and willingness to solve that problem.  Not sure there's enough of a market to support the work required to produce something works like we expect.

 

 

[obxa]

Only thing to watch out for with Capo and most of those Chord identifier apps, is they get confused with slash chords.  (E.G.  C over E) .  But hey, when you're in a crunch to learn a bunch of stuff, still nice to have some help. 

 

[RTPRocker]

If I have to do 'real' notation I have used Finale, Sibelius, even Logic for a lead chart. But if I am under a huge time pressure to create a cheat sheet chart, nothing beats the speed of hand-writing the song and then scanning it. Just write the chords, add slashes for timing as needed, and use a different color for any lead lines. The attached I did in about ten minutes. My record was preparing 25 songs I had never heard before for a performance in 2 days. 

One In A Million.pdf

[ElmerJFudd]

49 minutes ago, Mills Dude said:

On the topic of why there isn't some software that can instantly analyze what's being played and crank out a lead sheet with chord symbols, let me take stab.  I'm a programmer guy, 30+ years in the biz and a pretty good one, but as a referred to in another thread, I'm not a computer scientist.  

 

I'd say the 2 biggest reasons as to why this doesn't exist.

1. Money - Nobody with deep pockets willing to foot the bill
2. Lack of research interest for Computer Scientists

Natural language recognition has been a holy grail for computer scientists for a long time.  It's a set of difficult problems to solve and there's significant money to be had for the research required to solve those problems.  Today, language recognition is pretty good, much better than it was 10 or 20 years ago.  I can use google translate (or others) on my phone and it works wonders, even allowed me to get by in China for a few years.  Its everywhere now, "Hey Siri? ..." and it's changing the world. 

 

The big tech companies have put a lot of money and research to tackling the problem.  They have funded academic research and hire very smart computer scientists and pay them well to solve these problems.  Now, after about 20+ years of work and research, language recognition is quite good.

 

Music recognition seems like a much easier set of problems to solve and it probably is.  Even given that the primary use case is to use a midi enabled keyboard as input as opposed to attempting to analyze audio, it should be easier to accomplish and it probably is. The companies producing notation software have gotten quite good at interpreting midi input to notation but we all know its far from perfect.  Frequently requiring a long process of corrections that make most people think it would have been easier just to notate from the start, instead of trying to tame the stream of notes into something usable.

 

When I look at a written score, I know and understand the rules that say -- Oh that's an A/C# which is a very simple interpretation as opposed to trying to analyze a more complex piece.  Quantifying those rules into a set of algorithms -- well I don't know if I have the time, resources and ability to do that.   I'm sure the same people at Sibelius, Finale, etc. say the same thing.  Its a niche market and those guys are probably struggling to keep their operations in business dealing with us consumers who constantly balk at the price tag they put on their products.

 

Capo is probably the leader in this right now and I bet at the heart of that organization is a passionate coder willing to take the risk and dedicate the time and resources to making something that  works.   I hope they can be successful and continue.  The market is niche and in my view will continue to narrow.  On the positive note, none of that work will go to waste and if that company doesn't make it, someone else will pick up the mantle and move the football further down the field.  So maybe another 10+ years, we'll have software that just works the way we expect it to.

 

tl;dr -- While it seems that it should be straightforward to do, it takes a lot of money, research and willingness to solve that problem.  Not sure there's enough of a market to support the work required to produce something works like we expect.

 

 

Also music as a language is a big deal, it’s way more complicated than OCR or frequency analysis.   This is the kind of job that requires AI.   Getting a robot to play music as a human would from notation (especially hand written) and imagining how something we play might look in notation requires high level reasoning.  Starting with knowing what each instrument is capable of and not, instrumental technique, forgiveness of time stretching and contracting, forgiveness of center of pitch, recognizing what is a bend, half or full, etc.  It’s a large body of knowledge and the rules change with style.   

[jarrell]

I think the reason why there aren't currently apps that can listen to audio and write a good lead sheet, is that it is a very hard problem. While $$ may be a factor, I am not aware of research that does this very well.

 

There are three steps:  polyphonic pitch detection and musical analysis, and then chart design.

 

You do have Melodyne which can do polyphonic pitch detection, but it isn't perfect. Of course, MIDI input skirts this issue.

 

Translating the resultant pitch detection into a lead sheet with chord symbols, rhythmic slash notation for hits, etc., writing appropriate repeats, 1st and 2nd endings, D.S./D.C. and coda, solo section with just chords etc. - in other words, all the things you would normally do when writing a lead sheet, requires an extensive domain knowledge.

 

It'll happen, just like the singularity, but to work as well as a good musician writing good lead sheets, it will take some number of years before it happens.

[analogika]

One thing that really annoys me is that I haven't found any software solution that lets me freely combine notation and text.

 

E.g. write out

a hookline,

an A Theme, and

a B Theme, all with melody and chord changes. 

 

Then free space to write as text 

"A Theme (sin rep)" 
"B Theme (organ solo)"

"A Theme (with drum breaks)"

 

Then another line of notation and chords for the C part

 

Etc. 

 

The way I work now is to take a sheet music lined blank PDF and write it out by hand on the iPad, directly in the Files app. 

 

I'd really like to be able to just MIDI-record some of the more complex stuff for speed and combine computer notation with the above free layout. 

 

Nothing seems to be able to do this, unless I arrange screen-shot notation and text with a layout program like Pages. 

[cassdad]

1 hour ago, analogika said:

One thing that really annoys me is that I haven't found any software solution that lets me freely combine notation and text.

 

E.g. write out

a hookline,

an A Theme, and

a B Theme, all with melody and chord changes. 

 

Then free space to write as text 

"A Theme (sin rep)" 
"B Theme (organ solo)"

"A Theme (with drum breaks)"

 

Then another line of notation and chords for the C part

 

Etc. 

 

The way I work now is to take a sheet music lined blank PDF and write it out by hand on the iPad, directly in the Files app. 

 

I'd really like to be able to just MIDI-record some of the more complex stuff for speed and combine computer notation with the above free layout. 

 

Nothing seems to be able to do this, unless I arrange screen-shot notation and text with a layout program like Pages. 

Right, that’s what I mostly do….   For example, let’s say I want to notate (transcribe) the main riff/hook of a song.  I just play it into Logic Pro via MIDI and record it…..   then take a screen shot of the notation, and then paste that into Pages or any word processing program…..  then I can write out say, verse and chorus chords…..  then save the resultant document in Forscore, and then I have everything I need for that song.  YMMV.

[dixonge]

ok, so on Musescore, once you have your Midi keyboard connected and recognized (had to switch out a cable) you can go into 'Note' mode, pick a note (whole, half,  quarter, etc.) then play your chord. Those notes will then enter the score as designated (whole, half or quarter notes). You then connect everything as needed (dotted, tied, etc.) and then the app can play back. Entering the chord symbols starts with Cmd+K then type it out, hit space to move to the next note. That's about as fast as it gets. I'm so rusty at this I have to look up what things are called (repeats, codas, etc.) so I can look up the appropriate key shortcut.

 

You can then add notes (Verse, Chorus) and lyrics. In the end, you have a printable score! 

 

I wish I could just use Chord Symbols and slash fills, but that doesn't give me any way to remember when chords or notes aren't exactly on the beat. Will probably need to memorize those things any way, but hey, one step at a time...

[analogika]

41 minutes ago, cassdad said:

Right, that’s what I mostly do….   For example, let’s say I want to notate (transcribe) the main riff/hook of a song.  I just play it into Logic Pro via MIDI and record it…..   then take a screen shot of the notation, and then paste that into Pages or any word processing program…..  then I can write out say, verse and chorus chords…..  then save the resultant document in Forscore, and then I have everything I need for that song.  YMMV.

Yeah, exactly. Logic for notation, screenshot, paste into Pages in layout mode, add structural notes, export to PDF and add to (in my case) Set List Maker. Annoying as hell. 

[Bobadohshe]

I chart everything out and use Finale to do it. Tons of tunes in a short amount of time? No problem. And even though it seems like a little more work up front, it ALWAYS pays dividends down the road. Want to do a tune we haven't done in seven years? No problem. Pull it up on the iPad and we're good. Quick dirty leadsheets with just chords and forms? Still use Finale. It's just so useful and easy. Probably faster than doing it by hand and far neater and transpoable and rearrangable.

[Analogaddict]

I do everything in Sibelius. I can do a simple transcription faster by hand, but the ability to edit stuff after the fact makes it worth it. Copy/paste entire sections and the ability to change keys have saved my butt many times.

[Al Coda]

hand,- pencil, paper, rubber.

 

😉

 

A.C.

 

 

[Cabo]

I can do a lead sheet faster in Sibelius than by hand, plus it's quickly transposable and is much easier to read.  Adding text is very easy.

[Al Coda]

29 minutes ago, Cabo said:

I can do a lead sheet faster in Sibelius than by hand, plus it's quickly transposable and is much easier to read.  Adding text is very easy.

 

I wrote tons of notation, scores and sheets, when working as an M.D. in the past and when needed, there were copyists doing the final work.

The learning curves of these apps are unacceptable for me,- just only to become faster in notation than I am already "by hand".

Today, I write sheets only for me and not for others anymore.

No interest in learning applications for every task I can do otherwise already.

I can read my sheets, make photocopies, shrink ´em,- or scan the sheets to make a PDF or such.

 

But when it came to presentation ´cause I´d have to do that work for ensembles and would get big money for,- I possibly learned such app.

Leaves defining "big money" ...

But I´m retired from such pro work meanwhile.

 

A.C.

[Dave Number Four]

Learning the keyboard shortcuts for Musescore has paid off many times over for me. And I assume Sibellius or Finale would be similar. 

[stoken6]

21 hours ago, analogika said:

One thing that really annoys me is that I haven't found any software solution that lets me freely combine notation and text.

 

E.g. write out

a hookline,

an A Theme, and

a B Theme, all with melody and chord changes. 

 

Then free space to write as text 

"A Theme (sin rep)" 
"B Theme (organ solo)"

"A Theme (with drum breaks)"

 

Then another line of notation and chords for the C part

 

Etc. 

 

The way I work now is to take a sheet music lined blank PDF and write it out by hand on the iPad, directly in the Files app. 

 

I'd really like to be able to just MIDI-record some of the more complex stuff for speed and combine computer notation with the above free layout. 

 

Nothing seems to be able to do this, unless I arrange screen-shot notation and text with a layout program like Pages. 

Yeah I like my charts to do this, and I use Sibelius in a fairly simple way to achieve it. I'd score "hookline", "A theme", "B Theme", and then use multiple bar rests for the subsequent times they appear, with added text above/underneath as appropriate. Then score the C part, and refer back to it if needed.

 

Cheers, Mike.

[marino]

15 hours ago, Al Coda said:

hand,- pencil, paper, rubber.

 

Same here... + scanner at the end.

In all fairness, it's been a long time since I wrote any orchestral or large ensemle music... in that case, I would probably try to re-learn (or learn from scratch) some notation program.

I learned Sibelius once, but I was so used to write by hand, that it wasn't saving me any time - so I just let it go, many years ago.

[jerrythek]

I think it's obvious that those of us that grew up writing out music by hand are still faster and more comfortable doing it that way. If you don't have that skill already, it's a toss-up.

 

There are a lot of advantages to using software:

 

- Likely better legibility

- More durable (doesn't fade over time, doesn;t blow away off the stand on a gig (assuming you're using a "device")

- easy to transpose when needed

 

The issue is you need to put in the time to learn the task at hand. Those of us that grew up on pencil and paper put in that time, so some don't want to reinvest in a different way of doing a task we already know how to do.

 

When I left Korg and chose to go back to playing music, I decided to invest in my skills at using Sibelius to be able to write out charts. I put in a lot of time, all doing tasks I needed to do, and developed a decent proficiency at it. I think I might still be a touch quicker by hand, but that can't compare to the quality charts I make with the software. And as Bobadohshe mentioned - having a library of charts I've written for gigs has paid dividends when someone calls a tune, and I can do a quick search and find a chart I forgot I did for some one-off gig a few years ago.

 

The worst part has been starting to chart a tune and then when I go to save it, my computer prompts me with, "this file already exists". Doh!! Wrote one a few years ago and forgot about it!

 

Any way you do it, taking the time to write a chart means you'll listen to the tune carefully, and repeatedly, and it is a great form of practice/prep.

 

Jerry

 

 

[ElmerJFudd]

25 minutes ago, jerrythek said:

I think it's obvious that those of those that grew up writing out music by hand are still faster and more comfortable doing it that way. If you don't have that skill already, it's a toss-up.

 

There are a lot of advantages to using software:

 

- Likely better legibility

- More durable (doesn't fade over time, doesn;t blow away off the stand on a gig (assuming you're using a "device")

- easy to transpose when needed

 

The issue is you need to put in the time to learn the task at hand. Those of us that grew up on pencil and paper put in that time, so some don't want to reinvest in a different way of doing a task we already know how to do.

 

When I left Korg and chose to go back to playing music, I decided to invest in my skills at using Sibelius to be able to write out charts. I put in a lot of time, all doing tasks I needed to do, and developed a decent proficiency at it. I think I might still be a touch quicker by hand, but that can't compare to the quality charts I make with the software. And as Bobadohshe mentioned - having a library of charts I've written for gigs has paid dividends when someone calls a tune, and I can do a quick search and find a chart I forgot I did for some one-off gig a few years ago.

 

The worst part has been starting to chart a tune and then when I go to save it, my computer prompts me with, "this file already exists". Doh!! Wrote one a few years ago a forgot about it!

 

Any way you do it, taking the time to write a chart means you'll listen to the tune carefully, and repeatedly, and it is a great form of practice/prep.

 

Jerry

 

 

Correct.  Whatever you practiced is what’s fast and easy.  It took time to learn pencil and paper (even if you started at 5) and it takes time to be fluid with  software.   

[cedar]

What about Staffpad? Doesn't work for me but seems like it would for many.

 

If you're not familiar, Staffpad is software meant for tablets. (Used to be only for the Surface, but now I believe works on ipads).  Allows one to use a "pen" to write out onto a blank staff, and then converts it into a score that can be transcribed, copied, edited, with functions similar to Finale.

 

After I purchased the software, I realized that my handwriting is atrocious in all formats, so it really wasn't a solution for me.  But for those who have decent penmanship, might be a good compromise.

[bfields]

On 7/23/2022 at 5:10 PM, analogika said:

One thing that really annoys me is that I haven't found any software solution that lets me freely combine notation and text.

 

E.g. write out

a hookline,

an A Theme, and

a B Theme, all with melody and chord changes. 

 

Then free space to write as text 

"A Theme (sin rep)" 
"B Theme (organ solo)"

"A Theme (with drum breaks)"

 

Then another line of notation and chords for the C part

That's more or less exactly what I do with Musescore, by inserting horizontal and vertical frames for the text.

 

Though to be honest, I'm still figuring out how to make it work well.

 

No matter how I do it, it's a struggle to create a chart that captures the necessary detail and is also easy to scan while playing live.

[b3plyr]

2 hours ago, cedar said:

What about Staffpad? Doesn't work for me but seems like it would for many.

 

If you're not familiar, Staffpad is software meant for tablets. (Used to be only for the Surface, but now I believe works on ipads).  Allows one to use a "pen" to write out onto a blank staff, and then converts it into a score that can be transcribed, copied, edited, with functions similar to Finale.

 

After I purchased the software, I realized that my handwriting is atrocious in all formats, so it really wasn't a solution for me.  But for those who have decent penmanship, might be a good compromise.

I have Staffpad on my iPad Pro. It has been significantly improved since first released. One thing I have trouble with is simple slashes! Go figure. However, notes, etc., work very well.

[ElmerJFudd]

On the low end cost wise, has anyone played with Kawai’s Touch Notation? $11.99. 
 

https://apps.apple.com/us/app/touch-notation/id890358046

 

I’m curious if it is both simple and feature-full enough for lead sheet writing quickly.  
 

 

 

[obxa]

I'd be curious about that app, look promising. 

 

  It may be age related, but I think of most IOS apps the same: they do great light content creation while in a pinch, but I still consider Tablets better for consumption rather than creation.

You can absolutely do light word-processing, spreadsheets, email, and even simple notation.  But personally, I'd still default to using a bona-fide desktop, proper input keyboard, dual monitors, and access to my favorite java.