iPad app for IMSLP: Thoughts/Preferences

[gardano]

Hi all,

As I'm finishing up the iPad app (Padrucci), I have some questions about folks' preferences on specific topics. Here's the first.

I'm finding that the layout of some scans are inconsistent. Even numbered pages may be upside-down, while odd numbered pages are right-side up. They may have been scanned in landscape mode and are presented sideways.

In giving the user the ability to rotate pages, what would the preferable way to implement this be?
* Rotate all pages in 90° increments
* Rotate individual pages in 90° increments

-- I'm looking forward to your thoughts on this

Gardano

[nsilvers]

I would definitely say that I would prefer to be able to rotate pages individually. Perhaps I misunderstand you, but if every other page is rotated then by rotating all pages one would simply end up with the order of inverted pages swapped. So perhaps it would be best to have a really easy way of rotating a page 90º. Perhaps a dedicated button?
I am very much looking forward to trying this app — I can't wait!

[pml]

I would have thought with an iPad that you shouldn't have to actually rotate a page – the Padrucci app should be using the in-built accelerometers to work out which way is up. If you're reading a score which is mainly landscape rather than portrait, then you turn the iPad on its side. (If the long axis of the page doesn’t match the current iPad rotation’s long axis, then a sensible default would be to load it in the “wrong way” to make it clear that the user ought to rotate it for best readability – and separately, you have an application preference to allow the user to override this default behaviour, which appears pretty sensible to me!)

To be absolutely clear about what that implies, I think it would be pretty annoying if:
a) initially you're holding the iPad in portrait,
b) you open a score that is landscape and it opens displayed in portrait, so that you have to rotate for the better resolution AND redraw the page.

(Of course, if you want to zoom a page larger than what the iPad can display, then the iPad orientation naturally overrides the page orientation.)

Also worth thinking about, in terms of using an iPad to actually play works from – an option to tag pages that shouldn’t be displayed (or an easily-accessible option to display only even pages, or only odd pages), since there is an entire category of works – piano four-hands – where a player will normally read only every other page, and you probably don't want to have to have controls to "skip two pages forward/back" in addition to the usual default of ~ four buttons (skip to: first page/one page back/one page forward/last page).

Finally: think about having initial "Tool Tips" when you start the application, which could explain some of the more esoteric preferences that you might conceivably want. (As well as the option to turn them off, since some people hate having a "Tip of the Day" appear when they start an app.)

PML

[gardano]

PML, nsilvers

Good thoughts, all. I agree that 90° individually is the way to go. The downside is that if you have a piece with 100 pages, you have to rotate 100 times.

PML, unless you've locked you screen rotation to one or the other direction, relying on the orientation of the iPad won't work. If you have a score scanned in sideways, when you rotate the iPad, the page will rotate along with the ipad, so it will still be sideways, but too small to see. I suppose I could tell folks to lock their orientation for that (I'd certainly prefer to write less code rather than more code), but it would give the wrong impression. Apps should work for YOU, not the other way around, seems to me.

Zooming is something the programmer has to write in, it doesn't come "for free" in an app.

So the way the app works now is that it remembers your last manual rotation and the zooming level on each individual page.

I like very much the option of tagging a page to not display. There are plenty of scores where all the parts are in one big scan, and having to wade, page by page, through the cello and basso continuo parts to get to the flauto part every time one opens a score could be annoying.

I also like a "help mode" idea, although I'd have to do it better than ForScore, which I originally found totally baffling until I discovered I had to swipe with 2 fingers.

[pml]

Hi Angelo (or why else would you have chosen the nym of gardano, eh? ;-),

The downside is that if you have a piece with 100 pages, you have to rotate 100 times.

You mean you wouldn’t offer the user a “rotate all pages?” option after about the second time he or she went to rotate a page in the same score?

Surely the iPad would store some info about the way a user uses certain scores, such as, the default view for this score always omits page 1 or all of the odd/even pages? (e.g. piano 4-hands music)

I’m currently without an iPad, but having used several I’m aware of the physical switch to lock the rotation. (I know a friend who’s used one in performance using one of the many PDF reader applications, but she had to do some scanning and combining of files so that all of the separate concert pieces were gathered in one continuous “heft”.)

The point about the rotation lock is that it already tells you (by virtue of being on or off) something about how the user may want the application to behave, so it might be intelligent to map out all the sorts of ways a user might utilise that. Most scores I use tend to be portrait (an organist might have a majority of scores that are landscape), so if I were using Padrucci on a regular basis I’d probably have the rotation locked off in portrait mode; if I opened the occasional score in landscape, I’d prefer it opened rotated 90º one way or the other to the direction of the long axis of the screen, and I’d physically manipulate the iPad to match.

One thing that would make it far superior to a lot of PDF readers would be how it deals with page turns. With a number of PDF readers you get
1) a long delay while the next page is processed, before it displays
2) even worse, some readers blank the display in the middle of removing the previous page and drawing the next page, so if there is a pause in the rendering for some reason, you can be staring at a blank screen for an unconscionably long time; you don’t have this problem with printed paper!

Repeats. The normal method of score reading is forwards, but works like symphonies or da capo arias often involve multiple page turns backward at some point. A really useful gesture would be to have certain pages marked for “instant recall”, so that there’s a specific gesture you can use to turn back however many pages to the most recently tagged page. (You’d also use the same tagging to tag the starts of movements, or particular rehearsal spots, to allow jumping rapidly to say, movement 4 by only three or so taps.)

I know you said you wanted to code less rather than more, but some ideas are worth having implemented in the program (probably via a preferences screen):

* Performance mode. You don’t want the iPad making noises, or the battery running out, if you’re using it to perform; so a bit like “Airplane Mode” on the iPhone, activating a single touch “Performance Mode” does things like:
- turn off power-hungry networking features like WiFi, Bluetooth, 3G, etc.; obviously Padrucci would cache scores rather than pulling them from IMSLP every single time? (Obviously if you have a Bluetooth “clicker” to turn pages without you having to touch the iPad screen, then you would not disable it, again as a default preference)
- turns off the speakers. Don’t want loud mp3 playing by accident as a notification reminder!
- doesn’t have the iPad go to sleep after 1 or 2 minutes by default (as a slow point in a movement might not involve a page turn for that length of time).

* Playlists. Add specific IMSLP scores to any new playlist you create on the fly, such as “Bach concert 22/4/2011”, so that you can quickly jump between items on a list of scores without excessive navigation between pages.

If you’re programming gestures (like pinch zooming etc.) then these are worth explaining via tool tips in addition to being listed on a documentation page somewhere under the “About” menu item...

I gather you’re intending Padrucci to be a paid app?

Cheers, PML

[horndude77]

It may be worth investigating automatically orienting the page. This is a common operation for text documents. Leptonica provides some facilities for doing this (see http://www.leptonica.com/papers/skew-measurement.pdf and http://tpgit.github.com/Leptonica/flipdetect_8c_source.html). I imagine some research would be needed to make it work reliably for music. (In most cases the source scores should probably be fixed though.)

[gardano]

PML:

The rotation issue is somewhat complicated, which is why I asked for feedback on it to begin with :¬)

Yes, rotation lock can and should be important in viewing the PDFs. The issue I'm grappling with is documents that were scanned in landscape mode but display in portrait mode. I'll have to dig out an example to clarify what I mean. In the end, any iPad app should "do the right thing" out of the box, without having to force the user to perform any additional action -- so that's what I'm going for. In some cases, there's no way to know what "the right thing" is, because of the inconsistent manner in which some PDFs have been scanned.

I'm working hard on having a smooth transition from one page to the next. As the app stands now, the first time a page is viewed,there's a slight delay in viewing the page. Subsequent viewings are zippity fast though. I continue to work on making this as seamless and quick as possible. I agree that having a huge lag when viewing a page is unacceptable.

As to rotation in general, the simplest solution would be to have the option of rotating one page and another option of rotating them all.

I've been thinking hard on having some bookmarking capabilities, so that one could have the pagination honor repeats, da capos, etc. This unfortunately won't make it into the first version of the app, I believe. It's a pretty huge issue -- one which will require some dedicated development time to make work, because it would have to be integrated with some sort of automated page turning system, which I'm giving additional hard thought to. Automated page turning might have to rely on a bluetooth page turning device, or some other solution that I'm considering. As far as I can see, the only appendage that can do this turning is the foot. Unless you're an organist or other musician with an instrument that uses the feet too.

I'm implementing playlists. I haven't been using that term, but thank you -- that's exactly the right term to use!

By design, the app caches the scores, so once your scores are downloaded, you don't need to have a network connection at all. Since WiFi isn't used unless one is actively downloading a score, this power drain won't be an issue. I also plan on having a preference to disable the screen dimming/locking so that one isn't surprised in the middle of a performance! I hadn't thought of turning off notifications, but you are exactly right in that.

Turning off speakers is an issue that your suggestion of having a "Performance Mode" preference would solve. I can imagine people using Padrucci to read along with a score while the iPod app is playing it in the background, so turning off sound as a default wouldn't work for that case.

Pinching and zooming are working in the app, yes. A good way of explaining the things one can do on a given page of a score will be useful, I totally agree.

This will be a paid app, with a significant portion of the proceeds going right back to IMSLP. Buying the app will be donating to IMSLP too.

[gardano]

It may be worth investigating automatically orienting the page.

I'm orienting the page automatically as it is, but that doesn't help for pages that have been incorrectly scanned or rendered to PDF, unfortunately -- so it's not an option that can be depended on.

[pml]

I'm implementing playlists. I haven't been using that term, but thank you -- that's exactly the right term to use!

Pleasure!

Turning pages, I would have thought if the performer has a free hand (or a page turner), that they would be using a directional swipe across the iPad screen to achieve this? (Right to left = forward, left to right = backward)

Instrumentalists of all types without the services of a page turner might find it easy to have a foot-actuated clicker, but I can imagine other types of hands-free performers (singers, conductors’ non-baton hand) using hands to manipulate the iPad screen directly.

Regarding the page rendering delay: surely when the performer turns to page x, then unless there’s a repeat it would be sensible for the app to begin rendering page x+1 in the background? It’s rare in any piece of music that you spend less than one second on page x before turning to x+1 (mainly orchestral scores with just one bar to a page – I can think of several such scores).

PML

[gardano]

Regarding the page rendering delay: surely when the performer turns to page x, then unless there’s a repeat it would be sensible for the app to begin rendering page x+1 in the background?
PML

Yes, I'm rendering n-1 and n+1 in the background (don't forget, they can turn the page backwards too, and there's not enough memory in the iPad to keep all pages "live" in memory).

and yes, paging is done by swiping the screen to the left or to the right, as expected. I was more thinking of automated ways of turning the page. As a violinist, I automatically try to think of ways to achieve this sans hands.... Keep a device in one's breast pocket and hit it with the inside of the elbow (naw, that's just crazy-talk!)....

[pml]

True, but usually n–1 has already been rendered (and shouldn’t be flushed until it becomes n–2!)

I’m usually a vocalist or a conductor, and in the latter case one could imagine someone designing a USB device for the baton so that an extra bit of pressure somewhere on the pommel turns the page… hehehe

[gardano]

True, but usually n–1 has already been rendered (and shouldn’t be flushed until it becomes n–2!)

If you're on page 251, the prerendered pages are 250 and 252. If you page to 250, then without prerendering n-1, the prerendered page would be only 251. You still need to do a prerendering of n-1. (Of course, if the page is already prerendered, it's not prerendered a second time!)

It would be awesome to have a Bluetooth baton with a "next" "previous" and "macro (jump)" button indeed ;¬)

[Philidor]

Fabulous stuff. I suspect this app will be very popular... I see Apple's just posted record profits so the timing's perfect.

... shares increased by 4% in after-hours trading, and the company announced that during the Christmas quarter it sold 7.3m iPads – as many as it sold in the previous two quarters – taking the total sold to 14.5m copies since its introduction in April. The iPhone also saw strong growth, with 16.24m sold in the most recent quarter – its best ever.

[gardano]

14 1/2 million iPads alone are now out there in the wild...

[Wayne Borean]

I'll definitely use it. Both my wife and I are musicians, and have IPads.

As a matter of fact I predicted the day after Jobs announced the IPad that it would be a success just on sales to musicians. How many of us carry around 20 pounds worth of binders when we travel

One IPad to rule them all...

Wayne