Skip to content

Updated Jaws Scripts for Motif Editors

Some of you may know that Arturas Lenksas maintains a set of Jaws scripts for several of the Motif editor programs written by John Melas. Since I routinely receive requests for info about the editors, this post will contain some basic info to get you started with them.

The editors are programs that run on a Windows PC. Together with scripts, they provide an accessible way to control many functions of Yamaha Motif keyboards and modules. The editors do not make the Motifs “talk”. The editors are not a screen reader for the Motifs.

To use the editors, you must have a supported model of Motif and a supported version of Jaws.

The following Motif models are supported:

  • Motif XF: all models
  • Motif XS: both keyboards and the XS Rack
  • S series: S70 XS, S90 XS, and S90 ES
  • Motif ES: both keyboards and the ES Rack
  • MO: MO6 and MO8

The editor and scripts combination is not compatible with any models of the Motif classic or the MoX.

The scripts are designed to work with Jaws 10 or later.

The editorsmake the following activities possible:

  • Load new samples and voices in to the Motif.
  • Organize your existing voices, performances, master setups, arps, waveforms, and more.
  • Customize existing instrument voices (instruments and drum kits), or create new voices from scratch.
  • Customize existing performances, or create new performances (splits, layers, backing arps).
  • Edit all of the mixing settings of Songs and Patterns.
  • Create Master Setups, giving you quick access to your voices, performances, songs, etc, when in a live situation.

The editors and the scripts are both commercial products. Both can be run as demos, allowing you to try them before you buy.

To get started, download the editors from here.

Then, download the most recent version of the scripts here.

If you’d like to discuss the editors and the scripts, please subscribe to the MoAccess email list.

Site move / New Hosting Provider

Over the last few days, I’ve been moving bryansmart.com to a new hosting provider. The move should now be complete. The main site moved immediately, but it took a little longer to move the files that were hosted on the domain. For those of you that messaged me about broken links, please try again. If anything is still broken, let me know.

VoiceOver Punctuation Bug Passes Embedded Speech Commands

I made an interesting discovery today while investigating a strange punctuation issue with VoiceOver. Normally, I use VO either in no punctuation (for general work), or all (for coding, so these problems have escaped my attention until now). After trying out the text verbosity settings, the bugs are quite something to experience. The results revealed some large problems with VoiceOver’s punctuation processing/filtering, as well as a way to write embedded speech commands so that VoiceOver will pass them along to the Mac Speech Manager.

Before going further with the following examples, open VO Utility, and go to Verbosity, Text. Start with all punctuation, and set the repeated punctuation to always spoken.

Read these lines:

********** Oh my God, it’s full of stars!
(((4*3)*2)/3)
———- forwarded messages do this ———-
[[[[[[[[[[ whole lot of brackets ]]]]]]]]]]]

VO reads all of that punctuation, as it should. Now, set repeated punctuation to first two times and read those lines again.

notice that only the first line with the asterisks/stars are repeat-filtered? It seems that VO has a very limited list of punctuation that is eligible for filtering. Observe…

!!!!!!!!!!
&&&&&&&&&&
((((((((((
))))))))))
———-
[[[[[[[[[[
]]]]]]]]]]
“”””””””””
<<<<<<<<<<

??????????

Vo filters these, though:

@@@@@@@@@@
##########
$$$$$$$$$$
%%%%%%%%%%
^^^^^^^^^^
__________
==========
++++++++++
{{{{{{{{{{
}}}}}}}}}}
\\\\\\\\\\
||||||||||
;;;;;;;;;;
::::::::::
''''''''''
,,,,,,,,,,
……….
//////////

I guess the question is, do they leave the first group out for a reason, or have they been overlooked somehow?

Now, go back to your VO Utility window, set the repeat-filter back to always spoken, and change punctuation to none. You should now hear no punctuation, except, in the following, you will.

var1 & var2 &&&&&&&&&&
"""hello world"""
accessibility@apple.com @@@@@@@@@@@@@
#1 #2 #######
$1,000,000,000 is a lot of $$$$$$$$$$
500% is a lot of %%%%%%%%%%
^^^^
_lame bugs_ __________
c:\wow\this\is\bad \\\\\\\\\\

That's right. No punctuation is really some weak form of some punctuation, after all, or else this is a huge bug.

Now, set the repeat-filter to first two times, which should, at least, get rid of the cases where lots of punctuation is in a row. It reduces many cases, such as the following:

accessibility@apple.com @@@@@@@@@@@@@
#1 #2 #######
$1,000,000,000 is a lot of $$$$$$$$$$
500% is a lot of %%%%%%%%%%
^^^^
_lame bugs_ __________
c:\wow\this\is\bad \\\\\\\\\\

But still lets these totally slide:

var1 & var2 &&&&&&&&&&
"""hello world""" """"""""""

So, now that it is obvious that the whole punctuation processing code is not quite working as we expect, back to the bracket issue.

Set the punctuation to none, and set the repeat-filter to always spoken.

For the brackets problem, the key is three left brackets in a row.

This is no problem:

[[test

However, this is:

[[[ test

If, however, you set punctuation to either most or all, then three brackets are no problem.

[[[ test

Now, please go back to no punctuation before continuing.

What is, in fact, happening, is that, when set to no punctuation, VoiceOver is not filtering out left brackets. It filters them out when they're isolated, but, when you have two or more in a row, then the number, less one, are passed, as in the following:

[[[rate 175]]][[[pbas 0]]][[[pmod 0]]] please do not attempt to adjust your speakers.
[[[rate 300]]][[[pbas 60]]][[[pmod 200]]] Nothing is seriously wrong!
[[[rate 130]]][[[pbas 40]]][[[pmod 200]]] reeeeeeeeeeeeeeeeeeeeeeeeeealy!

I exploit this above to force VO to pass along embedded Speech Manager commands.

An actual command looks like:

[[rate 150]] Hello, world.

When I use three brackets to open and close, instead of two, then VO passes two along to the synth.

[[[rate 150]]] Hello, world.

The above doesn’t work in all punctuation mode, probably because VO uses a substitution dictionary to replace punctuation names.

[[

is replaced with

left bracket left bracket

Anyway, I anticipate that this will soon become a popular prank on instant messaging and e-mail lists. Not sure how many problems can be created by pranksters redefining embedded speech command delimiters, or by forcing the VO synchronization callback to run with odd values. Anyway, besides the potential pranks, it seems to me that the entire punctuation filtering approach should be examined.

VoxKeys release 1.3 now available, updates and bug fixes

Release 1.3 of VoxKeys, a collection of tools for speeding up the way that you operate your Mac with VoiceOver, has been released. You can download it from the VoxKeys project page.

The new release includes hotkeys that can recover speech if VoiceOver, or the operating system, stop responding. Bug fixes for the GarageBand support, the power status script, and others are included. VoxKeys now uses the Mac’s system voice, instead of speaking through VoiceOver.

Please see the help file for a full list of changes.

Accessibility of Motif Karma and Native Instruments Products

hennie de Wet wrote to ask about the accessibility of the new Karma software for the Yamaha Motif.

For those that haven’t heard about Karma, it is software that creates music in real-time that is influenced by the performance of a single player. It is something like an arpeggiator, and something like an automatic backing band. Karma-Labs recently announced a version of the Karma software for Windows that is compatible with the Yamaha Motif XS and XF keyboard workstations. There aren’t many good video demos yet, but here is one recorded at Winter NAMM. This is a significant development, as, previously, Karma has largely been available as a built-in technology for Korg workstations.

Of course, since you perform with Karma from the Motif, all of the performance controls will be accessible. However, access to the software through a screen reader would be required to create your own Karma performances. As far as I know, no one with a screen reader has tried the software yet. In fact, the software isn’t generally available yet to try. I’ll end up buying it, as I do with most everything else Motif-related, and will post about my experiences.

Karma-Labs is an extremely small software company. If small changes are needed to improve screen reader compatibility, they should be able to help without needing to spend months on a development study. However, if the changes are large, then they may not have the resources to make significant changes in order to support a few screen reader users.

Anyway, it is all guess work until someone actually is able to try the software with a screen reader. Hopefully, that should happen in the next month or so.

Hennie also asked about the support for custom user interfaces in Native Instruments products, such as Kontakt, and wondered if this capability could be used to improve screen reader compatibility.

Unfortunately, the answer is no. The custom interface/skinning support will allow you to replace some of the graphics used in the products, as well as perform some other small appearance tweaks, but those tweaks will not change the fact that all visual information that is produce by the plug ins is placed on the screen in ways that prevents a screen reader from being able to recognize it.

Native Instruments produces many different plug ins, and they must work in many recording environments. On Windows, each product must work with DAWs that support VST format for Cubase and other popular DAWs, RTAS format for Pro Tools, and even DXI format, in some cases, for Sonar. On the Mac, they must support VST format, AU format for DAWs like Logic and GarageBand, and RTAS for the Mac version of Pro Tools. Beyond that, some plug ins also support Rewire. That means that, for each product, they must produce over half a dozen versions that run on two entirely different operating systems.

One way that they are able to produce this many versions without requiring massive amounts of work is that they rarely use any of the built-in capabilities of an operating system for displaying a user interface. Instead, they directly draw their own interface to the screen. Each version only needs to know how to display graphics in the plug in window of the host that will use that version. In the future, when Native Instruments updates a product, and needs to change its interface, they can change it once, without having to change it for each version. Working this way saves them lots of time and expense, but it means that any software, such as a screen reader, that needs their software to work through proper operating system approaches for displaying a user interface, will only see their windows as a blank area of the screen.

Without changing the way that their user interfaces are fundamentally displayed, a Windows screen reader will never be able to read anything that is placed on the screen by a Native Instruments product. Many high-profile blind musicians, recording professionals, and even organizations have contacted Native Instruments over the years with offers of technical assistance and money in the hopes that some sort of technical solution could be created that would make it possible for a screen reader to operate their products. Most all of the proposed solutions involve changes to how their interface is displayed. This isn’t an acceptable situation for them, as it would disrupt their ability to edit an interface once, and have it display correctly in each version of the product that is created for a different plug in format.

On the Mac, screen reader compatibility is a completely separate technical matter from how information is displayed on the screen. Even software with custom graphical components can still work with a screen reader without any compromise to its appearance. Even programs that don’t use the Mac’s user interface facilities at all can still expose a hidden interface that is only visible to a screen reader. As a great example, the hidden interface approach was recently used by Avid to make most all of the Pro Tools interface accessible without requiring any changes to the appearance of the software.

However, native Instruments has long-sense stopped responding to any queries regarding the accessibility of their products. Perhaps they still believe that no technical solutions to the problem are possible, that any such solutions would cost more to create than they could earn back in sales, or are just tired of talking about the issue. Regardless, they are now completely unresponsive, and have been this way for years.

I know that Kontakt is popular. There are lots of libraries available for it, and many 3rd-party synths use it as an engine. However, my advice to anyone that is desperately searching for access to a Native Instruments product is to forget it, and to spend your money with their competitors. If you use Sonar, then there are thousands of additional voices for Dimension Pro that are available through expansion packs. Spectrasonics Omnisphere is an incredible and massive set of instrument voices, and is accessible through HotSpotClicker. If you’re recording on a Mac and using Pro Tools, then Avid’s Pro Tools Instrument Expansion Pack adds a large set of sounds through accessible instrument plug ins. For GarageBand users, there are thousands of high quality presets and gigs of samples available through Jam Packs. These are just some of the options. There isn’t a need to become stuck on a single vendor’s products. No one touched millions of people, won a Grammy, or made a fortune due to a specific preset.

Sorry to be harsh, but it seems like I read one query or another on the lists about the accessibility of Native Instruments products every few days. The posts are: what if this hack were tried, or if we used HotSpotClicker in this way, or if this person or organization got involved? It won’t work due to fundamental technical limitations, people with influence and money have attempted to assist in the past with no success, and Native Instruments doesn’t desire to take any action at any foreseeable point in the future. I wish that everyone would let this Native Instruments quest die. They don’t want our business, and there are lots of alternatives, so why beg for the privilege of supporting their work?

New Motif editor scripts available

Arturas Lenksas has released an update to his Jaws scripts for John Melas’s editors for the Yamaha Motif XS and XF keyboard workstations. The Motif is the most accessible modern keyboard workstation. The editors, when used with the scripts, make it possible for a blind person to edit instrument and drum kit voices, performances, master setups, mixing setups in the sequencers, many system settings, as well as organize data on the Motif with John’s librarian tool, and load your own samples with his Waveform Editor. Arturas’s scripts cost 30 Euros, and work with all of John Melas’s editors for the Motif XS and XF. Both the editors and the scripts will run as a demo, allowing you to try them before purchasing.

Some enhancements in this new release include the ability to quickly jump between more sections of controls than before, improved speech feedback in the effect editing windows, more values can be set based on live MIDI input, and speech feedback in the Waveform editor has been improved.

Download John Melas’s editors for the Motif
Download Arturas Lenksas’s Motif editor scripts

VoxKeys release 1.2 now available, adds GarageBand support

Release 1.2 of VoxKeys, a collection of tools for speeding up the way that you operate your Mac with VoiceOver, has been released. You can download it from the VoxKeys project page.

Top of the list of what’s new in this most recent release is a set of shortcuts that I’ve been working on for speeding up the operation of GarageBand. With the new GarageBand support, you can quickly jump to commonly used areas of the song window with a single key press, hear lots of information about the song and current track without having to move the VoiceOver cursor, directly adjust the current track’s volume and pan from the keyboard, and more. Please see the help file for all of the details.

Some problems have also been resolved with the Netflix support. VoxKeys detects the Netflix player based on its URL, and Netflix changed that URL right after 1.1 was released. Netflix support should work fine now.

Some people reported installation problems on the e-mail list that we weren’t able to fix. The installer has been entirely redesigned, so, if you had problems with 1.1, give 1.2 a shot.