Accessibility for sight impaired users

Using Zynthian
61

Organnery’s aeolus version includes sending out midi note messages when a stop is changed, so this message could be used to trigger a voice saying “oboe stop out” or “principal 8 on”.
I’m just wondering how long it could be when switching from one preset to another where more than 10 stops change …

the code can be checked out here : Organnery/aeolus: fork of aeolus engine with additions made for organnery - Audiotronic-git
As it not an official version, I’m not sure you want to pull it into zynthian, but one could if he wants. There are also more midi controls available as details in the midi chart MidiChart - Audiotronic-git

1 Like
62

There’s two solutions to this.

If using a touch screen,
You wouldn’t need midi messages to announce if using a touch screen, if the screen reader service is running and supports a touch layer, as long as the pre-requisites are there to identify a button or function, a single touch would indentify the stop / piston / coupler and it’s state (on / off)
This would mean that you could, providing you know the layout of the touch screen after getting used to it, you could navigate by positional memory and engage / disengage stops.

For a hardware solution, such as an assembled console,
if the screen reader is running already and observes changes in button state, because you can effectively treat a screen reader as recognition of keyboard or other input tools, you could therefore with mapping, either by midi or by other resources, have the screen reader announce a “per button state, this could be a minor challenge for a console where the stops are not motorised, just touch tabs or pull (spring retract) stops with LED’s. then of course, it would be midi stop.

You could configure the screen reader to announce in different ways providing you set out particular references to each button / control. Say as a good example, rather than announcing each stop name if you’re already knowledgeable of the stop list, it could announce “Man 1 - 1, 2, 5, 6, 8 on” which would be a good example of a pro-use screen reader output,

if the voice selection to the screen reader was set at a speech rate of 55% as an example, it is fast enough to be understood for someone like me and in doing so, take off the strain a touch.

lew

63

that would be the ideal solution. Aeolus window is quite minimal in design (squares for stops and controls, and basic font type, no decoration) so mapping with a specific organ definition should be a simple task.

for a hardware solution, if using Rocker tabs you would instantly feel the position of stops while passing over them with your fingers. It would not apply to all organs though.

If using midi output message as a source, doing so would need some “intelligence” to interpret the incoming messages. While not impossible at all, this would require a lot of work to study and cover all possible cases and precise timings.

yes I had clients show me how they use this on an iphone, and it was really impressive to hear the speech rate while they are scrolling through apps and menus !

1 Like
64

From the zynthian perspective a machine announcing it’s change over the MIDI (or OSC,) network would allow mimic displays and access from remote devices. The simplicity of the GUI and the standardization of 4 encoder, 4 switch interface transports well, and allows the construction of powerful structures that can exceed the limits of sound production from one Raspberry Pi.

The separation of devices that respond and devices acknowledge is what we possibly discuss here.

Acknowledging, locally , that a response has been received passes on a responsibility to only a local device.

1 Like
65

It might not be hard to build something that actually feels a lot like a traditional church organ console.

ALPS make motorised faders that include a touch sensor. Basically this is a linear potentiometer with a motor and with a separate terminal that connects to the sliding lever. They come in two lengths, with a travel of 60mm or 100mm.

Product list and specifications are on the ALPS site here:
https://tech.alpsalpine.com/prod/e/html/potentiometer/slidepotentiometers/rsn1m/rsn1m_list.html

It looks like these parts can be purchased for about €20 each from suppliers such as Mouser Electronics.

Normally in an audio mixer the control surface is horizontal, and you would slide the lever up and down to adjust the control. However, you could make a vertical control surface similar to an organ console. The faders are still mounted horizontally, but you could connect a conductive rod to the lever which runs parallel to the fader and emerges in the vertical front surface with a knob attached. These “pistons” would then move in and out to adjust the control. Using capacitive sensing you could tell which piston is being touched, which could trigger the announcement of the control name. When changing “patches” on the organ, the motors would move the pistons to the relevant positions, so that you can read the state of the patch by simply feeling the location of the pistons.

I imagine that you could mount a bank of say 10 or more of these piston controls in a thin flat box that may be only a few centimeters high. You could then stack several of these boxes vertically to get however many pistons you need for your console. I think this would give a much higher density of controls than you can get with horizontally mounted faders. Of course this is an overkill for the situation where each control is binary (on/off). I wonder if Aeolus can support a separate volume level for each stop.

1 Like
66

I like your idea, but here’s the problem, you’re confusing the linear faders to that of a hammond organ, not that of a church organ console. they do not use faders at any point because the stop action is purely on a midi scale 0 or 1 mode, off / on, no variation, therefore, a console with 80 stops as an example, would not only have to take on 80 of these faders as an example to engage / disengage stops, but also a suitable power rail, circuit boards, amongst other parts. I’m not an electrical engineer and am fully blind, so I can’t solder for toffee :slight_smile:

Tyes, having faders like this running as banks is a good idea in a tactile practice, as when off, the faders would be at 0 state (lowest level), when on, would be at top state (1), but on a linear scale, the instruction change has to convert 0 - 1 as values 0 - 127 and include ramp up / down motor speed so as not to “snap” the fader from 0 to 1 (0-127)

You’ve now put this thought in my head, you naughty thing you :slight_smile: lol hmmmm. if someone could utilise these, build a stop rail or 2 x stop panels using these to particular elements, that would actually be something rather unique and quite sexy.

are these touch sensitive, as in, do the fader heads have capacitive response? I ask as, if a surface were made using these, but to have a rather unique design to them, wooden ball knobs could be fitted to make the console a bit more desirable / traditional, rather than modern.

hmmmmm. food for thought.

any electronics guys out here up for a potential challenge?

lew

67

I think flying faders (which do have touch sense to allow disable of automation when manually overriding) is overkill. They are expensive and prone to failure.

If you want a purely manually configuration without ability to recall presets then a bank of rocker switches or push / pull switches world be easy to implement and relatively cheap. If you want automation then some form of solenoid could work.

68

What about finding a second hand transistor organ that supports midi out? That’ll have the switches for stops and is likely to have pedals and presets under the manuals. Like this Technics SX-GX7 £24.50

If you’ve got big bags of money then Johannus do some impressive stuff.

69

Good point.

I spent a lot of time going through this issue in terms of hardware based recall.

here’s where things are.

Standard rocker tabs do not have solenoids nor is there room for them to be added. There are tab stop systems, but they are per tab and about £100 each and require 2 controller boards and a rail for power.
Traditional draw stops which are electro-mechanical / electro-magnetic, are also expensive, require intensive installation, power rail and then the stop faces which are extra cost.

using momentary switches, is somewhat viable, if an LED is fitted to a side outside of each button, that then requires further control and custom wiring work. also the issue of suitable momentary switches which aren’t poor quality plastic, that’s just for the stop jambs, never mind piston rails. They would self cancel, the LED indication only assists a light scanner or if the bulb emits a temperature, like that of christmas tree lights, then the reaction to skin temperature translation is something you have to adapt to.

taking on an old transistor organ with stops isn’t easily viable, because certain stop rails are designed in particular ways, aka, boards with the controls soldered, they don’t transmit midi, they are purely CV based. I spent a long time looking in to this.

In all truth, a lottery win would help, or a grant to purchase the console that actually fits my needs which is worth over £50,000

until then, the best way is either having this custom panel I’m work on with a developer, a touch screen that would need screen reader or stop announce functionality, or god only knows what.

lew

70

that console would never do. As I’ve said before, I am a church organist, huge difference between a pipe organ and something like this. Thanks for the info on it, but it’s 2 x 4 octave keyboards, short pedalboard, the controls to that are midi spec but you can’t run it in particular ways, my build is a 3 manual plus so it doesn’t cover the needs.

yes, johannus is my main choice for a console if a lottery win or grant ever happened, either the Live 3 fully loaded or the ecclesia D570 which I’ve had the honour of working with.

lew

1 Like
71

This is what you get when you ask someone who’s main organ is a Bontempi. :grin:

2 Likes
72

There must be an sf2 of a bontempi organ out there somewhere …?

1 Like
73

But how would you emulate it’s buttons and spongy keyboard?

74

oh no. I’ve started a bontempi rant. not that plastic pile of doodah! cast it to the very gates of hell :wink:

that said, it did have an “interesting sound” came across one years ago with it’s figure 8 leg stand in a charity shop. stayed away from it.

75

Perhaps worth a read, the story of a visually impaired musician and RME’s Babyface Pro user, who wrote a reader compatible app that links to the RME via OSC interface protocol.

AZS OSC Helps Visually Impaired, Blind Users Access RME’s Audio Interfaces Jun 2020

The article links to the programmer’s Turkish blog page with a summary of his interface.
The same text is also on the forum for his “ASLOW” application.
AZ Accessible OSC (AOSC)

Summary of the the azslow introduction page:
"AZ Accessible OSC (AOSC) is an Open Sound Control. The interface is implemented using standard widgets and all functions are optimized for computer keyboard only control, making it accessible in terms of MSAA. "
(Unfortunately AOSC client is for Windows operation system, perhaps use a win emulator on the Mac.)

There are 2 Accessibility threads at the bottom of the azslow forum main page.
https://www.azslow.com/

76

Another feedback device possibility. In a live performance situation where voice feedback might not be as ideal.

The “Peterson BodyBeat Pulse Solo Metronome”, a vibrating device worn on the belt, or elsewhere, might be used for simple beep code status indications. By appearance it is probably much like a cell phone vibrator.

The transducer is triggered by audio or click output, it sells for $29.99 at Sweetwater

This same interface is used in their larger “Bodybeat” metronome, tuner and radio link that allows artists to share the beat up to 100 yards away. (another way to silently count down a song)
the control box meant to be worn on your belt, has a couple of tactile rubber buttons for start/stop and tempo tap. $132 at Amazon.

Body Beat Sync - The World’s Most Versatile Metronome

77

As @riban says, if you want a purely manual configuration then its cheap and easy to build a bank of rocker switches.

I guess the question is how important would it be to be able to save and recall the mechanical state of the switches? Apart from mixing consoles with motor-faders, I haven’t seen any other systems that actually do this. But I imagine it could be very useful if you are relying on tactile feedback to imagine the state of the console.

Circuit breakers often have the ability to switch themselves off, but I don’t know of a binary switch that can be electrically set in both directions. I guess one could design some bespoke system with solenoids, but it might not be cheaper or more reliable than using motor-faders.

Switch-bot make some gadgets for remotely controlling mechanical switches, but by the look of it I would say it contains a motor and some gears. I think this would be necessary to get sufficient force to throw the switch.

I am wondering if it might be possible to build a control surface using latching solenoids. These solenoids have an arrangement of magnets that makes them bistable. They will stay in one of two states, and you can flip then with a momentary current in one direction or the other. Depending on the holding force of the magnets, you could probably also flip them manually by pushing or pulling on the piston. Then you just need a basic sensor (a microswitch or opto-sensor) to read back the state. I have never seen solenoids used in this way, but don’t see why it wouldn’t work. Only disadvantage is that the throw of the piston is probably only about 10mm, so might need some lever arrangement to amplify the movement.

Another idea is to use a touch surface with an overlay. For example, imagine a 6 by 6 grid of virtual switches laid out on a touch surface. Now make an overlay (eg. out of cardboard or similar) which has holes corresponding to the switch locations and a suitable margin in between (say 10-15mm). You can now feel the location of the virtual switch without having to scan the otherwise featureless surface. You could even emboss names for each switch on the overlay in braille. The system could give haptic feedback about the switch state using a vibration motor. For example, continuous vibration for “on”, short pulses or no vibration for “off”. This is probably the cheapest option per switch. Touch screen digitisers are only a small proportion of the overall cost of a touch screen, and the display screen itself isn’t required in this scenario. Of course this requires some software programming, but I don’t think it would be particularly difficult.

78

There are operational reasons why this sort of thing isn’t done to do with remote control of motors and large equipment which requires interlocks. X-10 control had some semblance of this. In essence front panel controls shouldn’t be remotely controllable and physically alter from remote control. IT applies in mains equipment but the control mechanism would preclude general equipment of this sort without it being specifically built for a specialist ( which this is) application.

79

A dedicated hardware panel could be built with momentarily push buttons which toggle state with capacitive touch detection to identify when each is touched and an audible indication of the current state as it is touched which could be fed to headphones and / or via a built-in speaker. The switches could be low profile to allow swiping fingers across quickly. A voice could announce the identity of each switch if desired, maybe on linger, e.g. state indication when touched and name if long touched. This is quite plausible and could be relatively inexpensive.

80

One practical compromise, that might be readily applied to standard matrix switch controller (using momentary switches) is making a long press describe the current buttons state, with fast clicks actually applying the switch action.

I just ran across this this Haptic feedback device that a second hand could monitor, as an alternative to voice descriptions:
Pimoroni DRV2605L Linear Actuator Haptic Breakout £12.90 ($14 Digikey)
The Pimoroni DRV2605L Linear Actuator Haptic Breakout £12.90 ($14 Digikey) uses a Texas Instruments chip that contains software and waveforms licensed from immersion (apparently a leader in the haptic feedback field) These linear transducers should be able to impart a greater variety of signals than a monotonous eccentric rotating mass (ERM) motors used in cell phones.
(123 effect waveform description table)