Hi Folks. Here is a draft of the Speakup User's Guide I'm working on. It's missing a few things, but should answer a lot of the newbie questions that arise on the list. Gene The Speakup User's Guide By Gene Collins Last modified on Wed Mar 9 10:04:07 2005. Copyright 2005 (r) Gene Collins, and distributed with the Speakup source under Version II or later of the GPL (GNU Public License). See the file "copying" for further details. Preface The purpose of this document is to familiarize users with the user interface to Speakup, a Linux Screen Reader. If you need instructions for installing or obtaining Speakup, visit the web site at http://linux-speakup.org/. Speakup is a set of patches to the standard Linux kernel source tree. It can be built as a series of modules, or as a part of a monolithic kernel. These details are beyond the scope of this manual, but the user may need to be aware of the module capabilities, depending on how your system administrator has installed Speakup. If Speakup is built as a part of a monolithic kernel, and the user is using a hardware synthesizer, then Speakup will be able to provide speech access from the time the kernel is loaded, until the time the system is shutdown. This means that if you have obtained Linux installation media for a distribution which includes Speakup as a part of its kernel, you will be able, as a blind person, to install Linux with speech access unaided by a sighted person. Again, these details are beyond the scope of this manual, but the user should be aware of them. See the web site mentioned above for further details. 1. Starting Speakup If your system administrator has installed Speakup to work with your specific synthesizer by default, then all you need to do to use Speakup is to boot your system, and Speakup should come up talking. This assumes of course that your synthesizer is a supported hardware synthesizer, and that it is either installed in or connected to your system, and is if necessary powered on. It is possible, however, that Speakup may have been compiled into the kernel with no default synthesizer. It is even possible that your kernel has been compiled with support for some of the supported synthesizers and not others. If you find that this is the case, and your synthesizer is supported but not available, complain to the person who compiled and installed your kernel. Or better yet, go to the web site, and learn how to patch Speakup into your own kernel source, and build and install your own kernel. If your kernel has been compiled with Speakup, and has no default synthesizer set, or you would like to use a different synthesizer than the default one, then you may issue the following command at the boot prompt of your boot loader. linux speakup_synth=ltlk This command would tell Speakup to look for and use a LiteTalk or DoubleTalk LT at boot up. You may replace the ltlk synthesizer keyword with the keyword for whatever synthesizer you wish to use. The speakup_synth parameter will accept the following keywords, provided that support for the related synthesizers has been built into the kernel. acntsa -- Accent SA acntpc -- Accent PC apolo -- Apolo audptr -- Audapter bns -- Braille 'n Speak dectlk -- DecTalk Express decext -- DecTalk (Old) External dtlk -- DoubleTalk PC keypc -- Keynote Gold PC ltlk -- DoubleTalk LT or LiteTalk spkout -- Speak Out txprt -- Transport Speakup does support two other synthesizers, but because they work in conjunction with other software, they must be loaded as modules after their related software is loaded, and so are not available at boot up. These are as follows: decpc -- DecTalk PC (not available at boot up) sftsyn -- One of several software synthesizers (not available at boot up) See the sections on loading modules and software synthesizers later in this manual for further details. 2. Basic operation Once you have booted the system, and if necessary, have supplied the proper bootup parameter for your synthesizer, Speakup will begin talking as soon as the kernel is loaded. In fact, it will talk a lot! It will speak all the boot up messages that the kernel prints on the screen during the boot process. This is because Speakup is not a separate screen reader, but is actually built into the operating system. Since almost all console applications must print text on the screen using the kernel, and must get their keyboard input through the kernel, they are automatically handled properly by Speakup. There are a few exceptions, but we'll come to those later. Most of the Speakup review keys are located on the keypad at the far right of the keyboard. The numlock key should be off, in order for these to work. If you toggle the numlock on, the keypad will produce numbers, which is exactly what you want for spreadsheets and such. For the purposes of this guide, you should have the numlock turned off, which is its default state at bootup. You probably won't want to listen to all the bootup messages every time you start your system, though it's a good idea to listen to them at least once, just so you'll know what kind of information is available to you during the boot process. You can always review these messages after bootup with the command: dmesg | more In order to speed the boot process, and to silence the speaking of the bootup messages, just press the keypad enter key. This key is located in the bottom right corner of the keypad. Speakup will shut up and stay that way, until you press another key. You can check to see if the boot process has completed by pressing the 8 key on the keypad, which reads the current line. This also has the effect of starting Speakup talking again, so you can press keypad enter to silence it again if the boot process has not completed. When the boot process is complete, you will arrive at a "login" prompt. At this point, you'll need to type in your user id and password, as provided by your system administrator. You will hear Speakup speak the letters of your user id as you type it, but not the password. This is because the password is not displayed on the screen for security reasons. This has nothing to do with Speakup, it's a Linux security feature. Once you've logged in, you can run any Linux command or program which is allowed by your user id. Normal users will not be able to run programs which require root privileges. When you are running a program or command, Speakup will automatically speak new text as it arrives on the screen. You can at any time silence the speech with keypad enter, or use any of the Speakup review keys. Here are some basic Speakup review keys, and a short description of what they do. keypad 1 -- read previous character keypad 2 -- read current character (pressing keypad 2 twice rapidly will speak the current character phonetically) keypad 3 -- read next character keypad 4 -- read previous word keypad 5 -- read current word (press twice rapidly to spell the current word) keypad 6 -- read next word keypad 7 -- read previous line keypad 8 -- read current line (press twice rapidly to hear how much the text on the current line is indented) keypad 9 -- read next line kepad period -- speak current cursor position and anounce current virtual console It's also worth noting that the insert key on the keypad is mapped as the speakup key. Instead of pressing and releasing this key, as you do under DOS or Windows, you hold it like a shift key, and press other keys in combination with it. For example, repeatedly holding keypad insert, from now on called speakup, and keypad enter will toggle the speaking of new text on the screen on and off. This is not the same as just pressing keypad enter by itself, which just silences the speech until you hit another key. When you hit speakup plus keypad enter, Speakup will say, "You turned me off.", or "Hey, that's better." When Speakup is turned off, no new text on the screen will be spoken. You can still use the reading controls to review the screen however. 3. Using the Speakup Help System Speakup has a help system, which is compiled as a module. It is loaded automatically whenever the Speakup help system is invoked for the first time, and remains loaded after that, until speakup is unloaded. Note that if speakup was compiled into a monolithic kernel on your system, you will not be able to unload Speakup from your kernel. If you try to use the help system, and find that it is unavailable, then your system administrator has not installed the Speakup help module, which is called speakup_help. Complain to your system administrator about this. In order to enter the Speakup help system, press and hold the speakup key (remember that this is the keypad insert key), and press the f1 key. You will hear the message: "Press space to leave help, cursor up or down to scroll, or a letter to go to commands in list." When you press the spacebar to leave the help system, you will hear: "Leaving help." While you are in the Speakup help system, you can scroll up or down through the list of available commands using the cursor keys. The list of commands is arranged in alphabetical order. If you wish to jump to commands in a specific part of the alphabet, you may press the letter of the alphabet you wish to jump to. You can also just explore by typing keyboard keys. Pressing keys will cause Speakup to speak the command associated with that key. For example, if you press the keypad 8 key, you will hear: "Keypad 8 is line, say current." You'll notice that some commands do not have keys assigned to them. This is because they are very infrequently used commands, and are also accessible through the proc system. We'll discuss the proc system later in this manual. You'll also notice that some commands have two keys assigned to them. This is because Speakup has a built in set of alternative key bindings for laptop users. The alternate speakup key is the caps lock key. You can press and hold the caps lock key, while pressing an alternate speakup command key to activate the command. On most laptops, the numeric keypad is defined as the keys in the j k l area of the keyboard. There is usually a function key which turns this keypad function on and off, and some other key which controls the num lock state. Toggling the keypad functionality on and off can become a royal pain. So, Speakup gives you a simple way to get at an alternative set of key mappings for your laptop. These are also available by default on desktop systems, because Speakup does not know whether it is running on a desktop or laptop. So you may choose which set of Speakup keys to use. Some system administrators may have chosen to compile Speakup for a desktop system without this set of alternate key bindings, but these details are beyond the scope of this manual. To use the caps lock for its normal purpose, hold the shift key while toggling the caps lock on and off. We should note here, that holding the caps lock key and pressing the z key will toggle the alternate j k l keypad on and off. 4. Keys and Their Assigned Commands In this section, we'll go through a list of all the speakup keys and commands. You can also get a list of commands and assigned keys from the help system. The following list was taken from the speakupmap.map file. Key assignments are on the left of the equal sign, and the associated Speakup commands are on the right. The designation "spk" means to press and hold the speakup key, a.k.a. keypad insert, a.k.a. caps lock, while pressing the other specified key. spk key_f9 = punc_level_dec spk key_f10 = punc_level_inc spk key_f11 = reading_punc_dec spk key_f12 = reading_punc_inc spk key_1 = vol_dec spk key_2 = vol_inc spk key_3 = pitch_dec spk key_4 = pitch_inc spk key_5 = rate_dec spk key_6 = rate_inc key_kpasterisk = toggle_cursoring spk key_kpasterisk = speakup_goto spk key_f1 = speakup_help spk key_f2 = set_win spk key_f3 = clear_win spk key_f4 = enable_win spk key_f5 = edit_some spk key_f6 = edit_most spk key_f7 = edit_delim spk key_f8 = edit_repeat shift spk key_f9 = edit_exnum key_kp7 = say_prev_line spk key_kp7 = left_edge key_kp8 = say_line double key_kp8 = say_line_indent spk key_kp8 = say_from_top key_kp9 = say_next_line spk key_kp9 = top_edge key_kpminus = speakup_parked spk key_kpminus = say_char_num key_kp4 = say_prev_word spk key_kp4 = say_from_left key_kp5 = say_word double key_kp5 = spell_word spk key_kp5 = spell_phonetic key_kp6 = say_next_word spk key_kp6 = say_to_right key_kpplus = say_screen spk key_kpplus = say_win key_kp1 = say_prev_char spk key_kp1 = right_edge key_kp2 = say_char spk key_kp2 = say_to_bottom double key_kp2 = say_phonetic_char key_kp3 = say_next_char spk key_kp3 = bottom_edge key_kp0 = spk_key key_kpdot = say_position spk key_kpdot = say_attributes key_kpenter = speakup_quiet spk key_kpenter = speakup_off key_sysrq = speech_kill key_kpslash = speakup_cut spk key_kpslash = speakup_paste spk key_pageup = say_first_char spk key_pagedown = say_last_char key_capslock = spk_key spk key_z = spk_lock key_leftmeta = spk_key ctrl spk key_0 = speakup_goto spk key_u = say_prev_line spk key_i = say_line double spk key_i = say_line_indent spk key_o = say_next_line spk key_minus = speakup_parked shift spk key_minus = say_char_num spk key_j = say_prev_word spk key_k = say_word double spk key_k = spell_word spk key_l = say_next_word spk key_m = say_prev_char spk key_comma = say_char double spk key_comma = say_phonetic_char spk key_dot = say_next_char spk key_n = say_position ctrl spk key_m = left_edge ctrl spk key_y = top_edge ctrl spk key_dot = right_edge ctrl spk key_p = bottom_edge spk key_apostrophe = say_screen spk key_h = say_from_left spk key_y = say_from_top spk key_semicolon = say_to_right spk key_p = say_to_bottom spk key_slash = say_attributes spk key_enter = speakup_quiet ctrl spk key_enter = speakup_off spk key_9 = speakup_cut spk key_8 = speakup_paste shift spk key_m = say_first_char ctrl spk key_semicolon = say_last_char 5. The Speakup Proc System The Speakup screen reader also creates a speakup subdirectory as a part of the proc system. You can see these entries by typing the command: ls -1 /proc/speakup/* If you issue the above ls command, you will get back something like this: /proc/speakup/attrib_bleep /proc/speakup/bell_pos /proc/speakup/bleep_time /proc/speakup/bleeps /proc/speakup/caps_start /proc/speakup/caps_stop /proc/speakup/characters /proc/speakup/cursor_time /proc/speakup/delay_time /proc/speakup/delimiters /proc/speakup/ex_num /proc/speakup/freq /proc/speakup/full_time /proc/speakup/jiffy_delta /proc/speakup/key_echo /proc/speakup/keymap /proc/speakup/no_interrupt /proc/speakup/pitch /proc/speakup/punc_all /proc/speakup/punc_level /proc/speakup/punc_most /proc/speakup/punc_some /proc/speakup/punct /proc/speakup/rate /proc/speakup/reading_punc /proc/speakup/repeats /proc/speakup/say_control /proc/speakup/say_word_ctl /proc/speakup/silent /proc/speakup/spell_delay /proc/speakup/synth_direct /proc/speakup/synth_name /proc/speakup/tone /proc/speakup/trigger_time /proc/speakup/version /proc/speakup/voice /proc/speakup/vol In addition to using the Speakup hot keys to change such things as volume, pitch, and rate, you can also echo values to the appropriate entry in the /proc/speakup directory. This is very useful, since it lets you control Speakup parameters from within a script. How you would write such scripts is somewhat beyond the scope of this manual, but I will include a couple of simple examples here to give you a general idea of what such scripts can do. Suppose for example, that you wanted to control both the punctuation level and the reading punctuation level at the same time. For simplicity, we'll call them punc0, punc1, punc2, and punc3. The scripts might look something like this: #!/bin/bash # punc0: sets reading punc and punc level to 0. echo 0 >/proc/speakup/punc_level echo Punctuation level set to 0. #!/bin/bash # punc1: sets reading punc and punc level to 1. echo 1 >/proc/speakup/punc_level echo Punctuation level set to 1. #!/bin/bash # punc2: sets reading punc and punc level to 2. echo 2 >/proc/speakup/punc_level echo Punctuation level set to 2. #!/bin/bash # punc3: sets reading punc and punc level to 3. echo 3 >/proc/speakup/punc_level echo Punctuation level set to 3. If you were to store these four small scripts in a directory in your path, perhaps /usr/local/bin, and set the permissions to 755 with the chmod command, then you could change the default reading punc and punctuation levels at the same time by issuing just one command. For example, if you were to execute the punc3 command at your shell prompt, then the reading punc and punc level would both get set to 3. I should note that the above scripts were written to work with bash, but regardless of which shell you use, you should be able to do something similar. The Speakup proc system also has another interesting use. You can echo Speakup parameters into the proc system in a script during system startup, and speakup will return to your preferred parameters every time the system is rebooted. Most of the Speakup proc parameters can be manipulated by a regular user on the system. However, there are a few parameters that are dangerous enough that they should only be manipulated by the root user on your system. There are even some parameters that are read only, and cannot be written to at all. For example, the version entry in the Speakup proc system is read only. This is because there is no reason for a user to tamper with the version number which is reported by Speakup. Doing an ls -l on /proc/speakup/version will return this: -r--r--r-- 1 root root 0 Mar 21 13:46 /proc/speakup/version As you can see, the version entry in the Speakup proc system is read only, is owned by root, and belongs to the root group. Doing a cat of /proc/speakup/version will display the Speakup version number, like this: cat /proc/speakup/version Speakup v-2.00 CVS: Thu Oct 21 10:38:21 EDT 2004 synth dtlk version 1.1 The display shows the Speakup version number, along with the version number of the driver for the current synthesizer. Looking at entries in the Speakup proc system can be useful in many ways. For example, you might wish to know what level your volume is set at. You could type: cat /proc/speakup/vol 5 The number five which comes back is the level at which the synthesizer volume is set at. All the entries in the Speakup proc system are readable, some are writable by root only, and some are writable by everyone. Unless you know what you are doing, you should probably leave the ones that are writable by root only alone. Most of the names are self explanatory. Vol for controlling volume, pitch for pitch, rate for controlling speaking rate, etc. If you find one you aren't sure about, you can post a query on the Speakup list. 6. Changing Synthesizers It is possible to change to a different synthesizer while speakup is running. In other words, it is not necessary to reboot the system in order to use a different synthesizer. You can simply echo the synthesizer keyword to the /proc/speakup/synth_name proc entry. Depending on your situation, you may wish to echo none to the synth_name proc entry, to disable speech while one synthesizer is disconnected and a second one is connected in its place. Then echo the keyword for the new synthesizer into the synth_name proc entry in order to start speech with the newly connected synthesizer. See the list of synthesizer keywords in section 1 to find the keyword which matches your synth. 7. Loading modules As mentioned earlier, Speakup can either be completely compiled into the kernel, with the exception of the help module, or it can be compiled as a series of modules that can be loaded after the system is booted up. When compiled as modules, Speakup will not be able to speak the bootup messages, and the modules can only be loaded after the file systems have been checked and mounted. There is a third possibility. Speakup can be compiled with some components built into the kernel, and others as modules. As we'll see in the next section, this is particularly useful when you are working with software synthesizers. If Speakup is completely compiled as modules, then you must use the modprobe command to load Speakup. You do this by loading the module for the synthesizer driver you wish to use. The driver modules are all named speakup_<keyword>, where <keyword> is the keyword for the synthesizer you want. So, in order to load the driver for the DecTalk Express, you would type the following command: modprobe speakup_dectlk Issuing this command would load the DecTalk Express driver and all other related Speakup modules necessary to get Speakup up and running. To completely unload Speakup, again presuming that it is entirely built as modules, you would give the command: modprobe -r speakup_dectlk The above command assumes you were running a DecTalk Express. If you were using a different synth, then you would substitute its keyword in place of dectlk. But now, suppose we have a situation where the main Speakup component is built into the kernel, and some or all of the drivers are built as modules. Since the main part of Speakup is compiled into the kernel, a partial Speakup proc system has been created which we can take advantage of by simply echoing the synthesizer keyword into the /proc/speakup/synth_name proc entry. This will cause the kernel to automatically load the appropriate driver module, and start Speakup talking. To switch to another synth, just echo a new keyword to the synth_name proc entry. For example, to load the DoubleTalk LT driver, you would type: echo ltlk >/proc/speakup/synth_name You can use the modprobe -r command to unload driver modules, regardless of whether the main part of Speakup has been built into the kernel or not. 8. Using Software Synthesizers Using a software synthesizer requires that some other software be installed and running on your system. For this reason, software synthesizers are not available for use at bootup, or during a system installation process. In order to use a software synthesizer, you must have a package called Speech Dispatcher running on your system, and it must be configured to work with one of its supported software synthesizers. Two open source synthesizers you might use are Flite and Festival. You might also choose to purchase the Software DecTalk from Fonix Sales Inc. If you run a google search for Fonix, you'll find their web site. You can obtain a copy of Speech Dispatcher from free(b)soft at http://www.freebsoft.org/. Follow the installation instructions that come with Speech Dispatcher in order to install and configure Speech Dispatcher. You can check out the web site for your Linux distribution in order to get a copy of either Flite or Festival. Your Linux distribution may also have a precompiled Speech Dispatcher package. Once you've installed, configured, and tested Speech Dispatcher with your chosen software synthesizer, you still need one more piece of software in order to make things work. You need a package called speechd-up. You get it from the free(b)soft web site mentioned above. After you've compiled and installed speechd-up, you are almost ready to begin using your software synthesizer. Before you can use a software synthesizer, you must have created the /dev/softsynth device. If you have not already done so, issue the following commands as root: cd /dev mknod softsynth c 10 26 While we are at it, we might just as well create the /dev/synth device, which can be used to let user space programs send information to your synthesizer. To create /dev/synth, change to the /dev directory, and issue the following command as root: mknod synth c 10 25 Now you can begin using your software synthesizer. In order to do so, echo the sftsyn keyword to the synth_name proc entry like this: echo sftsyn >/proc/speakup/synth_name Next run the speechd_up command like this: speechd_up & Your synth should now start talking, and you should be able to adjust the pitch, rate, etc. In this section, we have assumed that your copy of Speakup was compiled with the speakup_sftsyn component either built into the kernel, or compiled as a module. 9. Using The DecTalk PC Card The DecTalk PC card is an ISA card that is inserted into one of the ISA slots in your computer. It requires that the DecTalk PC software be installed on your computer, and that the software be loaded onto the Dectalk PC card before it can be used. You can get the dec_pc.tgz file from the linux-speakup.org site. The dec_pc.tgz file is in the ~ftp/pub/linux/speakup directory. After you have downloaded the dec_pc.tgz file, untar it in your home directory, and read the Readme file in the newly created dec_pc directory. The easiest way to get the software working is to copy the entire dec_pc directory into /user/local/lib. To do this, su to root in your home directory, and issue the command: cp dec_pc /usr/local/lib You will need to copy the dtload command from the dec_pc directory to a directory in your path. Either /usr/bin or /usr/local/bin is a good choice. You can now run the dtload command in order to load the DecTalk PC software onto the card. After you have done this, echo the decpc keyword to the synth_name entry in the proc system like this: echo decpc >/proc/speakup/synth_name Your DecTalk PC should start talking, and then you can adjust the pitch, rate, volume, voice, etc. The voice entry in the Speakup proc system will accept a number from 0 through 7 for the DecTalk PC synthesizer, which will give you access to some of the DecTalk voices. 10. Using Cursor Tracking. In Speakup version 2.0 and later, cursor tracking is turned on by default. This means that when you are using an editor, Speakup will automatically speak characters as you move left and right with the cursor keys, and lines as you move up and down with the cursor keys. This is extremely useful, and makes editing files a snap. But there are times when cursor tracking can get in your way. So Speakup provides a toggle to turn cursor tracking on and off. You do this with the keypad asterisk key. Pressing this key repeatedly will cycle the cursor tracking on and off, and you will hear Speakup say, "cursoring off", and "cursoring on". Some folks like to turn cursor tracking off while they are using the lynx web browser. You definitely want to turn cursor tracking off when you are using the alsamixer application. Otherwise, you won't be able to hear your mixer settings while you are using the arrow keys. 11. Cut and Paste One of Speakup's more useful features is the ability to cut and paste text on the screen. This means that you can capture information from a program, and paste that captured text into a different place in the program, or into an entirely different program, which may even be running on a different console. For example, in this manual, we have made references to several web sites. It would be nice if you could cut and paste these urls into your web browser. Speakup does this quite nicely. Suppose you wanted to past the following url into your browser: http://linux-speakup.org/ Use the speakup review keys to position the reading cursor on the first character of the above url. When the reading cursor is in position, press the keypad slash key once. Speakup will say, "mark". Next, position the reading cursor on the rightmost character of the above url. Press the keypad slash key once again to actually cut the text from the screen. Speakup will say, "cut". Although we call this cutting, Speakup does not actually delete the cut text from the screen. It makes a copy of the text in a special buffer for later pasting. Now that you have the url cut from the screen, you can paste it into your browser, or even paste the url on a command line as an argument to your browser. Suppose you want to start lynx and go to the Speakup site. You can switch to a different console with the alt left and right arrows, or you can switch to a specific console by typing alt and a function key. These are not Speakup commands, just standard Linux console capabilities. Once you've changed to an appropriate console, and are at a shell prompt, type the word lynx, followed by a space. Now press and hold the speakup key, while you type the keypad slash character. The url will be pasted onto the command line, just as though you had typed it in. Press the enter key to execute the command. The paste buffer will continue to hold the cut information, until a new mark and cut operation is carried out. This means you can paste the cut information as many times as you like before doing another cut operation. You are not limited to cutting and pasting only one line on the screen. You can also cut and paste rectangular regions of the screen. Just position the reading cursor at the top left corner of the text to be cut, mark it with the keypad slash key, then position the reading cursor at the bottom right corner of the region to be cut, and cut it with the keypad slash key.
