I created this guide by taking a couple of guides I found online as a basis and adding knowledge I acquired through first hand experience while modifying the voices on my iDevice. I have personally repeatedly performed all the steps described below and extensively tested the outcome obtaining 100% positive results. Nevertheless, you do this at your own risk and I can not be held responsible if anything goes wrong or if performing these steps results in any Apple or Vocalizer license infringements. I used a PC running Windows 7 64bit, but the process should be exactly the same on any version of Windows. I suggest you read through the entire guide before beginning, make sure you fully understand each step and ask questions if you don't.
OK, here we go! 1: Open Cydia then search for and install OpenSSH. 2: On your device, make sure wifi is turned on and navigate to settings > wifi > network name more info (Where network name is the name of the network your connected to) and make a note of your ip address. 3: Using a program such as putty on Windows or the terminal on OSx, connect to your iPhones newly installed SSH server using the ip address we've just found. 4: Login with the following credentials: User: root Password: alpine 5: Type "passwd" without the quotes and change your password by following the onscreen instructions. This step is not essential but it is important for the security of your device. 6: You now need a method of transfering files to/from your device; Putty comes with a console app called psftp that will do this, however, WinSCP is much more friendly, so I suggest you use it instead. Unfortunately I'm unsure of what options you have if you're running OSx. You will need to authenticate again when logging in with WinSCP; the user will still be root however the password will now be the new password that you created in the step above. 7: Download the voice that you want from http://www.vocalizer-nvda.com/en/downloads.htm. Change the file extension of the downloaded file from nvda-addon to zip so you can access what is inside. 8: On WinSCP, navigate to /System/Library/PrivateFrameworks/VoiceServices.framework/TTSResources/ on your device. In there you should see subfolders with language and region codes, such as en-US (United States English), en-GB (British English) or es-MX (Mexican Spanish). The voice data inside any of these folders can be replaced with any voice you want. For the sake of this guide, we'll be replacing Daniel (en-GB) with Tom, but other combinations will work as well). Enter one of these folders, I.E., en-GB. You will see a file called broker.hdr, a few dat files, and a text file called user_rules.txt. This rules file is custom written by Apple, and it can be used as a dictionary (ever wondered why your iDevice pronounces artists like Rihanna correctly? Well, there you go). You can either leave it as is, or replace it with the file from the language you're installing, I.E., since I'm replacing Daniel with Tom, I could use the rules file from the en-US folder instead of the one inside en-GB. If the original voice and the one you are installing both speak the same language, I suggest you use the rules file that is larger. Now, go ahead and make a backup of this folder in case you mess something up - E.G. copy paste all the files inside the folder to a folder on your computer. 9: We'll now prepare the voice data. Lets look at the NVDA vocalizer file. When you open the adon's zip file/folder you downloaded, you will see a locale folder, which contains strings for NVDA, and a folder with the language code (for Tom it's enu), and you want to go there. Inside there will be only a speech folder so enter that as well. Now you'll see 2 folders, components, and vautov5. The components folder will have a bunch of dat files. You will definitely need these 3 (languagecode and voicename will change depending on what voice you're working with): clc_languagecode_cfg3.dat clc_languagecode_voicename_cfg3.dat select_voicename_bet2f22.dat The 4 files below these contain the voice data for each of the standard compressed, standard uncompressed, premium compressed and premium uncompressed variants: synth_voicename_dri40_155mrf22_270_06.dat, is the standard variant, which has a bunch of compression on it. synth_voicename_dri40_vssq5_f22.dat is the standard uncompressed version, which sounds very similar to the premium variant. next is synth_voicename_full_155mrf22_270_06.dat, this is the premium variant, again, the one which has compression on it. This is the one iOS downloads for Siri. And, not surprisingly, below that, is synth_voicename_full_vssq5_f22.dat, the premium uncompressed version, the largest one and the one with the best sound quality, just choose the one you want. Then we have the lexicon files: uselect_voicename_dri40.dat should be used with either of the standard variants, and uselect_voicename_full.dat is, not surprisingly, for the premium variants. Take the one appropriate to the variant you picked. You may have noticed we have everything... except broker.hdr. Well, you will need to create it using the files in the vautov5 folder. First, open vauto_pipeline_languagecode_voicename_22_bet2.hdr using a text editor such as WordPad. At the very top, type defaultvoice:voicename (with lowercase letters, no spaces and substituting "voicename" with the name of the voice you are working with) and press enter to push the original content of the file to the second line. Next, you will notice that the names of the other 4 files in the vautov5 folder correspond to the 4 standard and premium voice variants discussed above. Open with WordPad the one appropriate to the variant you picked and select and copy all of the text inside. Now, return to vauto_pipeline_languagecode_voicename_22_bet2.hdr, go to the very bottom of the text and paste the copied material from the other file there. For the sake of neatness, after pasting check that only one empty line appears on the very bottom, not 2. Finally, save this new text file, entering broker.hdr as the file name and selecting the plain text format, and that's it! You now have all the files you need for your new voice; for the premium, uncompressed version of Tom they would look like this: broker.hdr clc_enu_cfg3.dat clc_enu_tom_cfg3.dat select_tom_bet2f22.dat synth_tom_full_vssq5_f22.dat uselect_tom_full.dat user_rules.txt 10: It's now time to upload the new voice files to your iDevice. If you only plan to add a couple of standard voices or one premium voice then you can simply delete the files inside the folder of the voice you are replacing, I.E., en-GB, and paste the new voice files in there instead. If, however, you want to add a bunch of premium voices to your device, things get a little more complicated... The compact voices that come preloaded with iOS are stored inside the System partition of your device, which has a capacity of only 1.7 GB. For your device to run properly, you don't want this partition to run out of space, therefore, installing several premium voices requires that we store them somewhere else. To do this, just follow these steps: On WinSCP, navigate to /var/. Once you're in there, create a new folder by right clicking anywhere inside the var window, selecting the option "New" and then the option "Folder"; lets name the new folder CustomVoices, shall we? Now go into CustomVoices and create yet another new folder, this time naming it with the same language and region code as the voice you are replacing, I.E., en-GB. I bet you guessed what the next step will be, paste the files of your new voice into this folder. Next, navigate back to /System/Library/PrivateFrameworks/VoiceServices.framework/TTSResources/ and delete the folder with the old language files, I.E., en-GB (don't worry, remember you already saved a backup on your computer so it's no big deal!). We now need to create a link to the new en-GB folder so that iOS will know where to look for the voice files. Right click anywhere inside the TTSResources folder window, select the option "New" and then the option "Link". A small dialog window will open, for "Link/shortcut file" type en-GB, and for "Point link/shortcut to" type the address of the folder that contains the actual voice files, /var/CustomVoices/en-GB. That's all, you're done! You can access your new voice by selecting what ever language you replaced on VoiceOver's language roter, so in our case, Tom will now be under British English. You may repeat these steps as many times as you like, adding new voices to different folders or simply replacing the stock compact voices with their premium variants. Note that you no longer need OpenSSH installed once you've replaced your voices, so if you're feeling paranoid, you may wish to remove it through Cydia. Cheers! -- You received this message because you are subscribed to the "VIPhone" Google Group. To search the VIPhone public archive, visit http://www.mail-archive.com/viphone@googlegroups.com/. To post to this group, send email to viphone@googlegroups.com. To unsubscribe from this group, send email to viphone+unsubscr...@googlegroups.com. For more options, visit this group at http://groups.google.com/group/viphone?hl=en. --- You received this message because you are subscribed to the Google Groups "VIPhone" group. To unsubscribe from this group and stop receiving emails from it, send an email to viphone+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/groups/opt_out.
