Older: Playing a tune that has silence at the beginning gets the moving feedback out of synch. May 6, 1986 5:38:33 pm PDT VoiceInTextImpl.AddVoiceMarkerAtPrimarySelection does not check for a point selection. BoundsFault. August 21, 1986 6:11:39 pm PDT February 17, 1987 6:24:36 pm PST 1. Adding a textual marker refreshes the whole viewer after each character on the color display. For a long annotation, this can be very slow, unless you adjust the window smaller. On the bw display, it just blinks the display in a rather annoying fashion. 2. Where did Stephen put his new version when he was here in September? 3. The sound viewer frequently gets into a state where you have to close and reopen it in order to be able to select anything. Redraw works at least sometimes for this. 4. Trying to insert a temporary marker during playback is difficult, since the selection is being used to move the playback cue along. You are more likely to get the character that was selected before the playback operation was requested. Dan suggests that what you really want is to have one Mark operation insert a mark at the current playback spot, w/o any selecting. 5. Opening a Dictation Machine from a sound viewer and then clicking PlayFromSelection plays from the selection in the original viewer to its end -- it does not continue into the new sound viewer. It seems to know where a single selection is in some sound viewer, and clicking PlayFromSelection/ResumeFromSelection in any sound viewer plays/resumes from that single selection to the end of its viewer. Clicking ResumeFromEnd in some sound viewer resumes from the end of its own sound viewer only. 5a. It feels like DictationMachine should create a selection at the beginning of the new voice when recording begins. That way the user can repeatedly record and review without having to manually add selections. Of course the user could select something else later if he wishes. Another option: have a PlayFromMark, with DictationMachine automatically inserting Marks. 5b. When recording from the hands-free microphone, either the beep is too long or recording begins too quickly -- you get a bit of the beep at the start of recording. Recording from the handset is ok. 6. I can put voice in a Walnut Sender by clicking AddVoice in a regular Tioga viewer. Of course the interests for that voice will not be managed properly. 7. If you Mark a spot, open a new DictationMachine, and record as desired, you can't Store the DictationMachine at the original spot -- you can only copy it there. Also, if you have several marks, it may be difficult to recall which one to copy to - there aren't enough different colors for the editing history to help. (Using only a selection is even worse, of course, because you will instantly lose your place if you make any other selection anywhere.) 8. The color editing history is maintained over all sound viewers. That seems like the right choice. 9. Mark [left|middle|right] inserts a mark at the current selection (in any sound viewer) and scrolls that viewer to the top. If the selection is not in a sound viewer, it complains. 10. DeleteMarks [left|middle] deletes the selected mark (in any sound viewer) and scrolls that viewer to the top. It does not complain if no mark is selected. DeleteMarks [right] deletes all marks in its sound viewer and scrolls that viewer to the top. 11. If no character is selected during an AddVoice operation, TiogaVoice blows up when trying to create the talks bubble after recording completes. 12. Typing a CR in a sound viewer breaks the voice tube into multiple lines. These work correctly during playback. The multiple lines stay over a close/reopen, but editing the voice redraws at least the tail of the voice. Redraw also removes CRs. Interesting: inserting a CR sets the edited bit, but Redraw removes it if there have been no other edits. (How does it do that?) 13. Adding textual annotations to a sound viewer does not set its edited bit, so TiogaVoice refuses to save it. (I can't seem to reproduce this.) 14. Textual annotations on earlier pieces of voice are chopped off by textual annotations on later pieces of voice. Adding a CR does not allow the text to extend past the voice tube, but adding enough more voice will reveal the covered text. 15. Dan would prefer talks bubbles that live on a spot between characters and extend out to possibly obscure line/lines above. There could be text (names, permanent textual annotations, etc.) in these large bubbles. Seems to require planes (21/2D) in Tioga. Maybe you'd do a prototype in Gargoyle? They could be anchored to a group of characters rather than just one, say by reverse videoing the anchoring chars. 16. Saving the underlying Tioga document while a voice viewer is open records the talks bubble as open (sound viewer #n) permanently. Saving the sound viewer thereafter does not close the talks bubble. However, destroying the sound viewer does change it back. Maybe saving a sound viewer is not supposed to close the associated talks bubble. Storing it at a new text location closes the old location and opens the newly stored location. Here's what the above remark should say: 16a. Saving the underlying Tioga document while a voice viewer is open records the talks bubble as open (sound viewer #n) permanently. If you then destroy the Tioga viewer before the sound viewer, the file will have an open bubble in it the next time it is opened (whether or not there is a viewer of that name any longer -- and anyway it would have to be the same voice rope and textual annotations). You do get a message in the CommandTool: "detaching parent link for voice viewer 1". Then when you destroy the sound viewer later, you get: "Voice viewer 1 had no parent". You are not allowed to edit the erroneously "open" voice, because TiogaVoice thinks it's already open. 16b. Inserting new text earlier in the Tioga viewer than an open annotated spot causes the report: "source marker not found at expected position!!". So the following scenario is a problem: AddVoice, STOP, EditVoice, save Tioga doc (w/ open bubble), edit voice, insert char in Tioga doc before annotated spot, Save sound viewer. Fails with the above report. If you then try to Store the sound viewer at its old spot, you get "cannot add sound on top of sound". EditVoice, edit voice, delete char(s) in Tioga doc before annotated spot, {save Tioga doc (w/ open bubble),} Save sound viewer works ok! EditVoice, insert char in Tioga doc before annotated spot, edit voice, Save sound viewer: complaint in the Message Window: "unable to find source marker for voice window: voice not saved", along with the CommandTool report. If you insert 1 char and delete a few chars, it is somehow possible to get a BoundsFault from Tioga trying to find the $voiceWindow character property. 17. Once you have inserted a textual annotation, say by typing a space, you can never remove the little arrow, no matter how far you backspace. It's not just something cached, because destroying the sound viewer and recreating it displays the arrows again. RTiogaVoiceNotes.tioga Polle Zellweger (PTZ) March 3, 1987 7:16:20 pm PST Κ• voicelistP&PolleZ.pa#603701467&PolleZ.pa#603701676&PolleZ.pa#603723235&PolleZ.pa#603723286˜™Icode™2—J™J˜J˜J˜rJ˜Jšœƒ˜ƒJ˜J˜ J˜Jšœ…Πbiœy˜‚J˜J˜GJ˜J˜©J˜J˜τJ˜J• CharPropsg voicePolleZ.pa#603701467Artwork TalksBubble voicePolleZ.pa#603701676Artwork TalksBubbleƒXšœΏΟiœ–žœ˜ςJ˜J–ΣvoicePolleZ.pa#603723235Artwork TalksBubble textInVoice50:48:From Brad Myers's window paper - from microphonevoicePolleZ.pa#603723286Artwork TalksBubble textInVoice0:24:Brad again, from handset‚W˜σJ˜J˜ΙJ˜J˜›J˜J˜ΙJ˜J˜eJ˜JšœHžœž"œ;˜·J˜Jšœ<žœž"œZžœ1˜ώJ˜J˜“J˜J˜ϋJ˜J˜’J˜J˜ςJ™JšœτΟuœΟdœ©˜ J˜˜Ήšœ(˜(Jšœ¨˜¨J˜JšœΞ˜ΞJ˜šœžœa˜~Jšœ ˜ —J˜šœ žœH˜YJšœ…˜…J˜—J˜˜——J˜J˜‚—…—ξE