![]() |
![]() ![]() ![]() ![]() ![]()
|
next newest topic | next oldest topic |
Author | Topic: Problem descriptions and Questions from a new user [LONG] |
tuscland Member |
![]() ![]() ![]() Hi, I am Camille Troillard. I just had my Kyma last Friday and I was waiting for this moment since 4 years. Needless to say that I am now one of the happiest man in the world! Here is the result of some of my work along with the manual.
1. There are a few redraw glitches in the Automation pane when setting a parameter using the Enter key : the previous automation is still displayed until an explicit redraw command (like zooming) is issued. 2.a. It is not possible to use the insert-marker function on Windows XP, the marker are always placed at 00:00:00:00. 2.b. While paying, pressing the insert-marker button in the timeline window, then pressing ESC brought this error message. (see the stack-trace below) 2.c. Soon after the Capybara stopped responding (The Capybara is not responding, no error) and reinitialized successfully. Those problem are perhaps unrelated. This error is not reproducible. 2.d. The MIDI program change event doesn't work properly and make the Capybara crash (Signal too short). 2.e. All these previous problems seem related : removing the "Clicks" Sound solves all the previous problems. 3. Hover help message label is bogus on the Marker MIDI Program Change button : "If you send this Program Change number on channel nil", even if a channel number is selected in the Timeline MIDI Channel Number list control. Setting a MIDI Channel Number in the MIDI Configuration panel (under DSP menu) doesn't solve this problem either. 4. 'Freq/Time Scaling Vocoder' has no replaceable input (c.f. page 67 of the manual), so control-dragging Disk-Bass loop has a strange effect (and should not work?): the sound Freq/Time Scaling Vocoder is completely replaced by Disk-Bass loop. 5.a. There are several places in the Kyma program where there are 8 audio inputs/outputs displayed although I do not have the audio expansion card. If I choose in the preferences to display only 4 I/O's this doesn't change the number of I/O's everywhere. Is it a normal behavior ? I guess it is for keeping the ability to make surround mixes with only 4 outputs. 5.b. HOWEVER: as seen on page 68 and previous pages the Kyma should only display the correct available number of I/O channels. 6.a. Setting Track Level and Angle controls to live doesn't generate the proper VCS : on a Timeline with 2 tracks (second track contains 2 sounds), the generated VCS only has the controls for Track 1. If I set Tracks 1's controls to Automated, then I can see Track 2's parameters. Not both as expected on page 85 of the manual. 6.b. On another Timeline the problem arises again with a bogus "Panner" (instead of "Angle") parameters generated from a sound of Track 2 (which is the only sound of the track, and has been cached). 7. Changing the cache file number of channels doesn't invalidate any cache. It seems to be a desired effect, but it is inconsistent with sound recaching policy. 8. The sound CloudBank Resynthesis loops when played live in the Timeline, but when it is cached, it only plays one iteration of the sound. I expected the behavior to be the same in live or cached mode at the exception that the setting of the sound wouldn't be editable. NOTE: I tried to re-cache the sound and this time it has now its full length. This strange behavior is not always reproducible. 9. Recording sounds in cache is a bit paradoxal: if you need to save some processing power then you can use the cache function, but in practice the cache recording is almost always done "out of real time". Caching should be done in a sort of "offline mode", meaning of real time, with outputs muted and a modal window asking the user to wait or click the mouse to cancel. QUESTIONS: 1. By pressing '-' on a Timeline region, is it possible to remove the last effect applied to the region. When control-dragging a sound over a region, it is possible to replace the region's input with the dragged sound. Is there a possibility to restore the original input sound of the region befroe control-dragging just like when typing '-' when the region is selected ? 2. Recording Live Parameters from an audio source generates a lot of events which locks the host computer during and after recording. How is it possible to limit the number of generated events in order to control a Timeline Parameter from an audio signal (as described in page 81) ? STACK-TRACE: Kyma encountered this unexpected condition: Subscript out of bounds: 14 This may have occurred due to a problem in one of your Sounds. ---------------------------------------------- (0 0 0 0 0 0 0 0 0 0 0 0 0 ) (class Array) received message #subscriptBoundsError: sent from KymaDamageEvent(a KymaDialogWindow): 91940778 597@376 (25@93 corner: 121@116 134@93 corner: 230@116 243@93 corner: 339@116 117@55 corner: 213@56 117@74 corner: 213@75 117@56 corner: 118@74 212@56 corner: 213@74 ) (class KymaDamageEvent) received message #dispatchFor: sent from KymaKeyboardEvent(a KymaDialogWindow): 91940600 597@376 $ 1B 0 (class KymaKeyboardEvent) received message #dispatchFor: sent from IP: Logged |
SSC Administrator |
![]() ![]() ![]() Thanks, Camille. We've printed out your list and are working through one item at a time. More soon... IP: Logged |
HectorBenard Member |
![]() ![]() ![]() Just a quick note about point 9, caching in the timeline... you can always play subportions of the timeline when you're caching, so that you won't have a problem with being out of real time while doing it. Just mark the sound(s) you're caching and whatever else you want to hear with it and play that subportion of the timeline (see the Kyma manual). I have also encountered the redrawing glitch in the parameters of the timeline, as well as that strange message about "capybara not responding, no error". I've seen that message a few times in the last couple of days in two different capybaras. I'd be curious to know what that is about. In one of them it was happening when trying to start kyma for some reason (just at the point when you'd expect to hear Carla's voice), and it wouldn't initialize. In the other one it happened when playing some sinewaves through the oscilloscope. Congratulations on getting a Capy Camille, and have fun! H IP: Logged |
SSC Administrator |
![]() ![]() ![]() By the way, Camille, thanks for reading the book! ![]() Also, I like Hector's suggestion of compiling just a subportion when you are caching so you can still hear iit playing in realtime. Now to the points: >1. There are a few redraw glitches in the Automation pane when setting a parameter using >the Enter key : the previous automation is still displayed until an explicit redraw command >(like zooming) is issued. We tried each of these but could not replicate the behavior you report. Some questions: >2.d. The MIDI program change event doesn't work properly and make the Capybara crash >(Signal too short). You're right, something about the Clicks Sound doesn't seem to like the MIDI program change. We are looking into this one. Other Sounds seem to work fine with the MIDI program changes. Thanks for reporting it. >3. Hover help message label is bogus on the Marker MIDI Program Change button : "If you >send this Program Change number on channel nil", even if a channel number is selected in >the Timeline MIDI Channel Number list control. Setting a MIDI Channel Number in the MIDI >Configuration panel (under DSP menu) doesn't solve this problem either. FIXED! Thanks. (The help messages are formed when the window is first opened so it did not know which MIDI channel was selected yet). >4. 'Freq/Time Scaling Vocoder' has no replaceable input (c.f. page 67 of the manual), so >control-dragging Disk-Bass loop has a strange effect (and should not work?): the sound >Freq/Time Scaling Vocoder is completely replaced by Disk-Bass loop. >5.a. There are several places in the Kyma program where there are 8 audio inputs/outputs displayed although I do not have the audio expansion card. If I choose in >the preferences to display only 4 I/O's this doesn't change the number of I/O's everywhere. Is it a normal behavior ? I guess it is for keeping the ability to make >surround mixes with only 4 outputs. You're right, the menu for the Timeline Track inputs was always saying 8 channels. This has been fixed to show the number of channels that you select in the Preferences. Thanks! (DSP Status was already displaying the proper number of hardware inputs/outputs). >6.a. Setting Track Level and Angle controls to live doesn't generate the proper VCS : on a Timeline with 2 tracks (second track contains 2 sounds), the generated >VCS only has the controls for Track 1. If I set Tracks 1's controls to Automated, then I can see Track 2's parameters. Not both as expected on page 85 of the manual. I suspect that you may have skipped the step of selecting all of the Sounds first (using Ctrl+A). The automation menu affects *only* the selected Sounds' tracks. >6.b. On another Timeline the problem arises again with a bogus "Panner" (instead of "Angle") parameters generated from a sound of Track 2 (which is the only >sound of the track, and has been cached). Could you email this one to us so we could try it? >7. Changing the cache file number of channels doesn't invalidate any cache. It seems to be a desired effect, but it is inconsistent with sound recaching policy. Interesting question. Does changing the number of channels or bit depth refer to any *future* caches? Or should changing that information force you to redo all the caches you have already made?
Were you using a WaitUntil? >9. Recording sounds in cache is a bit paradoxal: if you need to save some processing power then you can use the cache function, but in practice the cache >recording is almost always done "out of real time". Caching should be done in a sort of "offline mode", meaning of real time, with outputs muted and a modal window >asking the user to wait or click the mouse to cancel. Often, though, one *would* like to hear the Timeline, even if it is running behind real time. I liked Hector's suggestion to ctrl-click only *some* of the Sounds and then using shift-space bar to play. That way you can monitor a subset of the Sounds in real time while caching. >1. By pressing '-' on a Timeline region, is it possible to remove the last effect applied to the region. When control-dragging a sound over a region, it is possible to >replace the region's input with the dragged sound. Is there a possibility to restore the original input sound of the region befroe control-dragging just like when typing >'-' when the region is selected ? I think you are asking whether it could replace the ORIGINAL ReplaceableInput (the one that the Sound had when you first dragged it into the Timeline). Say you have Sound-A in the Timeline. When you drag and drop Sound-B onto it, Sound-A becomes the ReplaceableInput for Sound-B. This is necessary in order to make the "-" work. The downside is that Sound-A's original ReplaceableInput is no longer remembered (there can be only ONE ReplaceableInput in the Sound). Changing this behavior would require giving up the "-" functionality. >2. Recording Live Parameters from an audio source generates a lot of events which locks the host computer during and after recording. How is it possible to limit the >number of generated events in order to control a Timeline Parameter from an audio signal (as described in page 81) ? Maybe we should offer an option of downsampling the Sound input to the MIDI rate so you do not end up with as many breakpoints? We *do* drop points that lie on a straight line connecting two other points, but when you are recording audio as a MIDI controller you still end up with lots of points. In the meantime, one thing you could try this: in your SoundToGlobalController, downsample the input Sound. For example: [AmpFollower] L: 10 would sample the AmpFollower once every 10 milliseconds, resulting in fewer breakpoints in the Timeline. IP: Logged |
SSC Administrator |
![]() ![]() ![]() Sorry, our website was down temporarily... [This message has been edited by SSC (edited 20 April 2005).] IP: Logged |
SSC Administrator |
![]() ![]() ![]() Problem with Clicks has been fixed! We'll make a new update and post it sometime in the next couple of days. Please check for it on the tweaky (http://www.symbolicsound.com/tweaky) in a couple of days. Thanks! IP: Logged |
tuscland Member |
![]() ![]() ![]() > By the way, Camille, thanks for reading the book! The book is great! > Also, I like Hector's suggestion of compiling just a subportion when you are This is a very good idea, the partial-timeline compilation is a handy feature. > Now to the points: No, I think it only happened is the Automation window of the Timeline. > * How long was the Timeline (total duration of the longest Sound)? And what was I'd say 60 seconds. This is the timeline from the example in the book that uses > >4. 'Freq/Time Scaling Vocoder' has no replaceable input (c.f. page 67 of the Uh ok!
Nope ;-) I tried many times, by selecting a few different controls but without > > >6.b. On another Timeline the problem arises again with a bogus "Panner" [see previous answer] > >7. Changing the cache file number of channels doesn't invalidate any cache. It I am thinking loud : That leads to that question : why is there the choice for a bit depth? This is > >8. The sound CloudBank Resynthesis loops when played live in the Timeline, but No. > >1. By pressing '-' on a Timeline region, is it possible to remove the last I think that I missed something about the concept of control-dragging a sound > >2. Recording Live Parameters from an audio source generates a lot of events That is what I thought doing later, but I think that the first option is IP: Logged |
pete Member |
![]() ![]() ![]() Hi Camille I just wondered if you realized that in almost all cases if you get the error message "Out of real time", it still continues working if you leave it, and that although the sound is chopped up while it is being prossesed it will end up sounding clean when you play back the cached version. Pete [This message has been edited by pete (edited 20 April 2005).] IP: Logged |
tuscland Member |
![]() ![]() ![]() Hi Pete, Yes, I saw that. Cam [This message has been edited by tuscland (edited 21 April 2005).] IP: Logged |
tuscland Member |
![]() ![]() ![]() Ok, so concerning the track parameters problem (1) and the incomplete sound cache problem (2) : * Problem 1 This one is easy to reproduce : a. Create a Timeline with "Freq/Time Scaling Vocoder" on track 1 and "Clicks" dragged from the prototypes on track 2. The problem is reproducible and the bogus control parameter is randomly chosen, depending on the sound that is used in the Timeline. I can send you the Timeline if you want it. Can I send it to symsound@SymbolicSound.com ?
If Timeline playback is prematurely interrupted, the sound caching is wrong, and it doesn't seem to be recached correctly neither.
IP: Logged |
SSC Administrator |
![]() ![]() ![]() Here's a possible explanation for problem 1: the track level controls are mapped (by default) to MIDI controller 7 on the track's channel. If the Clicks Sound had its Level control mapped to the same MIDI controller and channel, then both controls would be controlled by the same MIDI controller, but Kyma would draw only one control. Does this explanation fit? IP: Logged |
tuscland Member |
![]() ![]() ![]() quote: I think this explanation fits. Still, the behaviour is not the one that I expected. As I described, no custom MIDI parameter controls has been set, expect the ones that are set by default. I guess that if I choose to display the selected track's Level and Angle, I should get only those parameters in the VCS, is that right ? Is there anything special to care about when editing track's level and angle parameters? IP: Logged |
tuscland Member |
![]() ![]() ![]() I would to add something to my reply. As you pointed, "Freq/Time Scaling Vocoder" has its Volume parameter mapped to the same MIDI control as the Level's track. Choosing another control change for the Volume parameter solved the problem I described. However, I still think that when automating track's Level and Angle parameters, they should be displayed for the selected sound(s). If there is the same MIDI parameter, the correct behaviour is maybe to display both of the controls, in order to get the expected result? IP: Logged |
SSC Administrator |
![]() ![]() ![]() Yes, it should work as you describe. We will look into what it would take to fix this. IP: Logged |
All times are CT (US) | next newest topic | next oldest topic |
![]() ![]() |
This forum is provided solely for the support and edification of the customers of Symbolic Sound Corporation.