There is a wavelet transform plugin in the QM Vamp Plugins set (and therefore in the Vamp Plugin Pack). I'm not sure how useful it is. Probably not very, as it stands, because I can't recall ever hearing of anyone using it. I know very little about wavelets and can't really judge, but I would be curious to know what a more useful plugin might look like.
Wavelet conversion
Having this same issue two and a half years later in v4.5.1 64-bit on a Macbook pro on Ventura 13.2.1. Screenshots would be nearly identical.
ReplayGain tag support for volume normalization playback
hello using Sonic Visualiser shows me empty minutes of audio file. How can I mark from current point to the point of mouse to delete this audio part? I want to delete empty audio seconds/minutes. Is anyone interested in german localization of Sonic Visualiser?
This is a much needed feature for any serious spectral comparison analysis. It is a shame that it has not been developed in the last 15 years.
Ability to change measure/beat separator char for automatic labelling
Ability to change decimal separator on CSV export
Jack Transport in SV
Temporary UI freeze on play/pause when using pulseaudio
In v4.2, File > Export Image File... and File > Export SVG File... export incomplete images for a Spectrum Pane
Auto alignment doesn't work in version 4.2.
I just realized that it doesn't necessarily work the way I originally assumed, but perhaps instead finds the best points to alighnment and stretch vs. squoosh factors at each. So I guess I then have a question: within the suite of sonic visualizer apps, is there a way to export corresponding onsets times of two audiofiles that renditions of the same music?
I just that it doesn't necessarily work the way I originally assumed, but perhaps instead finds the best points to alighnment and stretch vs. squoosh factors at each. So I guess I then have a question: within the suite of sonic visualizer apps, is there a way to export corresponding onsets times of two audiofiles that renditions of the same music?
export timings from Sonic Lineup
Correct dB Scale
Font Size Reverts to Default in Exported Graphs
My responses are preceded with [JG] below. Thanks, Jean Goulet From: Chris Cannam cannam@users.sourceforge.net Sent: Tuesday, June 16, 2020 4:48 AM To: [sv1:bugs] 283@bugs.sv1.p.re.sourceforge.net Subject: [sv1:bugs] Re: #283 Many vertical lines on the spectrogram (there should be none) Thanks for the extra information! If I understand correctly, you are always recording the signal into Sonic Visualiser and displaying a spectrogram "live" as it records? [JG] Yes, live. Presumably if you load the...
Hi again - I've just put out v4.1 of Sonic Visualiser - I had hoped to take a look at this bug in time for that, but in the end I decided it wasn't practical as the release has been delayed a few times already and we really wanted to get it out of the door. It might be worth re-testing your situation using the new release, though I'm fairly sure there's nothing in it that should make a difference here. In the meantime, if you are able to answer the questions from my previous comment, that would be...
Thanks for the extra information! If I understand correctly, you are always recording the signal into Sonic Visualiser and displaying a spectrogram "live" as it records? Presumably if you load the original audio file directly into SV its spectrogram comes up clean? What is the audio recorded by SV like? (You can find the recorded audio files under File -> Browse Recorded And Converted Audio) - is that a clean signal or is the noise shown in the spectrogram actually present there as well? I'm trying...
Thanks for looking into this problem. The version is here: [cid:e638fd83-7b57-4357-a65b-f4909d577657] I have attached a sample tone of 250 Hz. However, note that the problem occurs regardless of the source audio (voice, music, tones, noise). The louder, the better, to see the problem. This file is being played into an audio device A (source) and transmitted to another audio device B. I am trying to use Sonic Visualizer to show the output of B in real time. The "record device" is set here: [cid:aebc81db-993e-428c-854c-4df11da6c4f4]...
Hi Jean - this is alarming, thank you for reporting it. Which version of Sonic Visualiser is this? For completeness, could I ask you to attach an audio file with which it exhibits this problem?
Many vertical lines on the spectrogram (there should be none)
piper-vamp-simple-server and vamp-plugin-load-checker not executable
rubberband devendoring required
@cannam this can be closed
@cannam thanks this works! :)
@cannam thanks, the patch suggested in https://sourceforge.net/p/sv1/bugs/282 fixes this problem as well, as now --as-needed is applied :)
The big advantages of qmake are (i) it's always there on any platform I'm building on, because it comes with Qt, (ii) it can produce Visual Studio project files on Windows rather than relying on make, and (ii) it isn't cmake (which I really dislike). I'm going to have to switch to something else at some point whatever happens, though, because qmake is gradually being deprecated by Qt. I had hoped to move to Qt's replacement build system qbs, but they abandoned that as well and are now moving to cmake....
Bug in configure I think. Try this
This basically means: exported $LDFLAGS are completely ignored by the build system. I've by now tried setting LDFLAGS_RELEASE in configure.ac to my global LDFLAGS, specifically exporting LDFLAGS when calling configure and make and setting QMAKE_LFLAGS to my global LDFLAGS. None of that works though
no full RELRO in piper-vamp-simple-server, sonic-visualiser and vamp-plugin-load-checker
This ticket can be closed. sonic-visualiser 4.0.1 builds both against capnproto 0.7.0 and 0.8.0
Overlinking issues in piper-vamp-simple-server, sonic-visualiser and vamp-plugin-load-checker executables
I can recommend not using qmake. So far it's the most troublesome build system (of upstream projects) that I had to deal with in the ~500 packages I maintain. Good cross-platform alternatives are meson and cmake. Aside from that: In 4.0.1 this is fixed, so I think this ticket can be closed! :)
I think you might be seeing this because of https://sourceforge.net/p/sv1/bugs/277/
Build fails with Qt-5.14.2: error: 'start' is deprecated: Use QElapsedTimer instead
Hi - it's the README that's wrong I'm afraid, the option really isn't supported. This is because the lab reader has no ability to include the source filename in the output, so a lab-one-file option would just merge lots of features without making it clear where they came from. The CSV writer is more fully-featured and can also write tab-separated files (--csv-separator '\t')
Lab output --lab-one-file
This is a quite useful idea. Would also like to obtain the feature. Yong
Expose peak frequency spectrogram data
This is happening in v4.
PowerCurve not working
Thanks for the note! I agree, -Werror shouldn't appear in the default release build. It looks as if it was sneaking through from the project files in the checker subrepo. I've updated and committed those so that it is no longer used in the default configuration. I prefer to keep -Werror in place in debug builds, because it is so useful as to be mandatory within the development cycle. Ideally we would have a third build type to distinguish between "in development" (with -Werror) and "occasional debug...
The -Werror compiler flag should be removed
I note that although the executable flag appears in the documentation for Qt 5.13, it does not appear in (at least the online copy of) the docs for any previous release. So it was either absent or undocumented previously. Undocumented seems likely, since the no_check_exist flag was also absent and I've been using that for ages. (That one is now documented, but wrongly - the docs call it no_check_exists, which does not work.)
Thanks for the report. I have had a surprising amount of trouble with the install target configuration in qmake. After some hunting, I think the situation is this. If the executables exist already in the build tree when qmake is run, then they are correctly identified as programs and installed using the QINSTALL_PROGRAM rule. This is the normal situation when doing repeated build/installs in a working development tree, which is no doubt one reason why I keep overlooking problems with the install...
piper-vamp-simple-server and vamp-plugin-load-checker not executable
Tests don't pass on ppc64
Playback Speed causes Crash
It's great to have sonicvisualizer working again. Thanks a lot Chris!
Following a further discussion off-tracker, it appears that: This problem has some connection to the use of a deprecated macOS API in PortAudio It affects any application using PortAudio right up to the current PortAudio code (this bit of PortAudio was last modified in the repo in late 2016) The problem is resolved by applying this patch from Tim Bunnell to update PortAudio so as to replace the deprecated API: https://lists.columbia.edu/pipermail/portaudio/2016-December/000992.html I'll look into...
Following a further discussion off-tracker, it appears that: This problem has some connection to the use of a deprecated macOS API in PortAudio It affects any application using PortAudio right up to the current PortAudio code (this bit of PortAudio was last modified in the repo in late 2016) The problem is resolved by applying this patch from Tim Bunnell to update PortAudio so as to replace the deprecated API: https://lists.columbia.edu/pipermail/portaudio/2016-December/000992.html I'll look into...
Hi Junqi - ok, I will see if I can find out anything. Meanwhile, can I please check: Do you have any audio devices installed other than the built-in mic/speaker? (e.g. USB devices) If you go to Preferences -> Audio I/O in SV, are there any other options you can select for "Audio playback device"? Is there anything useful in the log file at ~/Library/Application Support/sonic-visualiser/Sonic Visualiser/log/sv-debug.log ? (You could send me a copy or attach it here if you like) Although my ability...
Hi Chris, I tried the v4.0 but the problem still persist. So I believe there's still a compatibility issue for PortAudio for Mac Catalina... Chris Cannam cannam@users.sourceforge.net 于2019年10月24日周四 下午8:56写道: The situation with Catalina seems to be ... complicated and contradictory. I have only recently seen it in action myself - I sadly don't have access to enough hardware to be able to run test releases of operating systems as well as live build ones. I have made some test builds for a v4.0 release...
Hi Chris, I tried the v4.0 but the problem still persist. So I believe there's still a compatibility issue for PortAudio for Mac Catalina...
Hi Chris, I tried the v4.0 but the problem still persist. So I believe there's still a compatibility issue for PortAudio for Mac Catalina...
The situation with Catalina seems to be ... complicated and contradictory. I have only recently seen it in action myself - I sadly don't have access to enough hardware to be able to run test releases of operating systems as well as live build ones. I have made some test builds for a v4.0 release of SV (https://code.soundsoftware.ac.uk/projects/sonic-visualiser/files/) and will probably make an official v4.0 upload later today or possibly tomorrow. After that I'm expecting to gather remaining compatibility...
Port Audio Fail on MacOS Catalina (10.15) + vamp plugin fails to load
change order of panes
Thanks for the note! The code is compiled against Qt 5.3 headers, not 3.3. Qt guarantees backward compatibility across all 5.x releases, so compiling against 5.3 headers and running the executable with dynamic loading of any more recent 5.x will work fine as long as the two are built to compatible C++ ABIs, which in this case they are. The warning (printed by Sonic Visualiser code, not by a library) is purely about the specific path under which private resources such as cached data are stored - an...
Improper dependency in ubunto download
library mismatch in downloaded executable for Ubuntu
Thanks, this fixed it for me! :)
This should work in v3.3 on Windows (64-bit build only) and also on Mac. The Windows support is new in this release, Mac has been there for some time.
use different code hosting
BinTray downloads are now publicised and have been updated for each of the last few releases - closing this.
Support for Opus and AAC
The Opus decoding feature has now been released in v3.3. I'm closing this report; please open a new report for any problems with the released version. Note that AAC is also supported on 64-bit Windows (new in v3.3) and on the Mac (not new, it's been supported there forever). This uses platform libraries, which is why it isn't supported everywhere yet. I don't really like this inconsistency! Ffmpeg support is a possibility at some point, though that does come with its own complications.
appimage opens audio files from cli only with absolute path
I believe this was fixed in v3.2, but I forgot to update the tracker to that effect - sorry! - it should certainly work in the current v3.3.
rubberband devendoring required
Released in 3.3
Vamps Plug In Page Server Unavailable
Thank you for the report - it's up again now. In theory, we had an automatic monitoring system watching this site - in practice it didn't report the problem, so I need to look into what's gone wrong there!
Thank you for the report - it's up again now. In theory, we had monitoring for this site - in practice it didn't report the problem, so I need to look into what's gone wrong there!
Vamps Plug In Page Server Unavailable
I am an idiot! After doing all the real work, I then made the two Linux binary builds without the Opus support configured in. Sorry. Please find 3.3-pre2 at the same location. There are only Linux builds, as the code is unchanged from pre1 and it's only the Linux packaging process that had gone wrong. Separately I'm looking at how to ensure this particular problem doesn't happen again.
The Opus decoding doesn't works, it gives error when trying to open the files, wether they have the extension .opus or .oga, this is the error message that shows: Window title: Failed to open file Audio required Unable to load layer data from "/home/myuser/music.opus" without an audio file. Please load atleast one audio file before importing annotation data And this is what it shows in the terminal: AppImage root is /tmp/.mount_SonicVgW1XR2 MainWindowBase::stop: suspending Pane::Pane(0x2426d10) returning...
Add builtin EQ
open m4a files
Add link to Bug Report /Feature Request under "Help"
As you see in the attached image the time instants were delayed as the window was being redrawn durring scrolling.
Time instant misplaced when inserting while scrolling
v3.3pre1 win64: Notes layer only navigates horizontally
Be able to choose to save spectrogram with a transparent background
Please try the 3.3-pre1 test builds, listed at https://code.soundsoftware.ac.uk/projects/sonic-visualiser/files - this is the first preview release of a forthcoming version with Opus support.
I'd like to see that, too. Even better would be to simply fall back to e.g. ffmpeg for decoding, if a format is natively unknown.
Add pitch change function
Pitch change
Hi there - can you give me any more info here, please? I tried the measure tool on a few different systems and did find some problems, but I'd like to know whether I'm barking up the same tree. I found: on all systems, at some zoom levels, when you drag out a box using the measure tool, the left edge of the box (or whichever edge you started at) wanders leftwards across the screen as you drag the right edge instead of staying fixed. This is a side-effect of the work I did on supporting additional...
Measure Tool does not work
Have de-vendored this library in the current repo, for next release (3.3).
This was specifically about providing a .deb that people who use Ubuntu (any recent-ish version from the last few years) can download, install, and run, in the case where they want to get the current SV version independent of what version is in the distro repositories. This is entirely a response to the way some real users interact with the software - as with every awkward interaction with the real world it would be nice not to have to do it, but as you can see from the download figures, a number...
Hm, sorry, but this doesn't make any sense! If Ubuntu has broken packaging of librubberband2, they need to fix that, not you and in turn break your source code for every other distribution! I suggest, that if you want to provide .deb packages (instead of letting downstream take care of that - always the easier solution) to link them statically all the way (but not force this on everyone else by making it mandatory). If you want to provide a deb for 16.04 specifically (but non-static), you need to...
This precise problem is one reason I added the alternate BinTray downloads, e.g. https://bintray.com/sonic-visualiser/sonic-visualiser/download_file?file_path=sonic-visualiser-3.2.1.tar.gz Unfortunately I see that I have completely forgotten to link the source code file from the downloads page - the only BinTray links on there are to the binaries. That was stupid of me and I should sort it out.
This was an exasperated hack in response to an Ubuntu packaging problem: https://code.soundsoftware.ac.uk/issues/1853 I agree there are certainly better ways to deal with this - perhaps for 3.3.
use different code hosting
rubberband devendoring required
Thanks for the extra information - I see what you mean. The CSV logic saw some updates between 3.1.1 and 3.2 to make it more consistent in a number of ways, but this is one pitfall I hadn't seen. I'll have to see about adding an option to choose the delimiter.