Xenakios's Blog

Another progress report

xenakios — Wed, 01 May 2013 15:54:45 +0000

I haven’t been working on HourGlass much for the past weeks as I got distracted by another thing that hopefully makes a public appeareance soon. I certainly haven’t forgotten HourGlass, but this new project has given me some nice change after working for so long and much with the HourGlass codebase. The new thing is also an audio processing thing, but not a standalone application…That’s all I am willing to tell it this point…

Progress report

xenakios — Wed, 10 Apr 2013 08:51:15 +0000

I am planning to release the next HourGlass version within the next couple of weeks. I have been a bit sidetracked by another project, which will affect HourGlass at some point. As a teaser (but this won’t be part of the next public release) :

Ancient history, before HourGlass…

xenakios — Mon, 18 Mar 2013 23:06:13 +0000

Before I started working on HourGlass in “real C++” and using the Qt framework, there was this app, “Texturizer”, written in Microsoft’s “managed C++” and the .NET framework :

I unearthed the sources and made the app run and play some sound, just to see what the thing was like. I couldn’t even run the .exe directly now because it has the sound file to be played hardcoded in the source code, and the file in the .exe built in 2009 was no longer in that location on my hard drives… I have no idea what the app is really doing with the sliders etc, I didn’t bother looking in the source code that much for now…

People might think that perhaps HourGlass should have stayed that simple…But then there really wouldn’t have been much point for me personally to ever do it…I kind of knew I wanted and needed HourGlass to be something much more complicated than just a simple granulator.

HourGlass fragment voice send automation

xenakios — Sun, 17 Mar 2013 12:04:24 +0000

Not sure why it took me so long to work on this, I basically got it working in 15 minutes or something…(Besides the probabilities, the send volume levels are also automatable now.) Regarding the next public build, have some patience, I will get it out at some point…

HourGlass 1.3.0 released

xenakios — Tue, 19 Feb 2013 11:01:50 +0000

This build hopefully fixes the issue with the offline chain processor GUIs not being available after loading a HourGlass workspace file.

Windows : www.xenakios.liitin.net/public_html/HourGlass_1.3.0.zip

OS-X 10.6/10.7/10.8 (Beta, requires Intel 64 bit processor) : www.xenakios.liitin.net/public_html/HourGlass_1.3.0beta2.dmg

By the way, this release doesn’t of course reflect what I’ve done for the past month…There’s more to come relatively soon…

HourGlass effects crossfading

xenakios — Fri, 15 Feb 2013 09:21:28 +0000

In-development feature which allows setting up many effects in parallel chains and crossfading between them. (The chain volume envelopes are fully editable to any shape, I just didn’t show that in this video.) This is still missing some stuff like reordering the effects in the chains and automation of the volume morph parameter…

I will soon release a very minor fix release for HourGlass 1.3.0, which I suppose will be the “final” 1.3.0 release. And then later 1.3.1 which should have the sound file varispeeding and this effects rack/chainer thing.

Varispeed bending

xenakios — Sun, 10 Feb 2013 12:44:29 +0000

It’s been requested a few times that HourGlass could do plain old resampling based pitch bending/varispeeding, and I’ve sometimes wanted that myself too, so today I worked on that…This is now a process in the offline processing chain, instead of a new main processing mode. (In the video I am just playing the sound source preview looped, the granulation/fragmenting engine isn’t involved.) It might be possible to have it as a new main processing mode too (in addition to the Classic and Play Forward fragmenting modes), but I will have to consider if that’s too hard and messy to add…

HourGlass undo history flaw and a planned solution

xenakios — Sun, 27 Jan 2013 20:20:10 +0000

Currently it works like this :

So, each undo history step stores each envelope regardless of the envelope containing the same data or not…Hopefully I can in the near future improve it by introducing sharing of state data :

HourGlass 1.3.0 beta released

xenakios — Wed, 23 Jan 2013 19:09:56 +0000

Finally it’s here! There may be rough edges but as usual if you find bugs or quirks, let me know.

edit : I was in such a hurry to get this released I completely forgot there’s a feature that’s in a very unfinished state at the moment, the Audio Trimmer offline processing module. It has some passable basic functionality done but it isn’t really like it is supposed to be. So do explore it if you wish, but it might have all kinds of weird bugs and limitations at the moment.

Changes :

-Added triggable parameter gestures system. While you can operate it from the GUI 8×8 grid by clicking with the mouse (with the middle button, sorry Mac users, I think it doesn’t work there) or by assigning keyboard shortcuts or toolbar buttons to trigger the gestures, it probably works the best via MIDI or OSC remote control. This feature may be rather complicated at first to understand and use and probably warrants a separate Youtube video or blog post to explain it better. If you don’t plan to use HourGlass in a performance oriented way (live or by recording the audio output into a file), this feature likely isn’t going to interest you. (It might be developed later so that it’s possible to trigger the gestures from an editable timeline but at the moment it has no useful purpose when offline rendering the textures.)

-MIDI/OSC Remote control system was reworked. The scriptable value mapping system was removed (hopefully nobody misses that) and instead the incoming control values can be mapped using a visual envelope editor.

-The source sounds offline processing system can now remember audio peaks for all the intermediate processing steps, so that the envelope editors for the offline processing modules can now show the proper guide audio peaks. Note that when zooming in enough, the peaks or a waveform is not shown as that data is not available. (Because that would require storing the whole audio data for each processing step, which might be a bit too much memory wise for this…)

-Source sound buffers handling was improved. If the same audio file is loaded into HourGlass several times, only one real copy is kept around. Offline processing no longer makes a redundant copy of a file if the processing chain has all processing modules bypassed.

-Envelopes/parameters lifetime management was improved, hopefully eliminating many bugs (including crashing ones) relating to adding/removing the envelopes.

-Envelope editors can be locked to show a particular parameter envelope.

-When plugins are scanned, a visual indicator is shown that the scan is going on and if errors were encountered, a log is shown. On Windows the plugin scanner process won’t show the usual Windows crash dialog, to allow the scan to go faster if a plugin crashed. A solution for OS-X seems to also exist to not show the crash report dialog, but it’s more complicated to properly disable that, so it wasn’t yet implemented.

-Lots of more minor changes that will be added to this blog post later. Now I just want to get the new version out into the world after so much delay…

Download (download links changed on 18th February 2013 to another, unfortunately slower, location) :

Windows : www.xenakios.liitin.net/public_html/HourGlass_1.3.0beta.zip

OS-X (10.6/10.7/10.8, Intel 64 bit processor required) : www.xenakios.liitin.net/public_html/HourGlass_1.3.0beta.dmg

Forget pointers and embrace copying values

xenakios — Tue, 15 Jan 2013 16:45:28 +0000

I recently watched an amazing talk by Sean Parent on C++ at :

It’s rather technical (apart from the Photoshop demo around the 54 minute mark) and unfortunately with very bad audio quality. The rather ironic thing is that some of the stuff he is talking about would have been already available for me to use in the Qt C++ framework, which HourGlass uses heavily, but I never fully realized how easy the value semantics and copy-on-write stuff would have been to use. (Because it all seemed too good to be true.) Of course at this point in HourGlass’s lifetime I might not be able to benefit from those techniques so much, as using value semantics is a rather fundamental decision for how code should work. But for any following software projects looking into that stuff will be a must.

There are obviously points against value semantics and copy-on-write. For one, one might be tempted to implement features such as the Photoshop history, which can chew lots of memory when the user uses such features in certain patterns(*). A counterargument to that would be that the user would/should have already been aware that such features simply can’t come for “free”. And isn’t the RAM on modern machines there to be actually used, not to show impressive figures about how much of it is NOT used…?

The thing I didn’t yet mention is the concepts based polymorphism. Mainly because this is stuff that I didn’t yet try implementing in code myself as it is a bit difficult for me to understand well at this point. (For the value semantics and copy-on-write based stuff I already wrote a neat but useless image manipulation/drawing thing which implements the Photoshop style history brush thing.)

(*) For example if the user consistently makes changes to the current document that change everything in it compared to the current state, there’s nothing to be gained in sharing unchanged data.