whisper = psithyrizo, psithyros (Greek)
Whisper
VST VST is a trademark of Steinberg Media Technologies GmbH.
ASIO ASIO is a trademark and software of Steinberg Media Technologies GmbH.

Canvas view Pattern view Score view
Canvas View (for plugins and routing) Pattern View (for notes, parameters etc.) Score View (for arranging the score)

Whisper Virtual Studio

Whisper is a lightweight music sequencing application for PC hardware. It hosts VST plugins for intruments, effects and other realtime add-on functionality, while implementing a minimalistic user interface emphasizing the actual musical sequences. A lot of music creation functionality is already available in pluggable VST(i) tools (both freeware and commercial ones) created by other people, but we felt the need for yet another simple and clean, and a bit different, way to use those tools with each other, without adding too much things on top. The development of Whisper was started to fit this purpose.

Whisper is probably more a tool for a composer than it is for a typical musician. It's a tool for writing music, not for processing audio as such, or for live performances etc. (while it might suit some live applications as well). It's designed for people who are not overly familiar with traditional DAW applications, and/or who want to find a way to realize their musical ideas quickly by "writing them down" and just letting the software to play the music. Another goal is to utilize multicore home PC hardware efficiently in music creation.

The idea is to only do a bare minimum, but to still do it well enough. Whisper is a specialized application and clearly doesn't suit everyone, but it also enables music creation in quite a different way than most other sequencers and DAWs. The development of the application started from our own needs, but we're certainly happy if someone else can benefit from it as well!


The application is "helpware", meaning that you don't need to pay money for it (it's basically freeware), but we'd like to get useful feedback that helps us make it better. You are not currently allowed to redistribute the software, but you can ask for download rights to get it for your own use. It is possible that the application will be made freely distributable later, depending on the feedback from actual users. Currently the main criterion for whether you can get download rights or not is whether you actually have the time and determination (and a need) to learn to understand Whisper.


Audio examples (ogg)

The following audio files are draft quality examples of Whisper controlling a few commercial VST plugins (mainly Reaktor and Battery VSTs from Native Instruments, and a vocal synthesis plugin). All the music, including vocals, was written in the Whisper UI, and routed, mixed, and mastered together in Whisper, while the actual audio (also for the vocal parts) was synthesized in realtime by the respective plugins.

Note: The first of the tracks ("Evolution") is quite an outdated example, as it's actually from an early state of Whisper development. The rest of the tracks have been created with various beta builds of Whisper.

1) "Evolution"
2) "I Am Whisper"
3) "Slow Traction"
4) "Fluence"
5) "Like A Tree"

You might need OGG Directshow filters or Quicktime plugins available at Xiph.org to play the examples in Windows or Mac OS X.


Introduction

Making music does not necessarily have to require lots of physical equipment anymore, and doesn't have to be too complicated either. The technology to enable this has already been there for some time, but while such applications are created that allow for example "drag'n'drop" type music creation easily using loops and samples, etc., what is often missing is direct, detailed control of the musical sequences themselves. On the other hand, there are very advanced audio workstation applications that might have so much functionality that you easily get lost while trying to use them effectively.

Where many other tools are in their own ways still trying to imitate real, physical studios with all their various equipment, and also contain lots of functionality for handling physically recorded audio or samples, Whisper is designed to start from the essence of the music itself: the musical sequences, themes and patterns that make it, while still leaving one important part - the actual audio generation - to other tools.

Almost every new PC has at least a dual-core CPU. These processors actually have quite much processing power, and none of the basic office programs etc. really utilize that all so well. On the other hand, musical instruments like synthesizers etc. tend to be quite expensive in their physical form, while there are lots of virtual instruments and effects available in the VST plugin format. Using software in general uses less physical space, and gives you more flexibility than using hardware, while also enabling you to make the same software instruments and effects you've already bought perform better by upgrading the computer hardware a bit, if you want to, perhaps not requiring any updates to the software separately.


Basic ideology

Why shout when you can Whisper?Why use more than you really need?

Whisper is an application designed for writing music "like text" using the PC keyboard and a mouse (optionally with help of additional MIDI controllers) and playing the music in audio form (with the help of pluggable software synthesizers and effects), instead of recording tracks with microphones etc. It is not meant for combining pre-made loops and samples with each other, either, as it strives for a more direct way of music creation all the way from the ideas and themes and instrument selection to a digital recording.

The basic ideology behind Whisper is to only do what is really needed, but to do that well enough.

The keywords in Whisper development are:

1) Simple        -> no unnecessary bloat or too complicated user interfaces.
2) Non-intrusive -> doesn't decide too much for you, but helps you in your work.
3) Effective     -> utilizes the available HW and SW effectively, in realtime.
4) Direct        -> does not hide what the software is actually doing behind too many abstraction layers.

"Simple" and "direct" can in some cases also mean that you'll have to work more to get things done (as there might not be high-level automated functionality to do the things for you). In general, compromises will have to be made between having direct control over details, and having automated setup of larger entities. Also "effective" is a relative term, because software can be optimized for different environments, and a tool which works effectively in one environment might not do that in another. In the Whisper case, the software has mostly been optimized for multicore systems, for instance, and it probably doesn't work very well in single-core environments.


Who is it for?

If you want to write music (somewhat like text) in a way that lets you realise the compositions immediately in audio form, and can do with the limited amount of supported features and a very plain user interface (or are even looking for just that kind of "minimalism"), using Whisper as the host for your plugins might feel like a breath of fresh air (or like a Whisper ;-)). It might not suit most people at all, but for someone it could still be just the essential package.

For someone who always wants to play with lots of hardware: physical synthesizers, various electrical modules, racks and such things, Whisper probably doesn't have that much to offer. It doesn't currently support that equipment in the same way as many other tools. It is not really either for people who like loops and samples, or recording audio and editing it "offline", because it does not support any use of samples itself (you need to use plugins that support samples, if you want to use any). It's simply a VST host application for writing music, supporting ASIO audio output and recording to a PCM wave file (WAV). But if you're mainly interested in the core essence, the actual music itself, and do not necessarily even have any skills in handling "real" instruments or recording equipment either, you could possibly benefit from using it.

If you are more like a programmer- or writer-type of person with musical aspirations, Whisper could really be much better for you than some other host software, because of its plain short string-, numerical values-, and column/row based UI. On the other hand, you can still play MIDI instruments to create musical patterns, and even the basic input of note data is still handled by simulating a keyboard-based instrument using the PC keyboard, so anyone having some experience with keyboard instruments should feel at least a little bit familiar.


Influences from other software

Some of the UI principles in Whisper have roots in the tracker-type music software of the 90's, originating from the Amiga computing scene. But still, Whisper is otherwise very far from the trackers of the 90's (although you could even add some kind of actual tracker functionality to it in the form of a VST plugin). What is tracker-inspired is mainly the way that the music is input by using the PC keyboard so that you can input notes by "playing" with the keys, that the music is represented as "tables" of information (text and numbers), and that there are so-called "patterns" involved in the music creation process. There are differences to trackers too, and for instance the hexadecimal numeric format that's often used in tracker-type sequencers (while it has its benefits as well) is not being used in Whisper.

The Whisper UI has also been inspired by a very good freeware music application (itself also a tracker of sorts) called Buzz, the main similarities being visualising the instrument/effect routing in the form of a node diagram (which I think is a brilliant way to visualise that), and using multi-pattern sequencing (which is actually quite practical, but is implemented in quite a different way in Whisper than in Buzz).


The technology

Whisper is designed for contemporary PCs, so it tries to utilize multicore processors effectively, which means letting you use more instruments and effects simultaneously in realtime, depending on how many and how fast processors, and with how many processing cores, you have installed in your system.

The reason why Whisper is based on Steinberg's VST and ASIO technologies is simple: Those are the only available and suitable plugin and audio driver technologies that are both widely used and free to use without extra royalties (if you sign the license agreements with Steinberg). That makes them virtually the only good solution for a freeware application that also supports commercial plugins and audio equipment, on both Windows PCs and Macs. Not supporting other types of "free" plugins (more proprietary ones, like Buzz machines, or a new plugin type) is mostly a question of simplifying the application, keeping the development more focused, and not bringing more limitations to the architecture.

64-bit floating-point values are used in audio calculations in many places in Whisper. Also creating 64-bit versions of the binaries themselves has been though of, but as it would require implementing VST bridges, it is to be left for a possible later version of the software. When you need more memory, 64-bit plugins, or more protection against faulty plugins, you can try some of the available VST bridges (the best "non-commercial" one seemingly being called "jBridge" at the time of writing this.


Virtuality

Whisper is intended to enable as virtual a studio environment as possible, using as little extra hardware as possible, and to minimize the amount of needed software too (depending on what VST plugins the user chooses to use with it). It's designed to be both lightweight and simple, but still complete enough as a VST host, while letting the plugins themselves do any more specialized realtime work. Whisper tries to let you control the plugins in a direct and simple manner, and not to get into the way of creativity. It still tries to let you control all the necessary things when you want to, in an exact manner, whenever possible.

Automation

As it's designed for realtime music synthesis, Whisper tries to help you to conveniently automate all those things that you don't really want to do manually, or wouldn't be able to do anyway, like "turning controller knobs of several instruments at once in a synchronized way". The automation is done without hiding too much of what the instruments themselves are really being "told" to do, so that the control is (at least in theory) quite direct.

Sometimes this level of automation can make the music sound a bit "robotic" and "too controlled". There are some additional planned features which might help make the result more human and organic, but of course "exact" programming by numbers etc. will not give the same results as controlling your instruments physically. Still, you can always give your music more human touch by how you program it. Whisper will give you the possibility to let all the instruments play at the same time, instead of recording tracks separately and combining them afterwards, so you can think of Whisper as more of a virtual music box or a virtual conductor than a multitrack recorder.


Software support

In principle, Whisper supports most VST 2.4 -compatible plugins, at least in some way. Offline processing plugins are not supported (Whisper is basically a realtime application), and category shell plugins (collections of plugins inside the same DLL file) are not currently usable with it either. Some plugins (like many plugins based on SynthEdit) are not working well with Whisper especially when using multiple instances of them, and this is likely caused by deficiencies in the thread-safety implementation of the plugins themselves.

Whisper has special support for certain vocal synthesis plugins. It gives you the possibility to actually type the phonemes and control the parameters etc. directly in Whisper itself, effectively removing the need to use an external editor separately to create vocals. That support was essential for providing an integrated working environment, and to not "spoil" the directness aspect of Whisper. The wish to write vocal parts without using a separate application was actually one of the reasons to start the Whisper development in the first place. There is still a delay in all vocal processing, coming from the principles of how the technology works (and that delay can be noticed when playing back any music with Whisper), but Whisper tries to prevent it from having too much effect on the composing work.

Hardware support

Whisper does already have support for optional MIDI controllers as an input method, but mostly it just uses basic ASIO-compatible audio interfaces for output purposes (it's not doing anything to the input signals). The idea is to let the VST plugins do what physical instruments or effect modules would do in a physical studio environment (without the physical audio recording part), and use the (PC) hardware instead for all the processing. Whisper could also be modified to support physical instruments with a MIDI interface as pluggable synthesizers, but using that functionality most effectively would probably require the user to have a multichannel audio input interface supporting ASIO.


The reality

Whisper is available "on demand", meaning that you can request for download access to try it.

The application is being developed for Windows only. A possible Max OS X version (while an interesting and useful thing to have) is not going to be implemented very soon, as there are no Mac computers available to develop on (and it would also be too time-consuming in this phase to do any development on two different platforms). Also, some of the plugins used by Whisper development are not available for Macs yet either, and could not be used in Mac OS X even if Whisper itself was available for it.

The future

We want to keep the software simple enough also in the future, and to use only the optimal amount of time and effort on developing it. We might change the licensing and/or distribution methods, but most likely we want to still keep it free (no monetary cost involved for the users). We want to work on it in a restricted scale, and in our own pace. There are so many other applications already fulfilling the needs of many people better than Whisper would, that we don't want to add anything unnecessary to it. There's no need to make the project bigger, as it would also make things more complicated. Whisper is, after all, an application that is designed to fit quite a specific purpose.

Feedback and questions

If you have something to ask about Whisper, or have any opinions on something on the website, you can send email to the given feedback address. The mailbox name is simply "whisper", and the site name is "whispervst.org". And to make it a full email address, you of course have to add the "@" sign in between.