Voxengo Logo
On-line links: Products - Downloads - Purchase - Contacts - Forums
Voxengo Deconvolver HELP

Voxengo Deconvolver Screenshot



Contents

Introduction
Impulse capture method
Explanations on the GUI
Troubleshooting tips and additional information



Introduction

Recently, sampling (convolution) reverbs have become more and more in demand. With convolution, we have an opportunity to capture the sound of anything in the world that can generate a reverb and use these sound impulses freely in any situation imaginable. This enables us to use the sound of high-end reverb units, real-world rooms, halls, cathedrals, synthetic reverbs and other sources, including non-reverb ones, without any hassle and in a uniform way using only a single program or a plug-in module.

Although there are many different sources of impulse responses, we also face the difficulties of acquiring these so they can be used seamlessly in any software environment.

In many cases during some stage of the impulse capture, we typically have a rather large set of recorded test tones that were run through some device or mic'ed in some room. This poses the difficulty of recovering the impulses conveniently and with minimal user effort. The other problem we may face is the input or the output bit-depth incompatibility of the recorded and the recovered files. Some convolution plug-ins tend to support only a small subset of available bit-depths. And alike, existing deconvolution programs and plug-ins support only the given sample rates and bit depths, and tend to offer a very poor quality deconvolution.

Voxengo Deconvolver overcomes these problems. First of all, it supports almost all sample formats (bit-depths) of uncompressed mono/stereo WAV files. Secondly, it offers a very convenient environment in which to deconvolve large sets of recorded files. Voxengo Deconvolver also offers a true mathematical FFT deconvolution which delivers 100% exact deconvolution. At the same time, this puts a huge demand on the system memory: deconvolving a 25-second stereo file at 96 kHz may require up to 100 MB of memory.

Voxengo Deconvolver features:

  • True FFT deconvolution
  • Reversed test tone deconvolution technique
  • Minimum-phase transform option
  • Reads 8, 16, 24, 32, 64 bit PCM and IEEE WAV files
  • Writes 8, 16, 24 PCM and 32 IEEE WAV files
  • Multi-channel file support
  • Batch support
  • Built-in DC removal filter
  • Built-in test tone generator
  • Automatic stereo normalization
  • 64-bit processing
  • All sample rates supported


  • Impulse capture method

    Basically, there are only three things necessary to perform the capture of almost any impulse sound source, including rooms and hardware reverb units.

    1. The ability to playback the test tone through or within the impulse source. Making the room or field recording, you will need a speaker connected to a playback audiocard or a CD player to perform the playback of the test tone. When capturing a hardware unit, you will need to connect the hardware unit's inputs directly to a playback audiocard or a CD player.

    2. The ability to record the test tones which have passed through or within the impulse source. Again, making the room or field recordings, you will need a microphone connected to a recording audiocard or a field recording system. When capturing a hardware unit, its outputs should be connected directly to a recording audiocard or a field recording system.

    3. The ability to perform deconvolution of the recorded test tone. For deconvolution to work, you should record the full test tone duration without any cutouts. For reverberant impulse sources, you should record additional silence which should be at least as long as the expected reverb tail. When capturing hardware units additional silence should also be recorded as unit's impulse response can be lengthy. The recorded test tone should not be distorted or overloaded/clipped. You should pay attention to the playback and recording devices you use. They should exhibit a maximally linear and flat frequency response, and should have a good signal-to-noise ratio. Another possible requirement is that both playback and recording devices should be wordclock-synchronized.

    Nothing more should be done--other than the above-mentioned things--to create an impulse response. Later, after performing the deconvolution, you may need to edit the resulting impulses to fit your needs. For example, you may need to cut the leading and/or trailing silence. Also, you may need to add fade-ins and/or fade-outs. Voxengo Deconvolver does not require the recording to be "in sync" with the test tone - you may add as much pre- and post-silence as you need.



    Explanations on the GUI

    Test Tone File. Here you can specify any WAV file that contains the test tone used during capture. This can be a mono or stereo file with any sample rate.

    File Folder. This is the name of the folder containing the plain (non-deconvolved) recorded impulses. To select a folder, just select any file in it. In parallel, deconvolver will put the selected file into the file list which you can immediately process without first pressing the "Scan folders" button.

    Files to process list box lists all the files which shall be processed when you press the "Process" (or the "Invert") button. You can use the Del key to remove list entries. Several files can be selected for deletion. The Enter key (or mouse double click) can be used to open the currently selected file.

    Scan folders button scans the specified folder for WAV files and adds them into the file list. Before scanning starts, the file list is cleared. You can also use the "Clear list" button to clear the file list manually.

    Out bit depth selects which bit-depth resulting deconvolved impulses will have.

    Volume dB specifies the gain which should be applied to the output impulse file.

    Silence (sec) selects amount of the silence added before the start and after the end of the recorded file. Generally, this must be 0, but in some cases you might want to change this to any other value.

    Reversed technique switch enables an alternative method of performing deconvolution. This method works only for responses captured using sine sweep test tones created with Voxengo Deconvolver. Responses captured with other types of test tones may not work at all. This mode in some cases gives deconvolution of a better quality compared to a standard deconvolution method Voxengo Deconvolver uses. This is especially true with low signal-to-noise (SNR) ratio recordings such as room and field recordings. This is also true for hardware units with a limited frequency bandwidth and SNR. Please note that test tones created with fade-in and fade-out work best with the reversed test tone technique.

    MP Transform enables minimum-phase transform that takes place after deconvolution. Sometimes when you capture a non-linear equipment like speakers and amplifiers enabling MP transform will create much more realistic impulses, without pre-echo. This option can be also used with reverbs. In the end, you will get a perfectly timed reverb with zero initial delay and without pre-echo. However, this is not suitable if the left and right channels of the reverb impulse have different initial spike timings.

    Normalize to -0.3 dBFS switch enables automatic normalization of created impulses to -0.3 dBFS level.

    Low Cut, High Cut options allow you to apply low- and high-pass filters of the specified slope to the resulting impulse file.

    About button brings program's version and registration information.

    Test Tone Gen button brings test tone generator's dialog box.

    Process button starts processing of the files listed in the file list. Any file that could not be processed will be listed again after the batch finishes. This allows you to check/edit any such file in the default WAV file editor and continue the batch processing later.

    Output folder specifies the folder where output files should be created. This field is filled automatically after each new input file folder is selected. Please note that if you have enabled the "Include subfolders" option files in these subfolders will all be exported to the Output folder preserving the folder structure.

    Do not add "dc" suffix switch suppresses appending of the "dc" suffix to the output filename. Please note that this may overwrite the original file. Use this option with care!

    Ignore already processed files switch enables skipping of already processed files. Such files are identified by Deconvolver via special marker which is created for each output file which passes processing stage.



    Troubleshooting tips and additional information.


    Why do I get empty files?

    In most cases Deconvolver creates empty WAV files because the captured impulse file's tail is too short after the test tone stops playing. Make sure you record enough silence after the test tone ends. Ideally, the duration of this trailing silence should be 1.5 to 2 times more the expected reverb tail's length. For a pre-amp or other hardware unit you may additionally record at least 1 second silence.


    I tried to convolve the tube sound algorithm of a favourite VST eq I have without actually wanting its eq curves. Just the warmth. All I achieved was a phasey sound with delay (very good indeed but not what I had in mind!)

    I guess you are trying to capture IR from the source which is not suitable for capturing. Convolution (and hence Deconvolver) cannot work with 'tube' or alike sound sources. It only captures EQ and phase coloration. Tube distortion leads to a wrong impulse capture results.


    When using the Test Tone generator should 3 seconds always be used? Or under what conditions would it be better to use 1 second or 12 seconds?

    Answer depends on the audio device 'quality'. If it's good quality (low noise, good frequency response) 3 seconds should be enough - if it's a spring reverb, for example, you should use a longer test tone. I personally tend to use 6 second test tones.


    I've read on forum that impulse files above 16 bit are not worth it, but the default bit rate for the test tones are 24 bit. So if I use a 24 bit test tone must I also make the impulse 24 bit or is it ok to use a 24 bit test tone to produce a 16 bit impulse?

    You may mix 24-bits and 16-bits in any proportion. The end result will have the lowest bit resolution in the chain. For example, if audio device you are capturing is 8 bit, your final impulse will be 8 bit, too.


    Should I always Normalize to -0.3 dBFS level or are there some situations that I should not? Does normalizing lower the quality of the impulse in any way?

    Normalization is useful for editing the impulse manually (adding fades, cutting). Normalization should not be applied when you are building an array of impulses, because otherwise gain level differences between the impulses will be lost.


    When playing your test tone through a hardware reverb unit should I try to play the file at the highest possible volume that will not clip the hardware unit? Or does it not matter? I'm trying to get the highest quality impulses.

    It does matter whether you playback at the high volume or at the low volume. It's better to play at the highest volume to preserve bit depth. At the same time, some tests revealed that by using a longer sine sweep bit depth is preserved as well. It sounds pretty unbelievable, but it is possible to recover 16-bit impulse using a 8-bit soundcard just by using a suitably long test tone (however, the quality improvement is not huge considering you'll have to use a much longer test tone).



    Happy Impulse Capturing!



    Copyright © 2003-2007 Aleksey Vaneev

    All trademarks belong to their respective owners.