Processing Field Recordings Using Absentia DX and Izotope RX

Absentia DX can be a valuable asset in the cleaning and processing of field recordings. I use it with Izotope RX software to enhance the results.

Last updated: 30-July-2021

pexels-photo-414171.jpeg

This post is left here for archival purposes, but please be aware of these changes I’ve made to my processing: With the improvement in Izotope RX version 8 Standard, I’ve been able to eliminate the use of Absentia DX. Also Absentia DX has changed into a subscription service. While it is a good program, it now has little benefit for field recordings in my opinion.

File Structure

I keep a file structure for each of the field recordings that allow me to see what they are. The same file folders are in each field recording folder.

Field Files
  • ADX – This is where I put the file after being processed by Absentia DX.
  • ADX_RX –  This is the file after I have processed it in both Absentia DX and Izotope RX.
  • FINISHED – This is the finished file that is ready for distribution.
  • RAW – This is the file coming straight out of the recorder without processing.

Absentia DX

Izotope RX is my primary software used for cleaning of field recordings. Absentia DX is a separate program that is designed to clean up the dialog of a recording and is not meant for field recordings. However, I have found it useful in cleaning field recordings as well.

I use Absentia DX first in my workflow. I drag and drop a wave file onto the software and let it process it. I may do this with different settings so that I am able to pick and choose afterward when I edit the file in Izotope RX.

The settings I have tried and find to work on outdoor atmospheric-type (natural soundscape) sound files are shown below. I often check the “Broadband Reduce” box also, as I have found this quite helpful.

adx1

Absentia removes noises that Izotope RX will not. I save the resulting files with a “_adx” suffix.

Izotope RX

I then use Izotope RX:

  1. In the de-hum module, I use a preset called “Hearing A” that applies an A-weighted curve to the sound that replicates what humans can hear.
  2. I cut out the center of the recording to keep as it does not have the noises that the beginning and ending often do. I use the trim function (control-t) after listening to the track to find the noises unwanted in the beginning and ending.
  3. There are always clicks. Surprisingly, I cannot always hear them, but they’re there. I do the de-clicking using the de-click module in RX and choosing (in order):
    1. random clicks,
    2. random thumps, and
    3. short digital clicks.
  4. Next, I de-clip the peaks if any. Usually, there are not any, but some loud noises you want to keep such as thunderclaps and these will clip sometimes and need to be de-clipped using the RX de-clip module.
  5. Fade in and out the first and last ten seconds of audio. I use
    1. “fade in smoothly” and
    2. “fade out (end of song)” settings.
  6. Go in and trim out noises, or cut the clean sound and paste it over noises (preferred).
  7. Normalizing
    1. -6 dB for field recordings makes for a nice, realistic sound at this point. Realism is what I am going for in this kind of recording. I want the listener to feel like they were actually there.
      • I will often find subtle sounds come out such as footsteps and door slammings that need to be trimmed off of the front and back ends of the recordings and this is where I go back and trim and re-apply the fades to beginnings and ends if needed.
  8. (Optional) Experiment with this: Consider re-applying the de-hum module settings for hearing A-weighting. This re-application of the de-hum cleans up further the sounds that were hidden before normalizing. After normalizing, these sounds stand out and need to be taken down using the hearing A-weighting profile. I sometimes find this especially useful for field recordings where too much of the low end gives an artificial humming (for lack of a better word) or “closeness” feeling. By getting these sounds out, it helps with the overall perception of the sound and adds to the realism. The slight boost I have in the upper end also gives more “air” to the recording, further enhancing it.
  9. Now the file will sometimes have high and low amplitude parts. There may be a high part and the normalization has made that the -6 dB point (for the -6db normalization), but the rest is lower and only goes to -12dB for example. I then need to apply gain. I do this in RX by enabling the “view clip gain” control and then raising the gain gradually in these different areas. I do not want abrupt changes in the gain so I put in a lot of points and ramp carefully the gain. In the beginning and endings, I leave the gain at none. In some cases, I need to lower the gain as well in order to keep the overall within a good range.
  10. Finally, I re-apply the normalization to -3dB. This brings it all back up to a listenable level, while all of the low and unwanted noises have been filtered out. Now it is not only realistic but also listenable.

Finishing

Then I save the processed file as a wave file with a “_rx” suffix. I now have a wave file with adx_rx on the end of it so I can see how I processed it.

I export this to FLAC with compression of 5 and a dither setting called “noise shaping.”

This file should then be done and ready for use.

I will sometimes extend the length of the file by repeating it. This will allow me to make a 10-minute recording into a 30-minute soundscape. I do this using a technique in Audacity: Processing Field Recordings Using Audacity

One thought on “Processing Field Recordings Using Absentia DX and Izotope RX

  1. Pingback: Field Recording | The Audio Blog of Lars Lentz

Comments are closed.