{"id":35676,"date":"2026-04-05T11:03:49","date_gmt":"2026-04-05T18:03:49","guid":{"rendered":"https:\/\/www.podfeet.com\/blog\/?p=35676"},"modified":"2026-04-05T11:26:46","modified_gmt":"2026-04-05T18:26:46","slug":"auphonic-noise-filler-words","status":"publish","type":"post","link":"https:\/\/www.podfeet.com\/blog\/2026\/04\/auphonic-noise-filler-words\/","title":{"rendered":"Learning About Auphonic Cuts and Noise Suppression"},"content":{"rendered":"<p>A few years ago, I wrote a fairly lengthy explanation of how I use the awesome web service called Auphonic to sweeten the audio for the podcast. Everything you hear has gone through Auphonic. I put a <a href=\"https:\/\/www.podfeet.com\/blog\/2023\/05\/auphonic\/\">link in the show notes to the article<\/a> where I explained how Auphonic levels the audio between speakers, raises the audio to the official loudness standards (-16 LUFS), removes noise, maintains chapters, FTPs the audio to the server, and even creates the transcripts you get with the shows.<\/p>\n<p>While I was confident my settings were everything they should be, I started to notice things not sounding quite as expected. For example&#8230;<\/p>\n<p>During the recent episode where I talked about how much fun it was to be in a rehab hospital, I described the giant clunk sound the light made when you changed modes. I purposely recorded that awful sound while I was in the hospital so I could play it on the podcast. However, after running it through Auphonic, you never heard the sound at all. You could hear me set it up, but then there was just complete silence after that.<\/p>\n<p>Then remember when I had to change the sound that I play when I&#8217;m panhandling for donations? It was originally the sound of a cash register, but I discovered that you were never able to hear that either. I changed the sound to some music, and now you can hear it.<\/p>\n<p>In a very recent episode, I was interviewing the woman from Levoit, and at one point, she demonstrated the feature where the vacuum&#8217;s chamber gets sucked clean by its base station. I realized later that you never heard that sound.<\/p>\n<p>You have probably guessed that I may have my noise suppression settings in Auphonic set a little too aggressively.<\/p>\n<h2>Ask Mastodon<\/h2>\n<p>As one does, I immediately turned to Mastodon to ask for help. A lovely gentleman named <a href=\"https:\/\/mastodon.social\/@LoneLocust\">Eugene Glover<\/a> host of the <a href=\"https:\/\/fusionpatrol.com\">Fusion Patrol podcast<\/a>, jumped in to try to help. It was pretty clear that he was exercising Auphonic to a higher level than I was. While he didn&#8217;t pinpoint the solution to my problem, he taught me two incredibly useful things about using Auphonic.<\/p>\n<p>Because I like to drag out a story, I&#8217;m going to start with the solutions he showed me that didn&#8217;t directly fix my problem, but allowed me to fix the underlying problem and problems I didn&#8217;t realize I could solve. Let&#8217;s talk a little bit about the structure of Auphonic so I can explain.<\/p>\n<h2>Editing Productions<\/h2>\n<p>In Auphonic, you can create presets that define things like:<\/p>\n<ul>\n<li>Name of the show<\/li>\n<li>Artwork<\/li>\n<li>What kind of output files you require<\/li>\n<li>Where you want the output files to go<\/li>\n<li>The profile for things like noise reduction and loudness levels<\/li>\n<\/ul>\n<p>I have saved presets for the NosillaCast, Chit Chat Across the Pond, and Programming by Stealth, and they make it really easy for me to publish the show.<\/p>\n<p>Auphonic also has what&#8217;s called a Production. This is where you actually produce an episode of the show. You can start with a preset, which, of course, I do. In that Production, I can name the particular episode, and if I wanted to, I could tweak some of the settings, but since I&#8217;ve gone to all that work to create the presets, I usually don&#8217;t. I simply hit Start Production. Auphonic first ingests my recording, and then it goes to work while I do other things. Since my shows have been getting longer, it can take maybe 10 minutes.<\/p>\n<h2>Completed Production<\/h2>\n<p>When Auphonic is finished processing the audio, it brings up a window for the completed Production. This includes download links for the various files I&#8217;ve had it create, like the MP3 file and the HTML for the transcripts. Below that, it shows the waveform for the output file and the waveform for the input file.<\/p>\n<p>It&#8217;s fun to look at these waveforms because you can see the work that it has done to improve the audio. For example, if I&#8217;m recording with Bart, I can see that even though our levels are fairly close together when we record, they&#8217;re not exactly the same, and the levels tend to be low. After the production, the leveling has occurred, so we&#8217;re at the same volume, and the volume has been raised to the loudness standards.<\/p>\n<p>The first thing Eugene on Mastodon suggested was that I look at the finished production for the area where the sound had disappeared and edit the cuts. He sent me some screenshots where there were vertical gray bars on his waveform representing cuts that had been made to his audio. But I didn&#8217;t see anything like that in my productions.<\/p>\n<p>Eugene and I went back and forth a bit, and then I decided to contact the developer of Auphonic, a lovely gentleman named Georg. I learned a lot in talking to Georg, and the most important thing I learned is that <a href=\"https:\/\/auphonic.com\/help\/index.html\">he has documented the heck out of all of the features in Auphonic<\/a>. I&#8217;m 98% certain that I never even looked to see if there was documentation!<\/p>\n<h2>Cut Filler Words<\/h2>\n<p>Between Eugene and Georg, I discovered that hiding in plain sight was an option for Auphonic to automatically cut certain things out of your audio production. The section is called Automatic Cutting, and it has four options:<\/p>\n<ul>\n<li>cut silence<\/li>\n<li>cut fillers<\/li>\n<li>cut coughs<\/li>\n<li>cut music<\/li>\n<\/ul>\n<p>The one that caught my eye was &#8220;cut fillers&#8221;. <a href=\"https:\/\/auphonic.com\/help\/algorithms\/singletrack.html#automatic-cutting\">Georg&#8217;s documentation<\/a> explained:<\/p>\n<blockquote><p>\n  We have trained our filler word cutting algorithm to remove filler words, namely any kind of \u201cuhm\u201d, \u201cuh\u201d, \u201cmh\u201d, \u201c\u00e4hm\u201d, \u201ceuh\u201d, \u201ceh\u201d, and similar with English, German, and Romance language data. However, the algorithm already works quite well for many similar-sounding languages, and we will also train more languages if necessary.\n<\/p><\/blockquote>\n<figure style=\"float: center; margin: 10px\"><img decoding=\"async\" src=\"https:\/\/www.podfeet.com\/blog\/wp-content\/uploads\/2026\/04\/auphonic-automatic-cutting-cut-fillers.png\" alt=\"Auphonic automatic cutting with an arrow pointing at cut fillers.\"  title=\"auphonic automatic cutting - cut fillers.png\" width=\"799 \" height=\"543\"><figcaption style=\"text-align:center\">Cut Filler Words<\/figcaption><\/figure>\n<p>On a recent recording, my guest and I had a <em>lot<\/em> of ums and uhs, so I gave it a try, and it worked quite well. When I looked at the output waveform, I could now see the grey cut bars Eugene was talking about. Eugene says he checks them all because occasionally other words can be affected &#8230; but I was too lazy to check mine. Eugene explained that after he edits the cuts from the completed production, he reruns the production.<\/p>\n<p>While being able to edit a production after the fact to fix problems sounded swell, it prompted another question to Georg.  Auphonic is billed based on how many hours of audio recordings you upload. You can upload 2 hours a month for free, or they have plans starting at $11\/month if you pay annually.  My concern was that if I edited one of my shows like this after the fact, I&#8217;d use up another hour or more of my monthly credits.<\/p>\n<p>Again, Georg had a great answer. You&#8217;re only charged for the first upload. Every edit and rerun of the same production is free.<\/p>\n<p>Ok, cool. Now I can cut ums and uhs automatically, but that wasn&#8217;t the problem I set out to solve.  I was having trouble with intentional noise being removed from my recordings. I use Georg&#8217;s denoising algorithm, but it appeared I set it to be a bit heavy-handed. Now that I knew I could do experiments at no extra cost, I could figure out how to dial it in the way I wanted it.<\/p>\n<p>Another reason dialing back the noise reduction was an attractive idea is that with my current settings, the audio is so clean you can&#8217;t tell when I&#8217;m on a conference floor interviewing someone. Literally all of the surrounding crowd noise is gone. Steve in particular suggested that it&#8217;s more authentic if you can hear the rumble of the crowds, at least a little bit, when we&#8217;re on a show floor doing an interview.<\/p>\n<h2>Levoit Vacuum Noise<\/h2>\n<p>Now that I knew I could run productions repeatedly with the same audio, I was inspired to do some tests. The Levoit interview would be an ideal test for modifying the noise suppression settings in Auphonic. I could isolate the short section of the vacuum noise, so that in my tests, I would use up very little credit on my account.  It was also ideal because it not only had this specific sound I wanted you to be able to hear, but it also had conference noise. If I reduced the noise suppression too much, I could end up with the intentional noise coming through, but also make it harder for you to hear conference interviews.<\/p>\n<p>The Auphonic interface for a production has a drop-down for the level of noise suppression.  Full suppression is 100 dB, high is 24 dB, medium is 12 dB, and low is 6 dB.  The drop-down includes more options between those levels.  I had all of my presets set to 30 dB, which is above high.<\/p>\n<figure style=\"float: center; margin: 10px\"><img decoding=\"async\" src=\"https:\/\/www.podfeet.com\/blog\/wp-content\/uploads\/2026\/04\/Remove-noise-levels-in-db.png\" alt=\"Remove noise levels in db, described in the text above\"  title=\"Remove noise levels in db.png\" width=\"161 \" height=\"314\"><figcaption style=\"text-align:center\">Noise Suppression Levels<\/figcaption><\/figure>\n<h2>Experiments<\/h2>\n<p>I ran my first test at 30 dB of noise suppression, which is where I&#8217;ve had it set for quite some time. I wanted to see what the output waveform looked and sounded like. Then I tested it at 12 dB (medium) and finally at 6 dB (low).  Unfortunately, I can&#8217;t play each of those different files, because this recording will have its own noise suppression applied.  I did put the three 30-second audio clips in the blog post so you can listen to them there, and I also included screenshots where you can see visually how aggressive that 30 dB of noise suppression actually was.<\/p>\n<p>In the screenshots, you see the Output audio above the Input audio. In the 30 dB noise suppression version, the Input clearly shows the two times the vacuum was turned on. In between these peaks, you can see the rumble of the crowd. In the Output, it\u2019s nearly flatlined when the vacuum goes off, and there\u2019s no rumble of the crowd at all.<\/p>\n<figure style=\"float: center; margin: 10px\"><img decoding=\"async\" src=\"https:\/\/www.podfeet.com\/blog\/wp-content\/uploads\/2026\/04\/Levoit-jf-Defaults-30db-24db-is-high-noise-annotated.png\" alt=\"Levoit 30db (24db is high) noise annotated - as described\"  title=\"Levoit-jf-Defaults-30db-(24db-is-high)-noise-annotated.png\" width=\"799 \" height=\"404\"><figcaption style=\"text-align:center\">30 dB Noise Suppression Erases Vacuum &#038; Conference Rumble<\/figcaption><\/figure>\n<audio class=\"wp-audio-shortcode\" id=\"audio-35676-1\" preload=\"none\" style=\"width: 100%;\" controls=\"controls\"><source type=\"audio\/mpeg\" src=\"https:\/\/www.podfeet.com\/blog\/wp-content\/uploads\/2026\/04\/Levoit-jf-Defaults-30db-(24db-is-high)-noise.mp3?_=1\" \/><a href=\"https:\/\/www.podfeet.com\/blog\/wp-content\/uploads\/2026\/04\/Levoit-jf-Defaults-30db-(24db-is-high)-noise.mp3\">https:\/\/www.podfeet.com\/blog\/wp-content\/uploads\/2026\/04\/Levoit-jf-Defaults-30db-(24db-is-high)-noise.mp3<\/a><\/audio>\n<p>Next, I dropped the noise suppression to 12 dB or medium. You can see in the screenshot tiny peaks for the vacuum noise and an itty-bitty little amount of noise from the rumble of the room. Listening to it, it wasn&#8217;t yet what I was looking for.<\/p>\n<figure style=\"float: center; margin: 10px\"><img decoding=\"async\" src=\"https:\/\/www.podfeet.com\/blog\/wp-content\/uploads\/2026\/04\/Levoit-12db-med-noise-annotated.png\" alt=\"Levoit 12db (med) noise annotated.\"  title=\"Levoit-12db-(med)-noise-annotated.png\" width=\"799 \" height=\"368\"><figcaption style=\"text-align:center\">At 12 dB We Can Start to Hear the Vacuum But Not the Conference Rumble<\/figcaption><\/figure>\n<audio class=\"wp-audio-shortcode\" id=\"audio-35676-2\" preload=\"none\" style=\"width: 100%;\" controls=\"controls\"><source type=\"audio\/mpeg\" src=\"https:\/\/www.podfeet.com\/blog\/wp-content\/uploads\/2026\/04\/Levoit-jf-Defaults-30db-(24db-is-high)-noise.mp3?_=2\" \/><a href=\"https:\/\/www.podfeet.com\/blog\/wp-content\/uploads\/2026\/04\/Levoit-jf-Defaults-30db-(24db-is-high)-noise.mp3\">https:\/\/www.podfeet.com\/blog\/wp-content\/uploads\/2026\/04\/Levoit-jf-Defaults-30db-(24db-is-high)-noise.mp3<\/a><\/audio>\n<p>Finally, I tried 6 dB, which is considered low noise suppression. With this setting, the background rumble of the show floor is just perfect \u2014 not distracting, but you know where we are. And more importantly, you can hear the vacuum.<\/p>\n<figure style=\"float: center; margin: 10px\"><img decoding=\"async\" src=\"https:\/\/www.podfeet.com\/blog\/wp-content\/uploads\/2026\/04\/Levoit-6db-low-noise-annotated.png\" alt=\"Levoit 6db (low) noise annotated.\"  title=\"Levoit-6db-(low)-noise-annotated.png\" width=\"799 \" height=\"394\"><figcaption style=\"text-align:center\">A Nice Light Rumble and Audible Vacuum at 6 dB<\/figcaption><\/figure>\n<audio class=\"wp-audio-shortcode\" id=\"audio-35676-3\" preload=\"none\" style=\"width: 100%;\" controls=\"controls\"><source type=\"audio\/mpeg\" src=\"https:\/\/www.podfeet.com\/blog\/wp-content\/uploads\/2026\/04\/Levoit-6db-(low)-noise.mp3?_=3\" \/><a href=\"https:\/\/www.podfeet.com\/blog\/wp-content\/uploads\/2026\/04\/Levoit-6db-(low)-noise.mp3\">https:\/\/www.podfeet.com\/blog\/wp-content\/uploads\/2026\/04\/Levoit-6db-(low)-noise.mp3<\/a><\/audio>\n<p>And how about the noise the light made in the rehab hospital? As you may recall, it had three positions that I could change using a long strip attached to the light above my bed. You were robbed of hearing it because I was over-exercising the noise suppression. I now bring you that glorious noise!<\/p>\n<audio class=\"wp-audio-shortcode\" id=\"audio-35676-4\" preload=\"none\" style=\"width: 100%;\" controls=\"controls\"><source type=\"audio\/mpeg\" src=\"https:\/\/www.podfeet.com\/blog\/wp-content\/uploads\/2026\/04\/lights-rehab-6db.mp3?_=4\" \/><a href=\"https:\/\/www.podfeet.com\/blog\/wp-content\/uploads\/2026\/04\/lights-rehab-6db.mp3\">https:\/\/www.podfeet.com\/blog\/wp-content\/uploads\/2026\/04\/lights-rehab-6db.mp3<\/a><\/audio>\n<p>Can you even imagine if you were my roommate in the hospital and you&#8217;d <em>finally<\/em> gotten to sleep and I decided to turn my light off???<\/p>\n<p>The rehab hospital also had kind of a background hiss to it, I assume, formed by all the different machines that were doing things in all the different rooms. You can still hear that in this recording, and maybe that level of noise wouldn&#8217;t always be ideal. Don&#8217;t be surprised if I bump that 6 dB up a little bit at some point, but at least you can hear that glorious thunk now!<\/p>\n<h2>Bottom Line<\/h2>\n<p>The bottom line is at least tri-fold:<\/p>\n<ol>\n<li>I&#8217;m amazed at how much more capable Auphonic is than I realized. Looking at the documentation, I don&#8217;t think I&#8217;ve even scraped the surface.<\/li>\n<li>I&#8217;m pretty darn sure there are several of you in the audience who know a lot about this subject and are going to write me &#8220;Well actually, Allison&#8230;&#8221; messages. I really hope you do because I love to learn!<\/li>\n<li>Many thanks to Eugene Glover for getting me to poke at this and learn so much. He didn&#8217;t have to help me, but people are just nice that way.<\/li>\n<\/ol>\n","protected":false},"excerpt":{"rendered":"<p>A few years ago, I wrote a fairly lengthy explanation of how I use the awesome web service called Auphonic to sweeten the audio for the podcast. Everything you hear has gone through Auphonic. I put a link in the show notes to the article where I explained how Auphonic levels the audio between speakers, [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":35685,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"jetpack_post_was_ever_published":false,"_jetpack_newsletter_access":"","_jetpack_dont_email_post_to_subs":false,"_jetpack_newsletter_tier_id":0,"_jetpack_memberships_contains_paywalled_content":false,"_jetpack_memberships_contains_paid_content":false,"footnotes":""},"categories":[147],"tags":[2058,8070,8071,8072,8069],"class_list":["post-35676","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-blog-posts","tag-auphonic","tag-cutting-filler-words","tag-eugene-glover","tag-masotodn","tag-noise-suppression"],"jetpack_featured_media_url":"https:\/\/www.podfeet.com\/blog\/wp-content\/uploads\/2026\/04\/Levoit-6db-low-noise-annotated-1040x520-1.png","jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/www.podfeet.com\/blog\/wp-json\/wp\/v2\/posts\/35676","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.podfeet.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.podfeet.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.podfeet.com\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.podfeet.com\/blog\/wp-json\/wp\/v2\/comments?post=35676"}],"version-history":[{"count":8,"href":"https:\/\/www.podfeet.com\/blog\/wp-json\/wp\/v2\/posts\/35676\/revisions"}],"predecessor-version":[{"id":35690,"href":"https:\/\/www.podfeet.com\/blog\/wp-json\/wp\/v2\/posts\/35676\/revisions\/35690"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.podfeet.com\/blog\/wp-json\/wp\/v2\/media\/35685"}],"wp:attachment":[{"href":"https:\/\/www.podfeet.com\/blog\/wp-json\/wp\/v2\/media?parent=35676"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.podfeet.com\/blog\/wp-json\/wp\/v2\/categories?post=35676"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.podfeet.com\/blog\/wp-json\/wp\/v2\/tags?post=35676"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}