Hi, I’ve processed many narration audio tracks and they sound great, and then normalized to -3 dB to be ACX compliant:
However, when I add the audio to my video editor Camtasia and render the final videos, for some reason Camtasia reduces the dB, and it sounds quieter. And when I import the MP4 video into Audacity to extract the audio and then do an ACX check, the peak is now usually around -4.5 dB:
I could try normalizing the original audio track with a higher peak level, like maybe -1.5 dB, so that when Camtasia reduces it, it ends up close to -3. But this would be a lot of work over 200 tracks/videos, and I am unsure of quality impacts.
Any suggestions? Would just living with the peak of around -4.5 db be acceptable? Or would that be “perceived” to be too quiet, and therefore not as good by viewers? (Many viewers of this content will use headphones to watch the videos.)
Those RMS & peak requirements are ONLY for ACX. Nothing bad happens when you get close to 0dB, only if you go over, or “try” to go over in integer formats.
As you may know, MP4 (like MP3) is lossy compression and things can change. Usually the RMS stays about the same but the peaks can change. But I think Camtasia is doing something unexpected.
There are various standards for streaming platforms and broadcast. Most use an LUFS standard rather than RMS. And, all of the popular streaming platforms use loudness normalization so they will adjust your loudness up or down, except they won’t push the peaks into clipping so some programs can be quieter than the standard (and quieter than “everybody else”).
That and even though you passed ACX-Check, you’re still a bit off.
Your first example has perfect peaks, but the RMS (loudness) is just about to fail too loud. The loudness numbers are -18dB (loud) to -23dB (quiet). Audiobook Mastering pushes your loudness to a mid point of -20dB (just right) and then trims your peaks to -3.5dB.
I suspect if you did that, it would be slightly quieter, but everything would match and the video post processing may leave you alone.
Thanks @DVDdoug and @kozikowski . This changes my approach. The videos are going on Udemy, the big training course site, where most videos are much “louder” than mine. And as has been noted, louder is often perceived to be better quality. I have had comments mine were too quiet.
I thought -3 dB was a universal limit. However, having done more research as a result of your answers, I now understand that up to 0 dB is fine. So I want my videos to be significantly closer to 0 dB, to be perceived as comparable to the others on the training site.
The problem is that Camtasia is a wildcard. I have turned off all auto normalization options, but it is still greatly reducing the dB. Increasing the dB of the input track to get a louder final result would be chasing my tail. So I had an idea. I have simply doubled the audio track, so it renders in the video twice over. The result is much “louder”:
However, there are a few excessive peaks in this final result, e.g. see this excerpt showing the worst parts in the entire video:
Now, I cannot apply a limiter to this and put it back in Camtasia, because the software will just significantly reduce the whole thing again. So I am thinking of just living with it. My question then is: how bad is this? It sounds ok to me, but I know I don’t have a very good ear. Here is the excerpt:
What do you think, is this acceptable, or does it hurt the ears, or is it somewhere in between?
That’s not going to pass ACX Check. It’s significantly louder than an audiobook presentation. I picked peaks at -.25dB rather than 0dB (-0.0dB) because if you’re too close to overload, post production processing can cause ticking and other distortions.
Note the limiter is Soft Limit.
That processing should sound pretty much like you except louder.
Do they have any preferences for how to make the MP3 submission? ACX has some very firm ideas about that. I created a good quality audiobook-style MP3 from my tests and it passes very well.
Does Camtasia have a help forum? It might be a mono-stereo thing. Or, there are other video editors.
ACX check is showing peaks just-barely over 0dB. It’s highly-unlikely that you’ll hear any distortion. I wouldn’t worry about it. It’s not like you’re trying to make a high-fidelity music production.
Also, MP4 can go over 0dB without clipping. Audacity (and ACX check) is showing potential clipping. The viewer/listener will slightly clip their digital-to-analog converter if they play it at “full digital volume”.
Turns out there is a standard. -16LUFS. But they don’t follow up with a limiter to take care of peak distortion. They depend on the performance not going into overload by itself after Normalization.
The poster points out the wide variation of loudness in the field.
Hi @kozikowski , the key factor here is that the final product is an MP4 video, not an audio file. So it does not have to be ACX compliant. I was wrong about that. So I have a lot more flexibility. It only needs to not be “bad”.
I could clean the audio as you note, however once I put it back in the video editor and render the final video, my video editor software Camtasia would reduce the audio significantly again, putting me back where I started. I would be chasing my tail.
Now if Audacity could take a video as input, process the audio, and then export the video and cleaned audio together, that would be awesome. But of course Audacity cannot work with audio and video together.
Which leads me to my proposed solution. I cleaned up the audio, to ACX standards or pretty close, and then rendered the video with two copies of the audio track, to make it “louder”. This then is an extract of the audio from the final video produced this way, with some of the worst clips:
It sure is loud. And it has a very small number of peaks above 0 dB. An example of the worst is in the excerpt. However it is fine it is not ACX compliant, for this video use case. My main question is: is it ok to the ear? To me, it sounds ok. If you were watching a 20 minute video with this as the audio, would it be ok, or would the small number of above 0 dB portions be too annoying?
Audacity can open a video file and extract the audio track. But it won’t put it back. That’s where you need your video editor.
I’m not expecting the presentation to meet ACX when the goal is podcast or other application. That’s why I posted the revised tools above.
Post some of the Camtasia video after the volume reduction. You can’t go over 2MB in the forum. If it won’t let you post the video, open it in Audacity and export a WAV (Microsoft) of the sound track.
That’s where I’d be looking for answers. How did the other contributers do it?
Way in the back of all these exercises is the requirement of Fewest Steps Possible. It’s nice that you applied a bunch of effects to your presentation, but you’ll need to apply them forever and remember which ones you used.
Switching to another video editor is not doable, since there are about 80 videos, for 8.5 hours total, with more than 1,400 images and animations. Moving all that to another video editor would be months of work.
I processed the original audio with Audacity a few years ago, and you gave me a lot of assistance. I have chains set up that capture all the work to make it as simple and repeatable as possible.
However, now I want to simply make the videos “louder”, comparable with what I realize is the standard on the site with other videos. There is no standard approach from the other Udemy instructors. The whole point of Udemy is no gatekeeping or standards, everyone creates however they wish, and then the student ratings decide what rises to the top.
The video editor dropping db about -1.5 is really annoying. But it is what it is. Thus my attempted solution to render with two copies of the audio track. I really appreciate your comment “It’s highly-unlikely that you’ll hear any distortion. I wouldn’t worry about it.” If that excerpt I posted does not sound terrible to you, it is more than good enough for me. I’ll go with that.
If they accept it without the correction, I’d be tempted to go with it as is.
There is a YouTube explainer video on Camtasia Studio Video
They seem to think it’s possible to adjust the Camtasia audio volume.
Camtasia 2019 audio leveling The audio leveling is used to try to match your sound by raising all the wave-forms in your sound. The volume leveling effect will raise the volume of your entire audio based on the settings you choose.
Camtasia 2020 audio leveling The volume leveling in 2020 is the same as for version 2019. It works similarly.
Hi @kozikowski , yes, Camtasia has a volume leveling function, but it is very simple. Nothing compared to Audacity.
Yes, lowering 1.5 db might not seem like a lot, but my goal is to raise the dB. So Camtasia for some reason has decided to go in the opposite direction to what I want.
For comparison, here is an excerpt from the final video when my original ACX compliant audio track is rendered, too quiet for my ears, and much quieter than many other videos on Udemy.
And here again is the result from the video when the same track is doubled:
I like the second one much better. And if you say there is no noticeable distortion, then it should be safe to go with the second?
