If you came here just to see the video, here you go:
Delivered by Dr Nabilah. If you enjoy these videos, do show your support by subscribing to this channel.
But if you want to know more about the process I used and the relative difficulty in producing this, read on. As regards to the overall process and what I’ve learned so far (and what I got wrong) will be disclosed in another post. This will generally only talk on the final settings and edits.
To start, this is the audio source I’ve been working with:
Not great at all. The loudspeaker really effed it up to put it mildly, especially within the confines of the small CME room, and made the speaker’s voice sound not too natural. Added to that was the problem of feedback. Recording using just voice memos on an iPhone wasn’t a problem in general. It was just the acoustics of the source.
But the audio was certainly clipping as seen in the levels here:
Levels were quite high with clipping in the beginning, then lower down (as the loudspeaker stopped working) with no more clipping
In the end the audio was broken up into a few parts that I edited separately:
The early part of the talk where the speaker was talking with the mic
The latter part of the talk where the audio actually improved once the loudspeaker stopped working
Significant audio from the audience
Only the second part was easy to edit. There was some hum at the low frequencies which I attenuated, and left the rest untouched via channel EQ on Final Cut Pro X. As regards to compression, I used one of the presets then turned on the limiter at -6 dB and further reduced the output to -6dB, getting the track down to peak at -12 dB.
Settings (EQ) of the 2nd part of the audio track
Settings (compressor) of the 2nd part of the audio track
As regards to the first part, the editing was much more difficult. I’m not sure how much more improvement I could get, but I got approval from my audiologist wife (who has some interest in audio engineering but hates music), who definitely noticed the improved clarity versus the original track. I think it’s pretty clear that despite edits, the first part (at least to me) won’t sound better than the second part. But at least it’s now clear.
With the first part, editing the EQ included attenuating both the low and high frequencies, in contrast to the second part which only required attenuation of the low frequencies to remove the hum. To my ears the loudspeaker made the voice sound a bit tinny when the hum was attenuated; hence the need for the high frequencies to be attenuated too.
Settings (EQ) of the 1st part of the audio track
Settings (compressor) of the 1st part of the audio track
I wasn’t so exacting with the 3rd part; you can only do so much from audio far away and at the opposite trend in audio levels, too low. Bringing the audience’s voices up also brought along noise. So I only edited those that were personally deemed more important in the flow of the presentation or Q&A. In the edits I did feel that the voices were heard “just enough” to make out the flow.
Some low frequency attenuation to remove the hum, and more pronounced attenuation at the higher frequencies to reduce the tinniness of the audience’s voices once the levels were brought up. And since the audience’s voices were low to begin with, there was a lot more attenuation needed once the output signal was pushed up.
Corresponding compressor settings
I do feel that the edits done were meaningful and brought the levels close together. You may have to turn up the speakers a bit, but that is much better than having the audio signals clip and distort. See how the levels more or less stay the same after the edits.
I felt that the amount of feedback wasn’t too bad in the end, and I needed to look up how to deal with it in a free manner anyways; I didn’t feel that the effort needed to fix it would be worth it at this time. One interesting thing was that I used the same compressor settings throughout the whole video, with only changes in the output gain to keep the audience’s voices louder & peak at -12 dB.
As regards to the video component, this presentation had a greater need for pointers to make the presentation clear. Me being a novice decided using simple arrows will do. This presentation also had some ultrasound videos, which I decided to be displayed full screen. I’ve no idea if color correction is needed as I used just the presentation slides to make the video (as my laptop is nearing the end of its “useful life”).
Do tell me how you feel about the differences in the voices after the edits. After all, this is all self-learned (but certainly an interest of mine).