This is very helpful. I've heard comb filtering "explained" and told to "listen for it," but never heard it so clearly demonstrated. Now I actually know what to listen for. Thank you!
Watching this two years later - and I have to say this is outstanding. I've never had this explained properly before. Its now something I won't forget in a hurry.
Excellent job explaining the impact of reflected sound in a room! I am an audiophile building a new family room that is being acoustically treated. The folks interested in my hobby talk about the impact of room acoustics on your stereo and your video clearly illustrates the audio impact of reflected sound. Thank you!👍
Spot on. Great explanation. The very best modern era example of comb filtering is nearly ALL podcasts and audiobooks. There are very, VERY few recorded today free of it. It's so pervasive (even from professional producers), that most listener's ears have accepted it as normal. These recordings are so bad, it's very difficult for anyone that works with audio to listen to them. ESPECIALLY for an extended period of time. The next time you snuggle in with a good podcast, or audiobook and think to yourself..."Gee that sounds bad," but don't know why...that's comb filtering at work.
I could hear comb filtering on the radio after pressing the Mono button. When a sound card plays back, it usually introduces a 1 sample delay between channels. Programmes on repeat were passed through another generation of DAC at the radio station, resulting in 2 samples or about 40 µs of delay. When summed to mono, there was a noticeable loss of treble. If you mix EQ'd and dry signal together, you get more cancellation because the minimum phase EQ introduces a group delay around the steep lowpass filter. I think a shallow filter is more realistic. Your sweep generator has aliasing and and there is a short repeated buffer at the end of the clip.
I am a bass player & have always wanted to run two bass cabinets on either side of the drummer, thinking that my bass sound would spread more 'evenly'. I never did that because I was told about comb filtering. Good thing I didn't~ my bass sound 50 feet out from the stage would have been a big mess! Now I understand why~ Thank you
I'd have to say in this case that theory and practice may differ because there will be so many reflections in the room that any comb filtering may be masked. The only way you'll really know is to try it! DM
I was suprised by that spectragraph software (plugin?) and the lumpy effect at the bottom end of the sweep, didnt the writer do any tests?, really instills confidence in this being "Pro Tools" in function rather than just in name. As for the comb filtered sound, sounds like listening to the end of a decent lenth of poly-pipe.
i've heard at the end of the sweet, sound that were like ending on a frequency and then other sound starting at lower frequency and continuing where the other was ending, in some king of an overlap, excuse my English, this is the best i know
Fantastic explanation, great video. Now I'm curious, about something. Around 12:30 when you added 5ms(?) delay, you commented and said it sounded "dreadful". This sound, I call sewer pipe, as it sounds like you are in a tube. I hate satellite radio and highly compressed mp3's because of something similar. I wasn't expecting comb filtering to be an artifact of satellite radio. Is it similar or possible? Why do overly compressed audio files sound like they are in a tunnel like this, is it the same thing?
I think these might be different distortions that possibly sound similar. People's subjective experiences may differ. This is a topic I may look into further in future. DM
That sound is so cool. It sounds like its inside ur head. Is that binural panning? I use it often but i havent ever heard it like that before. I habe ever listened to it in this manner either. I havent watched the rest of the video yet, but just wanted to say that was pretty cool!
That distortion may be caused by some kind of bit depth or sample rate conversions in daw. It may be sample rate aliasing or something, or quantum error of internal daw/plugin bit depth. My guess, but it's only a guess.
My guess is that the problem is in the digitally-generated sweep. But I kind of like this sound and wouldn't mind having it in a plug-in to apply to any signal source as an effect. DM
I don't know of a way once it's burnt into your recording. It's best to avoid in the first place. Here's some useful info... www.dpamicrophones.com/mic-university/the-basics-about-comb-filtering-and-how-to-avoid-it
My guess for now is that it's in the sweep, which is digitally generated and could be some kind of zipper noise effect. I'm hoping for an analogue function generator for my birthday next time around. DM
@@AudioMasterclass I could be wrong but I'm pretty sure it's digital aliasing, caused by the daw's sample rate being too low for the highest frequencies, I imagine if you can lower the sample rate of your interface it would be more noticable
@@TheJakeofspades100 While we do not rule anything out, if Pro Tools or any DAW couldn't handle 20 kHz without aliasing that would be all over the internet. Our feeling is that the distortion is in the sweep.
that "distortion" you heard at the end of the file was Aliasing... Tory to play a higher sample rate version of the aduofile in a higher sample rate session and you should hear less of that. Was the wave generated by a software synthesizer - they all do anti-aliasing differently.
Yeah I heard it in the original test tone. It sounded like comb filtering was kicking in from about 10k-18k. And that was through my iPad speakers without headphones. It sounds like you are getting crossover noise in your recording from the internal components of your system, possibly a noisy GPU or power supply is contaminating the signal. You might want to try a new M1 Mac Mini and check your power setup to try to get a dedicated run/outlet from your breaker as that will help eliminate electrical noise.
You can hear this kind of sound when a rocket or fighter jet accelerates rapidly upwards to high altitude. Echoes from the ground and surroundinterrain combining with direct line of sight waves?
I'd say that nothing is prohibited but using a delay under around 40 milliseconds isn't likely to sound good. Under 20 ms definitely so. For a special effect perhaps - maybe a robot or space alien voice - then it could be useful. DM
I don't remember covering that in the video, but since comb filtering happens when a signal mixes with a delayed version of itself I would expect that in a loudspeaker where the drive units are not time aligned that around the crossover frequency the signal would come from both drive units and one would be delayed, therefore comb filtering. But since the level drops away from the crossover frequency then the effect may not be obvious, or indeed perhaps not audible at all. An experiment would be interesting but I have no plans to do this myself. DM
If you are sending below 80 Hz to your subwoofer and above 120 Hz to your main speakers then there's around half an octave missing, or at least at a lower level. If you like the sound then that's OK for entertainment, but it would be less than ideal for monitoring. DM
Nose hair trimmer? Is this your bowl of M&Ms with the brown ones taken out? 😂 Excellent science communication tutorial. It's hard for many to understand. Perhaps you can show how this translates in practice i.e. mic placement and boundary distance in a real setting? Tx
![Everything You Need To Know About Comb Filtering [with Audio Examples]](http://i.ytimg.com/vi/0wvlrBx3U4c/mqdefault.jpg)
This is very helpful. I've heard comb filtering "explained" and told to "listen for it," but never heard it so clearly demonstrated. Now I actually know what to listen for. Thank you!
You're welcome. Thank you for your comment.
That's it, Delay Noise by 5 milli seconds to get an EDM Banger
Watching this two years later - and I have to say this is outstanding. I've never had this explained properly before. Its now something I won't forget in a hurry.
All very clearly explained AND audible - thanks!
You're welcome.
Excellent job explaining the impact of reflected sound in a room! I am an audiophile building a new family room that is being acoustically treated. The folks interested in my hobby talk about the impact of room acoustics on your stereo and your video clearly illustrates the audio impact of reflected sound. Thank you!👍
Spot on. Great explanation. The very best modern era example of comb filtering is nearly ALL podcasts and audiobooks. There are very, VERY few recorded today free of it. It's so pervasive (even from professional producers), that most listener's ears have accepted it as normal. These recordings are so bad, it's very difficult for anyone that works with audio to listen to them. ESPECIALLY for an extended period of time. The next time you snuggle in with a good podcast, or audiobook and think to yourself..."Gee that sounds bad," but don't know why...that's comb filtering at work.
Some youtube channels have this and sadly some audio engineering channels as well.
I could hear comb filtering on the radio after pressing the Mono button. When a sound card plays back, it usually introduces a 1 sample delay between channels. Programmes on repeat were passed through another generation of DAC at the radio station, resulting in 2 samples or about 40 µs of delay. When summed to mono, there was a noticeable loss of treble.
If you mix EQ'd and dry signal together, you get more cancellation because the minimum phase EQ introduces a group delay around the steep lowpass filter. I think a shallow filter is more realistic.
Your sweep generator has aliasing and and there is a short repeated buffer at the end of the clip.
I am a bass player & have always wanted to run two bass cabinets on either side of the drummer, thinking that my bass sound would spread more 'evenly'. I never did that because I was told about comb filtering. Good thing I didn't~ my bass sound 50 feet out from the stage would have been a big mess! Now I understand why~ Thank you
I'd have to say in this case that theory and practice may differ because there will be so many reflections in the room that any comb filtering may be masked. The only way you'll really know is to try it! DM
Very nice discussion. This explains a lot issues I 've with table tops, laptops, and nearby windows....Thanks for the insight.
excellent demonstration
I was suprised by that spectragraph software (plugin?) and the lumpy effect at the bottom end of the sweep, didnt the writer do any tests?, really instills confidence in this being "Pro Tools" in function rather than just in name.
As for the comb filtered sound, sounds like listening to the end of a decent lenth of poly-pipe.
I think that distortion was nyquist distortion from antialiasing (or lack thereof) in the soundcard/ASIO software what-have-you
i've heard at the end of the sweet, sound that were like ending on a frequency and then other sound starting at lower frequency and continuing where the other was ending, in some king of an overlap, excuse my English, this is the best i know
Fantastic explanation, great video. Now I'm curious, about something. Around 12:30 when you added 5ms(?) delay, you commented and said it sounded "dreadful". This sound, I call sewer pipe, as it sounds like you are in a tube. I hate satellite radio and highly compressed mp3's because of something similar. I wasn't expecting comb filtering to be an artifact of satellite radio. Is it similar or possible? Why do overly compressed audio files sound like they are in a tunnel like this, is it the same thing?
I think these might be different distortions that possibly sound similar. People's subjective experiences may differ. This is a topic I may look into further in future. DM
That sound is so cool. It sounds like its inside ur head. Is that binural panning? I use it often but i havent ever heard it like that before. I habe ever listened to it in this manner either. I havent watched the rest of the video yet, but just wanted to say that was pretty cool!
That distortion may be caused by some kind of bit depth or sample rate conversions in daw. It may be sample rate aliasing or something, or quantum error of internal daw/plugin bit depth. My guess, but it's only a guess.
My guess is that the problem is in the digitally-generated sweep. But I kind of like this sound and wouldn't mind having it in a plug-in to apply to any signal source as an effect. DM
@@AudioMasterclass Yeah, sounds like a radiowaves) Aliasing maybe
so informative! thank you!
Thank you so much for this through tutorial!
You're welcome. DM
I need to know if i can remove comb-filtering or reflective delays on a mono vocal signal..with software obviously..if possible.
I don't know of a way once it's burnt into your recording. It's best to avoid in the first place. Here's some useful info... www.dpamicrophones.com/mic-university/the-basics-about-comb-filtering-and-how-to-avoid-it
As the waveform narrowed I heard the distortion too, and I'm using relatively cheap ear buds, so yes something is there.
My guess for now is that it's in the sweep, which is digitally generated and could be some kind of zipper noise effect. I'm hoping for an analogue function generator for my birthday next time around. DM
I heard it too
@@AudioMasterclass That sounds like a super idea 😀
@@AudioMasterclass I could be wrong but I'm pretty sure it's digital aliasing, caused by the daw's sample rate being too low for the highest frequencies, I imagine if you can lower the sample rate of your interface it would be more noticable
@@TheJakeofspades100 While we do not rule anything out, if Pro Tools or any DAW couldn't handle 20 kHz without aliasing that would be all over the internet. Our feeling is that the distortion is in the sweep.
that "distortion" you heard at the end of the file was Aliasing... Tory to play a higher sample rate version of the aduofile in a higher sample rate session and you should hear less of that. Was the wave generated by a software synthesizer - they all do anti-aliasing differently.
How do they use this in physical modeling synthesis?
Make the delay time of one copied track variable for example plus and minus 10 ms and you got a nice instant flanging effect!
You're right. There might be a whole new musical genre here. DM
Yeah I heard it in the original test tone. It sounded like comb filtering was kicking in from about 10k-18k. And that was through my iPad speakers without headphones. It sounds like you are getting crossover noise in your recording from the internal components of your system, possibly a noisy GPU or power supply is contaminating the signal. You might want to try a new M1 Mac Mini and check your power setup to try to get a dedicated run/outlet from your breaker as that will help eliminate electrical noise.
You can hear this kind of sound when a rocket or fighter jet accelerates rapidly upwards to high altitude. Echoes from the ground and surroundinterrain combining with direct line of sight waves?
7:54 Anybody else hearing pitches and different perfect intervals when he switches between no delay and delay in the white noise chapter?
Great poems, William Blake) -Nobody
Thank you Sir.
You're welcome.
so....no delays under 20 ms?? thanks for all amazing info
I'd say that nothing is prohibited but using a delay under around 40 milliseconds isn't likely to sound good. Under 20 ms definitely so. For a special effect perhaps - maybe a robot or space alien voice - then it could be useful. DM
@@AudioMasterclass thanks, i should definitely be more careful after watching this
How does having a tweeter forward in a speaker cabinet produce comb filtering?
I don't remember covering that in the video, but since comb filtering happens when a signal mixes with a delayed version of itself I would expect that in a loudspeaker where the drive units are not time aligned that around the crossover frequency the signal would come from both drive units and one would be delayed, therefore comb filtering. But since the level drops away from the crossover frequency then the effect may not be obvious, or indeed perhaps not audible at all. An experiment would be interesting but I have no plans to do this myself. DM
Basically, comb filtering annoys me when unintentionally applied for a whole frickin MOVIE!
What happens if I send crossover80Hz to my subwoofer and 120Hz crossover to main speakers, will I lose much bass?
If you are sending below 80 Hz to your subwoofer and above 120 Hz to your main speakers then there's around half an octave missing, or at least at a lower level. If you like the sound then that's OK for entertainment, but it would be less than ideal for monitoring. DM
@@GladeSwope As in my previous reply "at a lower level". If you like it you should stick with it. DM
I ve wandered lonely as a cloud
Nose hair trimmer? Is this your bowl of M&Ms with the brown ones taken out? 😂 Excellent science communication tutorial. It's hard for many to understand.
Perhaps you can show how this translates in practice i.e. mic placement and boundary distance in a real setting? Tx
“How not to do a Haas effect”
The precedence effect and Haas's interpretation of it are interesting and useful in stereo. But flatten the channels into mono - Comb filtering!
intersting 18:13
Amazing video! Thank you
You're welcome. DM
Great video! Thanks!
You're welcome.