Screencap below is a bunch of 5-10 sec clips of audio books (speech) straight off YouTube. The first 4 are posted by the same account, 5 by a different account. The [horizontal] dark bands make no sense to me -- not harmonic series/artifacts, improbable that HW on the analog side would have such pronounced notches. Don't vary throughout a single file, but often vary from file to file. I'm confused. Ideas?
P.S. Tried doing something similar (two bands below "my cuts") -- barely noticeable to my tin ear.

 

its probably the way the encoder throws away harmonic information, or could even be with the original mic used , pre-amp, and characteristic of the voice. it gets real complicated real fast regarding possible variables that would affect such a thing.

 

I am confused by your post. Can you share the Youtube link?

 

Which part? Unsure how a link would make things any clearer, but here.

 

Sorry, I couldn't reply earlier. To me, this looks like the effect of subtractive EQ/filter. Also, watermarking adds information across the spectrum and looks like noise, so it cannot produce "dark bands". Furthermore, in my screengrab, this is not as pronounced.

 

Behringer watermarks....bastards.

What can ya do.

 

A watermark could be anything you want it to be<=click, including crazy EQ settings (band/notch filtering). The question's why would anyone do that. If it's shitty encoder/hardware (as @MMJ2017 suggests), why would these bands change from file to file?

Not saying this is some cunning new watermarking strategy, just that I don't expect to see quirky harmonic profiles in lofi speech (with original likely being electrical transcription discs, if not crusty Nth gen carts/cassettes).

 

They are just comfort cuts.

It is not uncommon to cut ~2500 / ~4300-5000 / and smiley face the 1400 - 720.

Some people call this warming things up, some people call them comfort cuts, reducing the frequencies that the majority of people on this planet are most sensitive to helps reduce ear fatigue a bit (especially with headphone users where HRTF can amplify the effect of ear fatigue and tinnitus with the said frequencies.)

The guy who done, did, ur editing just ham fisted is all.

 

Last edited: Sep 21, 2017

why would you act destructively to the audio like that?
(are you honestly telling me that you have a region of the wave from 4.3khz to 4.7khz that is 100 db too loud? , but the rest is fine? WTF?)

do you know what is literally happening when you make that shape?
you are twisting the phase of the signal in knots in small tiny places, which is destructive toward the transients of the signal. for the life of me cannot see any intelligent benefit of doing such a thing ( what could it possibly accomplish?) if you instead put 2 plugins one tube pre-amp analog modeling, and second analog modeling of tape machine will giving you actual results without being destructive toward the signal.

to cut tiny slim notches out of your audio ,it would be like recording video with a high end camera , and then proceeding to go in and make individual rows of pixels black or white.

when you do any act with a plug or unit look at the waveform before and after (when not very experienced, in order to learn about which adjustments do what specifically)

an equalizer is just adjusting the phase of your audio signal , but instead of flipping it , like the button does, your changing the phase in certain frequency response and Q size .( which has effect of perception lowers the volume or raises it of that area of frequency response)

you should only put a unit or effect on a sound to do a specific thing that the signal needs, you listen and see what it needs .( get the signal optimized to its best sound possible)p.s.
the signal would never need to be completely silent in the area of 4.3khz and 4.7khz for example if it was a vocal it would sound like you had a lisp, if it was an instrument it would sound farty crappy. wow. lol hahahahaha
most bizarre thing i have ever seen with an eq in my life. (its just a friggin volume control for certain regions of your sound homie)

the way you use an eq is to lower or raise by the amount in reverse in the region and q size the same. say i had a vocal and from 2khz to 3khz was too loud by 1.5 db you lover in that range by that amount etc. that is what eq can accomplish and all it can accomplish. say you record a bass drum that is too quiet from 50hz to 95hz by 3 db you raise from 50hz to 95kz by 3 db. that is all eq can do homes. yo dont boost tiny regions by 100 db or cut tiny regions by 100 db . there is no example where that is the exact thing you audio needs . it seems like you are doing things like that because you don't know what actually needs to be done to a sound at any particular point in time or moment you are in. if there ever any certain results you want to know how to do you can ask me anytime and many others around here also homie.

 

Last edited: Sep 24, 2017

Really? No phase shift occurs with lin-phase EQs

Maybe, http://www.moultonlabs.com/images/gallery_images/articles/equal_loudness_contours04.gif (equal loudness graph)

 

@junh1024

correct but that image is for 100 SPL, that is way too loud to calibrate an audio book for.

Between 60 - 87 SPL is more common for the majority of audio outside of loud club mixes.

35 SPL quiet library
60 SPL median human output
87 SPL peak loudness of properly calibrated Dolby theater system (also the point at which most people can't really tell if things are any louder by hearing alone.)

 

"...A linear-phase EQ eliminates phasing by time shifting the frequencies (all the frequencies) so that the phase is not heard. In simple terms it will delay your signal so that all the frequencies within the track are kept in phase..."


"
IS THIS BAD?

Well all linear phase equalizers will introduce some major latency issues to your song. See this time shifting that a linear-phase EQ does causes a delay in the sound. Yes you will lose any phasing at the frequency points that you EQ (and surrounding affected frequencies), but you will also have an issue with transient smearing as a trade off. This is not always a good trade and in my option this smearing makes the track sound plastic like."

WHY NOT USE THIS ALL THE TIME?

In most cases it would be unnecessary because the audible artifacts that you get from a minimum phase EQ won't be detected in your track anyway. Secondly it is not an “all the time tool” but more of a specialty tool that can work along side your other production tools. As a matter of fact there are even some minimum phase EQ's that can be switched over from minimum to linear-phase on the fly.

Linear-phase equalizers are mostly used in a mastering setting as most mastering houses will have a Weiss EQ1-LP somewhere in the building. This does not mean that a linear-phase EQ can only be used in mastering situations only, but it does mean that in most cases a linear-phase EQ is going to be used sparingly instead of being used on every track of a song where an equalizers is needed.

I personally do not use a linear-phase EQ often, but in most cases I find that I use one to handle sub frequencies on tracks I am asked to mix. Sometimes I run into a muddy mix down there (very low frequencies) that my client could not hear on their system that needs to be cleaned up. In these situations I find a linear-phase EQ to be just what is needed for the job.



SO DO I NEED A LINEAR PHASE EQUALIZER?

Well I can not answer that question for you because I don't know all your needs. But I can tell you that if you don't already have and use one that you may not need one. If you do decide that you need a linear-phase EQ to fill your production toolbox then you can choose a dual purpose EQ that works as a minimum phase as well as a linear-phase EQ. Or you can choose from the many freebies that are out there. Filtrate LE and Cockos' ReaFIR are two of my favorite free linear-phase EQ plugins. Download one of them and try it out for yourself.



CONCLUSION

A linear phase EQ is one of the most mis-understood audio effects out there. I won't pretend to think that this article explained it all to you. Or that I am all-knowledgeable when it comes to all the inter workings of the linear phase EQ. There is a lot to learn about this little gem of the studio. I would recommend that you try out a demo or grab a freebie and just start experimenting with it.

 

Linear Phase EQ does introduce time shifted artifacts into audio, this is the pre-ringing people always talk about. It most pronounced in infra-low and low frequencies. In mid to mid high frequencies the the pre-ring transient might be 1-2 ms ahead of the the actual transient which is too close for the majority of humans with average to above average hearing to detect. In the lower registers where it can be between 8ms -5ms it is detectable with people who have above average hearing and are listening for it, which is why a natural phase EQ is preferable for LF work as phase distortion on a proper filter is less obvious and sometimes even expected as we've all grown up hearing it anyways.

 

Cheap mics often have "gunk" gathering at certain points in the spectrum, which when removed look exactly like the notches you see. Looking at the original file the original guy would have seen a constant bright strip of noise in that narrow freq range, all thru the entire file. So he notices it's garbage and cuts it away.

Another source of these is sometimes comb filtering effects due to delay, when multi mics are used. Sometimes the "gunk band" comes from an environmental resonance that the mic's capturing.

 

Does your DAW have delay compensation? Mine does. It should be a non-issue.

You can increase or decrease the buffer size for some FIR EQs, so possibly a non-issue. Also audiability threshold as @mozee mentioned.