Been searching for a video like this for years! Thanks for really breaking down this graph and demonstrating how to figure out a loudness equivalence between two frequencies. 🙏🏽
is 12:56 correct? Almost all music already bakes a significant tilt in, and, while bass boosting is often desirable, a pefectly flat system (that remains flat down to 40hz / 30hz) is usually regarded as very good, nonetheless, those systems also require a big subwoofer amp. I am guessing that the power requirement is actually because of the weight of the moving mass and the power of the magnet that tries to bring it back to resting position as fast as possible to stop inertia, creating lots of resistance. EDIT: This is not to say that I don't appreciate the explanation btw, thanks for the great vid!
Yes, it is correct. You have to remember that this is what we as humans always experience. Thus, when a track is mixed down, it is mixed by ear. They don't add bass to follow this curve. They adjust levels by how it sounds. In order for it to sound right, you end up with a lot more energy in the lowest octaves. On playback, this requires a perfectly flat frequency response (i.e., the effect is already "baked in").
@@ElectronicswithProfessorFiore exactly, but you are still able to play flat sweeps or white noise without overdriving upper frequencies (despite them playing at equal SPL). Or am I wrong and audio systems do usually overdrive when playing flat signals due to the manufacturers making the expectation that the signal is going to be tilted?
@@StarOnCheek I think there is some confusion between SPL and loudness. Many people take these things to be the same but they are not. Loudness is the human perception of SPL. Equal loudness curves explain how humans perceive SPL. If you produced two tones, one bass (40 Hz) and one treble (4 kHz), at the same SPL, the treble tone would sound louder. Another way of looking at this is to create a long term average of the energy present in music. In most forms of Western music, the majority of the energy is found below 500 Hz. Another item to consider is the power rating of woofers and tweeters in home loudspeakers. It would not be odd to find a woofer rated at 50 or 100 watts being paired with a tweeter rated at 5 or 10 watts. Check this other video I did: th-cam.com/video/8S-HdK6KGzA/w-d-xo.htmlfeature=shared
But what if i'm not having single tones but a wide band signal lots of harmonics or even noise? how can i convert a real acoustic noise to an phon value if i want to? I can't just apply a weighting curve as EQ because the curve is amplitude dependent?
The issue here revolves around what you're trying to accomplish. One example is hearing damage caused by excessive levels in an occupational environment; in other words, something like noise on a factory floor. In a case like that, a weighting filter would be used and we wouldn't talk about phons, but rather dBA (or something similar). It is understood that a scale like dBA is a proxy for loudness but is not identical to it (i.e., dBA does not equal phons). It gets the job done, though. If you wanted to, you could repeat the Fletcher-Munson experiment using band-limited noise. You wouldn't really have phons as a result, but it may prove useful. I do not recall specific experiments in this area, but I am sure that some work has been done.
@@ElectronicswithProfessorFiore Thanks for the answer! For example some PC fans i have seen loudness ratings in phons and i was curious how i/they would measure/calculate this. I know about dB(A) weighting curve. that's easy to apply and most industrial and safety regulation are given in dB(A) But i don't think dB(A) weighting is very appropriate for low volume sound at the border of auditability. "dB per 1/3 octave band" is also a common measurements but "1/3" seams arbitrary to me, the smaller the bins the higher the resolution but the lower the dB per bin. Can i overlay an Fletcher-Munson graph on a FFT or RTA to see if a signal is below audible volume? Thanks for the educational videos and taking time to answer to comments.
@@lambda7652 That's interesting. I haven't seen any fans spec'd using phons, only dBA. dBA seems to be the default, at least in the USA (in the UK, I understand that ITU-468 is a popular standard). You can "overlay a..graph on an FFT or RTA", but the issue there is getting the calibration correct. You need to know how the voltage coming out of your source (e.g., mic) corresponds to an SPL. At a minimum, you'd need a sensitivity spec like this one from a Shure SM-57 spec sheet: "Sensitivity (@ 1 kHz, open circuit voltage) -56.0 dBV/Pa (1.6 mV)". You then have to compensate for the fact that most mics (other than special measurement mics) do not have a flat frequency response.
How do I set this in reality? I can set target curve but how do I set it for this? If I listen at 75dB the the graf show 116 dB at 20hz. Should I set the target curve like this: 116dB - 75dB = 41 dB Should I set target curve at 20hz at +41? And 25hz show 111dB, so 111dB-75= 36dB. So plus + 36dB on 25Hz Etc That will crack my speakers 😅 I dont understand how to set it like this.
There might be some confusion here. This is not for setting up your home or car stereo system. Instead, it is explaining how human hearing works, and it sheds light on the sort of power differentials that are needed at different frequencies. The audio engineers and producer have already set the relative levels needed for a particular piece of music. The only mod you might make is if you are not listening at the same level they used during mix down. Unfortunately, it is extremely rare for those folks to tell you what SPL they used!
@@ElectronicswithProfessorFiore Thank you for your respons and for confirming that I confused all this. I thoght at first that I could use this for my home stereo. So this is only for producing/ mixing (?) Target curve for home stereo should be pretty flat then? With a little bump in the bass if thats what I prefer? Or do you know if there is something simalar for hifi- target curves?
@@jimmydelen2023 If what you're looking for is highest fidelity, then you want the flatest response out of the system. Of course, what you like is up to you, so listen, adjust, and decide from there.
Been searching for a video like this for years! Thanks for really breaking down this graph and demonstrating how to figure out a loudness equivalence between two frequencies. 🙏🏽
Thank you so much for this! I’ve been struggling with Physics and this video really helped me understand this topic I am very grateful!
Thanks for this video.
It brings lights to me for frequency balance mixing.
I’ll start experimenting…
Great job explaining this 👏🏼👏🏼👏🏼
is 12:56 correct? Almost all music already bakes a significant tilt in, and, while bass boosting is often desirable, a pefectly flat system (that remains flat down to 40hz / 30hz) is usually regarded as very good, nonetheless, those systems also require a big subwoofer amp. I am guessing that the power requirement is actually because of the weight of the moving mass and the power of the magnet that tries to bring it back to resting position as fast as possible to stop inertia, creating lots of resistance.
EDIT: This is not to say that I don't appreciate the explanation btw, thanks for the great vid!
Yes, it is correct. You have to remember that this is what we as humans always experience. Thus, when a track is mixed down, it is mixed by ear. They don't add bass to follow this curve. They adjust levels by how it sounds. In order for it to sound right, you end up with a lot more energy in the lowest octaves. On playback, this requires a perfectly flat frequency response (i.e., the effect is already "baked in").
@@ElectronicswithProfessorFiore exactly, but you are still able to play flat sweeps or white noise without overdriving upper frequencies (despite them playing at equal SPL). Or am I wrong and audio systems do usually overdrive when playing flat signals due to the manufacturers making the expectation that the signal is going to be tilted?
@@StarOnCheek I think there is some confusion between SPL and loudness. Many people take these things to be the same but they are not. Loudness is the human perception of SPL. Equal loudness curves explain how humans perceive SPL. If you produced two tones, one bass (40 Hz) and one treble (4 kHz), at the same SPL, the treble tone would sound louder. Another way of looking at this is to create a long term average of the energy present in music. In most forms of Western music, the majority of the energy is found below 500 Hz. Another item to consider is the power rating of woofers and tweeters in home loudspeakers. It would not be odd to find a woofer rated at 50 or 100 watts being paired with a tweeter rated at 5 or 10 watts.
Check this other video I did: th-cam.com/video/8S-HdK6KGzA/w-d-xo.htmlfeature=shared
Explained Very Good 👍
Thanks
But what if i'm not having single tones but a wide band signal lots of harmonics or even noise? how can i convert a real acoustic noise to an phon value if i want to?
I can't just apply a weighting curve as EQ because the curve is amplitude dependent?
The issue here revolves around what you're trying to accomplish. One example is hearing damage caused by excessive levels in an occupational environment; in other words, something like noise on a factory floor. In a case like that, a weighting filter would be used and we wouldn't talk about phons, but rather dBA (or something similar). It is understood that a scale like dBA is a proxy for loudness but is not identical to it (i.e., dBA does not equal phons). It gets the job done, though.
If you wanted to, you could repeat the Fletcher-Munson experiment using band-limited noise. You wouldn't really have phons as a result, but it may prove useful. I do not recall specific experiments in this area, but I am sure that some work has been done.
@@ElectronicswithProfessorFiore Thanks for the answer!
For example some PC fans i have seen loudness ratings in phons and i was curious how i/they would measure/calculate this.
I know about dB(A) weighting curve. that's easy to apply and most industrial and safety regulation are given in dB(A)
But i don't think dB(A) weighting is very appropriate for low volume sound at the border of auditability.
"dB per 1/3 octave band" is also a common measurements but "1/3" seams arbitrary to me,
the smaller the bins the higher the resolution but the lower the dB per bin.
Can i overlay an Fletcher-Munson graph on a FFT or RTA to see if a signal is below audible volume?
Thanks for the educational videos and taking time to answer to comments.
@@lambda7652 That's interesting. I haven't seen any fans spec'd using phons, only dBA. dBA seems to be the default, at least in the USA (in the UK, I understand that ITU-468 is a popular standard).
You can "overlay a..graph on an FFT or RTA", but the issue there is getting the calibration correct. You need to know how the voltage coming out of your source (e.g., mic) corresponds to an SPL. At a minimum, you'd need a sensitivity spec like this one from a Shure SM-57 spec sheet: "Sensitivity (@ 1 kHz, open circuit voltage) -56.0 dBV/Pa (1.6 mV)". You then have to compensate for the fact that most mics (other than special measurement mics) do not have a flat frequency response.
How do I set this in reality?
I can set target curve but how do I set it for this?
If I listen at 75dB the the graf show 116 dB at 20hz.
Should I set the target curve like this:
116dB - 75dB = 41 dB
Should I set target curve at 20hz at +41?
And 25hz show 111dB, so 111dB-75= 36dB. So plus + 36dB on 25Hz
Etc
That will crack my speakers 😅
I dont understand how to set it like this.
There might be some confusion here. This is not for setting up your home or car stereo system. Instead, it is explaining how human hearing works, and it sheds light on the sort of power differentials that are needed at different frequencies. The audio engineers and producer have already set the relative levels needed for a particular piece of music. The only mod you might make is if you are not listening at the same level they used during mix down. Unfortunately, it is extremely rare for those folks to tell you what SPL they used!
@@ElectronicswithProfessorFiore
Thank you for your respons and for confirming that I confused all this. I thoght at first that I could use this for my home stereo.
So this is only for producing/ mixing (?)
Target curve for home stereo should be pretty flat then? With a little bump in the bass if thats what I prefer?
Or do you know if there is something simalar for hifi- target curves?
@@jimmydelen2023 If what you're looking for is highest fidelity, then you want the flatest response out of the system. Of course, what you like is up to you, so listen, adjust, and decide from there.
Thanks sir
Not an expert in this by any means but there’s a new (revised) ISO standard, IOS 226:2023
Thanks for that. I took a look at the curve set from a prelim doc and there do not appear to be any noticeable changes.