|
Post by JKCashin on Feb 21, 2021 20:49:36 GMT -5
Is it just me or does the new re-implemented loudness curve (modelled after the Fletcher Munson studies I believe) really sound nice?
It goes against my purist nature, but I can't hel;p but liking it. Sigh
|
|
|
Post by markc on Feb 22, 2021 4:54:17 GMT -5
I like it too!
I disagree that it is not purist though.
Our ears/brains have varied frequency sensitivity at different volumes and the FM curves take an average person's ear/brain sensitivity and applies it to audio to compensate for the variation depending on the volume you are listening at.
In this way, audio played at a low volume sounds more similar to the same audio played at a louder volume, only louder (obvs!)
Without it, you get a skew, and listening at low volume, the bass disappears most of all
|
|
|
Post by megash0n on Feb 22, 2021 6:11:24 GMT -5
I like it too! I disagree that it is not purist though. Our ears/brains have varied frequency sensitivity at different volumes and the FM curves take an average person's ear/brain sensitivity and applies it to audio to compensate for the variation depending on the volume you are listening at. In this way, audio played at a low volume sounds more similar to the same audio played at a louder volume, only louder (obvs!) Without it, you get a skew, and listening at low volume, the bass disappears most of all I'm following you. I had it on for a few days, but turned it off. I'm not sure if this is what is causing the pops or not. For general TV watching, it did seem to provide a lot of missing audio detail. Almost like turning the sub on or off with your mains set to small. But also more in the vocals as well. There are attributes I like about it. I would like to know how it is implemented as I have doubts it is a hundred percent correct. Unfortunately, my default is doubt now. The one area I am doubting is due to playing a stereo audio track of rain. We do this every single night, and have for over a year now, with the G3P. I have an automation task that sets the volume at -42, surround and height trims to +5, and the mode to All Stereo. Then rain starts being played over HDMI from a HTPC. All this uses the network API fwiw. (Except playing the audio obviously) When loudness is enabled, the tone and pitch is so drastically different I cannot listen to the rain. It sounds like someone grabbed the 5 band eq sliders, pushed all the bass down and the treble side all the way up. So, I'm wondering if it behaves more correctly on non-upmixed content or it that matters. Lastly, we can say all day long what shouldn't matter, or how it is supposed to be, but that isn't the world we live in drinking this cool aid. DTS wasn't right for how many FW releases? Was that issue every really acknowledged in the 2.1 release? I could go search, but if I'm not mistaken... You were part of the vocal crew about the DTS issues as well.
|
|
cawgijoe
Emo VIPs
"We made too many of the wrong mistakes." - Yogi Berra
Posts: 4,896
|
Post by cawgijoe on Feb 22, 2021 6:48:21 GMT -5
I don’t use loudness. Never have and never will. I’ve never had a need for it. I did turn it on and it seems to work. Did not wait to hear any pops or ticks.
|
|
|
Post by 405x5 on Feb 22, 2021 8:41:30 GMT -5
I never read subwoofer AND “Loudness” implementation in these discussions. That’s where the problem is. Loudness is old hat. Been around for sixty years at least but subs weren’t used in the old days the way they are now. With no sub. Loudness has SOME merit at times. With a sub.....a nice disaster to avoid.
Bill
|
|
|
Post by AudioHTIT on Feb 22, 2021 11:39:10 GMT -5
I appreciate having Loudness as an option, and use it on occasion, especially when playing ‘background music’. As markc mentioned, when we listen to instruments and sounds played back at a lower volume than they naturally produce, we perceive them with less bass and treble — the change in bass can be pretty drastic, treble less so. The Fletcher-Munson curves attempt to quantify how this perception changes at different volumes. The problem with virtually any variable implementation is choosing a level where the curve is flat, because each of our systems have different amplifier and speaker sensitivities. If you have a sensitive amp (driven to full power with a lower voltage), and speakers (driven to higher SPL with a lower voltage), then you’d want the Loudness curve to go away at a lower setting of the volume control (attenuator), say -30 dB. If you have lower sensitivity amp and speakers you might want the setting to be say -15 dB. In the perfect world our systems would all be set to produce a specific SPL at say -20 or 0 dB, then this setting would be consistent for all of us. But since our audio world isn’t perfect, manufacturers (Emotiva), has to choose a compromise that works/sounds best to most of us (which of course means some will like it more than others). P.S. The Fletcher-Munson curves are the results of the original studies on this phenomenon, there are newer studies that are slightly different termed “Equal Loudness Contours”.
|
|
|
Post by megash0n on Feb 22, 2021 11:46:45 GMT -5
I appreciate having Loudness as an option, and use it on occasion, especially when playing ‘background music’. As markc mentioned, when we listen to instruments and sounds played back at a lower volume than they naturally produce, we perceive them with less bass and treble — the change in bass can be pretty drastic, treble less so. The Fletcher-Munson curves attempt to quantify how this perception changes at different volumes. The problem with virtually any variable implementation is choosing a level where the curve is flat, because each of our systems have different amplifier and speaker sensitivities. If you have a sensitive amp (driven to full power with a lower voltage), and speakers (driven to higher SPL with a lower voltage), then you’d want the Loudness curve to go away at a lower setting of the volume control (attenuator), say -30 dB. If you have lower sensitivity amp and speakers you might want the setting to be say -15 dB. In the perfect world our systems would all be set to produce a specific SPL at say -20 or 0 dB, then this setting would be consistent for all of us. But since our audio world isn’t perfect manufacturers (Emotiva), has to choose a compromise that works/sounds best to most of us (which of course means some will like it more than others). Interesting insight. This sparked a thought. Do you think it would be fairly straightforward to utilize a 20-20k measurement sweep (all automated) to perform tests at say... 4 different volume levels in order to "map out" the SPL ranges and basic analysis of in room response at those levels? I don't think you could do much with the varying dips and peaks at different SPL levels due to other room correction systems at play. It more so would be a rough estimate of "at what point am I flat" to be used for "how" to implement this loudness function. Also, there's the whole topic of how one's implemented house curve comes into play as well. My intention here is to spitball for a much larger topic of critical thinking and white boarding for such endeavors.
|
|
|
Post by AudioHTIT on Feb 22, 2021 12:02:30 GMT -5
I appreciate having Loudness as an option, and use it on occasion, especially when playing ‘background music’. As markc mentioned, when we listen to instruments and sounds played back at a lower volume than they naturally produce, we perceive them with less bass and treble — the change in bass can be pretty drastic, treble less so. The Fletcher-Munson curves attempt to quantify how this perception changes at different volumes. The problem with virtually any variable implementation is choosing a level where the curve is flat, because each of our systems have different amplifier and speaker sensitivities. If you have a sensitive amp (driven to full power with a lower voltage), and speakers (driven to higher SPL with a lower voltage), then you’d want the Loudness curve to go away at a lower setting of the volume control (attenuator), say -30 dB. If you have lower sensitivity amp and speakers you might want the setting to be say -15 dB. In the perfect world our systems would all be set to produce a specific SPL at say -20 or 0 dB, then this setting would be consistent for all of us. But since our audio world isn’t perfect manufacturers (Emotiva), has to choose a compromise that works/sounds best to most of us (which of course means some will like it more than others). Interesting insight. This sparked a thought. Do you think it would be fairly straightforward to utilize a 20-20k measurement sweep (all automated) to perform tests at say... 4 different volume levels in order to "map out" the SPL ranges and basic analysis of in room response at those levels? I don't think you could do much with the varying dips and peaks at different SPL levels due to other room correction systems at play. It more so would be a rough estimate of "at what point am I flat" to be used for "how" to implement this loudness function. Also, there's the whole topic of how one's implemented house curve comes into play as well. My intention here is to spitball for a much larger topic of critical thinking and white boarding for such endeavors. So, in my opinion, you wouldn’t want to measure frequency response of the room in order to compensate with Loudness, that should be the job of ARC like Dirac. But, you would want to know the SPL your system produces at a given volume setting (with a typical input setting). Assuming our ears are similar (though not true, we have to start there to implement Loudness), the frequency compensation would be the same for every room because we assume that’s already been compensated for with ARC. At this point however, there is nothing we can do to compensate for our various system sensitivities, and how Loudness affects them. One possibility would be a setting that would allow us to choose where on the volume control the compensation goes away (sort of a Loudness ‘Break Point’). During the beta I took measurements of the ‘electrical’ changes to frequency response at different levels, which is one way to find out what compensation is being implemented at different levels. They are no longer accurate as Emotiva chose a different break point than I measured at, and I’m not in a position to take new measurements now. I’ll try to post what I took when I get a chance. Edit JKCashin, might want to fix the spelling in the title of this thread (I’m reminded of Luddites 😁)
|
|
|
Post by megash0n on Feb 22, 2021 12:54:41 GMT -5
Interesting insight. This sparked a thought. Do you think it would be fairly straightforward to utilize a 20-20k measurement sweep (all automated) to perform tests at say... 4 different volume levels in order to "map out" the SPL ranges and basic analysis of in room response at those levels? I don't think you could do much with the varying dips and peaks at different SPL levels due to other room correction systems at play. It more so would be a rough estimate of "at what point am I flat" to be used for "how" to implement this loudness function. Also, there's the whole topic of how one's implemented house curve comes into play as well. My intention here is to spitball for a much larger topic of critical thinking and white boarding for such endeavors. So, in my opinion, you wouldn’t want to measure frequency response of the room in order to compensate with Loudness, that should be the job of ARC like Dirac. But, you would want to know the SPL your system produces at a given volume setting (with a typical input setting). Assuming our ears are similar (though not true, we have to start there to implement Loudness), the frequency compensation would be the same for every room because we assume that’s already been compensated for with ARC. At this point however, there is nothing we can do to compensate for our various system sensitivities, and how Loudness affects them. One possibility would be a setting that would allow us to choose where on the volume control the compensation goes away (sort of a Loudness ‘Break Point’). During the beta I took measurements of the ‘electrical’ changes to frequency response at different levels, which is one way to find out what compensation is being implemented at different levels. They are no longer accurate as Emotiva chose a different break point than I measured at, and I’m not in a position to take new measurements now. I’ll try to post what I took when I get a chance. Edit JKCashin, might want to fix the spelling in the title of this thread (I’m reminded of Luddites 😁) I'm with ya and agree. The points you made were the important ones. In my head I was thinking through "what may happen to the measured response as we move higher in SPL" as a simple data point for reflection. Not at all as a replacement for Dirac. In my head, with 5 mins of thinking, it seems like an interesting data point to reflect on, but after a week on the whiteboard, it may be found to be useless. Your message about electrical output was my first thought a few days ago. But, not sure what mechanisms are in place to make this a functional solution for the masses. Does the G3P have the hardware to monitor, with precision, the voltage levels of each output?
|
|
|
Post by AudioHTIT on Feb 22, 2021 14:32:54 GMT -5
... Your message about electrical output was my first thought a few days ago. But, not sure what mechanisms are in place to make this a functional solution for the masses. Does the G3P have the hardware to monitor, with precision, the voltage levels of each output? No, but the compensation should be the same for all outputs (because it’s compensating for your ear/brain, not the room), I believe I used the Left Channel output and a pink noise source via TV. I then ran the signal through the balanced in of a Focusrite 2i2 and then USB to my iPad Pro 11 running AudioTools. When I post the curves you’ll see there was an ‘anomaly’ (not perfectly flat without compensation), but the curves overcome most of that at most levels. So I don’t claim my graphs are precise, but do show Loudness generally doing what it’s supposed to. Now ... time to ⛷
|
|
|
Post by JKCashin on Feb 23, 2021 0:43:31 GMT -5
I like it too! I disagree that it is not purist though. Our ears/brains have varied frequency sensitivity at different volumes and the FM curves take an average person's ear/brain sensitivity and applies it to audio to compensate for the variation depending on the volume you are listening at. <snip> Never thought of it that way.... glad to see someone else is enjoying it too. Mine pops, with or without Loudness <snip> I did turn it on and it seems to work. Did not wait to hear any pops or ticks. Would be interested to know if you *DO* hear pops/ticks with it on only. I hear them either way. <snip> P.S. The Fletcher-Munson curves are the results of the original studies on this phenomenon, there are newer studies that are slightly different termed “Equal Loudness Contours”. Yeah, I did some reading on it. I originally called the thread "Fletched-Munson Loudness Curve" but realizing it was not and there have been updates, I just called it the "Loudness Curve" Edit JKCashin , might want to fix the spelling in the title of this thread (I’m reminded of Luddites 😁) Yup... just fixed it!
|
|
|
Post by megash0n on Feb 23, 2021 6:01:10 GMT -5
I like it too! I disagree that it is not purist though. Our ears/brains have varied frequency sensitivity at different volumes and the FM curves take an average person's ear/brain sensitivity and applies it to audio to compensate for the variation depending on the volume you are listening at. <snip> Never thought of it that way.... glad to see someone else is enjoying it too. Mine pops, with or without Loudness <snip> I did turn it on and it seems to work. Did not wait to hear any pops or ticks. Would be interested to know if you *DO* hear pops/ticks with it on only. I hear them either way. <snip> P.S. The Fletcher-Munson curves are the results of the original studies on this phenomenon, there are newer studies that are slightly different termed “Equal Loudness Contours”. Yeah, I did some reading on it. I originally called the thread "Fletched-Munson Loudness Curve" but realizing it was not and there have been updates, I just called it the "Loudness Curve" Edit JKCashin , might want to fix the spelling in the title of this thread (I’m reminded of Luddites 😁) Yup... just fixed it! auto correct is brutal on this forum. 🤪
|
|
Lsc
Emo VIPs
Posts: 3,342
|
Post by Lsc on Feb 23, 2021 9:20:02 GMT -5
I don’t use loudness. Never have and never will. I’ve never had a need for it. I did turn it on and it seems to work. Did not wait to hear any pops or ticks. I also do not use loudness either and don’t plan on it. I wasn’t sure why Emotiva felt the need to add this feature but I see it now. It’s being used. I wasn’t sure if other processors have this feature or not.
|
|
KeithL
Administrator
Posts: 9,929
|
Post by KeithL on Feb 23, 2021 10:10:58 GMT -5
If you want to make an accurate measurement of the response of any electronic system or device ... then you measure it ELECTRONICALLY. (In-room measurements are far too subject to variations and external influences to allow you to accurately measure the response of electronic systems.)
Then, if you like, you can sum that information with the response of your room correction, to get the total response. Of course, if you want a truly accurate rendition of the original performance, then you must simply play it at the original level at which it was recorded. In serious photography, if you want to know that the colors are correct, you have your model hold up a color checkerboard with calibrated color squares on it. If you wanted a recording to be similarly accurate you would record a level tone and response sweep at the beginning of the recording. (You would provide a graph of how that sweep measured in the actual venue - and listeners would calibrate their room to that same response.)
Of course, in many cases, with both pictures and recordings, the end result is significantly altered from the original, so absolute accuracy is somewhat moot. (And, with multi-track recordings, there is technically no "complete original" at all.)
And, in this case, the goal of "Loudness compensation" is to be able to play something at an arbitrary level and have it sound "perceptually correct"...
I appreciate having Loudness as an option, and use it on occasion, especially when playing ‘background music’. As markc mentioned, when we listen to instruments and sounds played back at a lower volume than they naturally produce, we perceive them with less bass and treble — the change in bass can be pretty drastic, treble less so. The Fletcher-Munson curves attempt to quantify how this perception changes at different volumes. The problem with virtually any variable implementation is choosing a level where the curve is flat, because each of our systems have different amplifier and speaker sensitivities. If you have a sensitive amp (driven to full power with a lower voltage), and speakers (driven to higher SPL with a lower voltage), then you’d want the Loudness curve to go away at a lower setting of the volume control (attenuator), say -30 dB. If you have lower sensitivity amp and speakers you might want the setting to be say -15 dB. In the perfect world our systems would all be set to produce a specific SPL at say -20 or 0 dB, then this setting would be consistent for all of us. But since our audio world isn’t perfect manufacturers (Emotiva), has to choose a compromise that works/sounds best to most of us (which of course means some will like it more than others). Interesting insight. This sparked a thought. Do you think it would be fairly straightforward to utilize a 20-20k measurement sweep (all automated) to perform tests at say... 4 different volume levels in order to "map out" the SPL ranges and basic analysis of in room response at those levels? I don't think you could do much with the varying dips and peaks at different SPL levels due to other room correction systems at play. It more so would be a rough estimate of "at what point am I flat" to be used for "how" to implement this loudness function. Also, there's the whole topic of how one's implemented house curve comes into play as well. My intention here is to spitball for a much larger topic of critical thinking and white boarding for such endeavors.
|
|
cawgijoe
Emo VIPs
"We made too many of the wrong mistakes." - Yogi Berra
Posts: 4,896
|
Post by cawgijoe on Feb 23, 2021 10:17:21 GMT -5
Never thought of it that way.... glad to see someone else is enjoying it too. Mine pops, with or without Loudness Would be interested to know if you *DO* hear pops/ticks with it on only. I hear them either way. Yeah, I did some reading on it. I originally called the thread "Fletched-Munson Loudness Curve" but realizing it was not and there have been updates, I just called it the "Loudness Curve" Yup... just fixed it! auto correct is brutal on this forum. 🤪 I used it for about an hour today with Directv. Switching channels....no pops or ticks. It works with Dolby Digital 5.1. It does not work with Dolby 2.0...at least with DTV...get a message "No Loudness". I did not try it with the ATV or any other source. Had to head into work. It does work and I can see a use for it, say for FM radio or music if you don't want to bother others....low volume but adds a "boost". I will probably never use it.
|
|
|
Post by megash0n on Feb 23, 2021 11:28:57 GMT -5
If you want to make an accurate measurement of the response of any electronic system or device ... then you measure it ELECTRONICALLY. (In-room measurements are far too subject to variations and external influences to allow you to accurately measure the response of electronic systems.)
Then, if you like, you can sum that information with the response of your room correction, to get the total response. Of course, if you want a truly accurate rendition of the original performance, then you must simply play it at the original level at which it was recorded. In serious photography, if you want to know that the colors are correct, you have your model hold up a color checkerboard with calibrated color squares on it. If you wanted a recording to be similarly accurate you would record a level tone and response sweep at the beginning of the recording. (You would provide a graph of how that sweep measured in the actual venue - and listeners would calibrate their room to that same response.)
Of course, in many cases, with both pictures and recordings, the end result is significantly altered from the original, so absolute accuracy is somewhat moot. (And, with multi-track recordings, there is technically no "complete original" at all.)
And, in this case, the goal of "Loudness compensation" is to be able to play something at an arbitrary level and have it sound "perceptually correct"...
Interesting insight. This sparked a thought. Do you think it would be fairly straightforward to utilize a 20-20k measurement sweep (all automated) to perform tests at say... 4 different volume levels in order to "map out" the SPL ranges and basic analysis of in room response at those levels? I don't think you could do much with the varying dips and peaks at different SPL levels due to other room correction systems at play. It more so would be a rough estimate of "at what point am I flat" to be used for "how" to implement this loudness function. Also, there's the whole topic of how one's implemented house curve comes into play as well. My intention here is to spitball for a much larger topic of critical thinking and white boarding for such endeavors. Agree with your logic, but where do you have voltage monitoring hardware embedded in your signal path? Secondly, isn't the basis for the whole thought experiment regarding "what you hear"? You don't hear whether the output voltage is 1V or 2V. Sure, you may hear a difference, but you don't understand what that means without additional context. I'm not arguing against your points. I'm only suggesting there might be an interesting way to provide your customers with an optional function that could be executed from the G3P that could better tailor the Loudness functionality for the individual consumer. As it has been stated, there are many variables in play, and one size does not fit all. Well, unless I completely miss the point of the theory which is absolutely possible. You checked a box off and delivered a feature you promised. Good job. We hope it is working correctly. My suggestions are simply an attempt to raise the bar to provide a more tailored experience. To provide value to the customer. To go that extra mile. That's it. My suggestion isnt the best way. It's the best way, logically, I could think of that would provide value to the customer while minimizing operational cost to Emotiva. It's just a thought. 😊
|
|
|
Post by megash0n on Feb 23, 2021 11:38:14 GMT -5
auto correct is brutal on this forum. 🤪 I used it for about an hour today with Directv. Switching channels....no pops or ticks. It works with Dolby Digital 5.1. It does not work with Dolby 2.0...at least with DTV...get a message "No Loudness". I did not try it with the ATV or any other source. Had to head into work. It does work and I can see a use for it, say for FM radio or music if you don't want to bother others....low volume but adds a "boost". I will probably never use it. At this point, I can only conclude that the individual components inside each batch of processors must be different at times. Like.. I found a lot of 20,000 XYZ units. Let's use these. We follow a standard, so it all should work the same. If you follow me. Think.. PA-1 It is very clear that many of you have rock solid units, and it is understandable why some of you guys think the other half is crazy. 😂😂😂 I don't know how else to really explain it. The instant I installed 2.2, it is popping and making weird noises worse than... What was it? V1.7? There was one FW that did this a lot. Now, it is only slightly louder at times, so I haven't had any scares of 120dB pops. But, it's pretty bad and very often. Could be sources, could be cables, could be a lot of things. What is certainly true though is that every FW release has some change in behavior and sonic signature. One can take any position, and stand firm on some version is " Correct ", but why does it keep changing? What is correct at this point? In a little over a year, my unit behaves differently, and sounds differently, with every FW update. Anyways, just something to ponder I guess.
|
|
|
Post by audiobill on Feb 23, 2021 19:41:11 GMT -5
This all seems so strange...
Only one question, but how does it sound; does the sound please you?
|
|
|
Post by autocrat on Feb 23, 2021 19:54:13 GMT -5
FWIW, my XMC-1 is currently back in my system (Do.Not.Ask) and if I turn Loudness on I get little pops and crackles eg. when I change TV channels. Same problem in this and the G3 units I suspect.
|
|
|
Post by JKCashin on Feb 23, 2021 20:49:10 GMT -5
This all seems so strange... Only one question, but how does it sound; does the sound please you? Not sure who this is addressed at... I like the Loudness sound
|
|