|
|
Post by hsamwel on Mar 18, 2022 7:37:34 GMT -5
Should be using??? Who ever tells you such BS? You can use or they recommend, nothing more. Yes, this is highly irrelevant to the issue. Btw if you use a Harman curve does it behave the same? It should really.. Yes I got that response from Dirac support more than once. Fact is, I have actually read the Harman AES papers so I know the origin of the curves and they are irrelevant. If someone wants to boost their bass they can just boost their bass. I actually have my Fronts ramped up 3db below 100Hz but that's to balance some other irregularities in response above that point, as well as mitigating some of the BM bug issues. I end up with a flat curve when the curtain issue is sorted out. Found this online.. Talking about Harman curves.. /Quote Is there such curve for speakers? target curves for loudspeakers are pretty well researched. That's in fact what the Harman Target is based on (measure a loudspeaker that performs close to the loudspeaker target curve, but measure it with a measurement head instead of a regular microphone, and then make a headphone that produces the same result when measured with that measurement head). While the exact details are subject to preference, the overall goal for a loudspeaker is: have a flat, linear frequency response when measured on-axis in anechoic conditions (="only the sound projected directly to the front") have a smooth sound power output (="sound in all directions, not just to the front") in total, when you measure the loudspeaker not in anechoic conditions but in-situ (meaning in your listening room, including all reflections and reverb), you should end up with a frequency response resembling a straight line that goes down 10 dB from 20 Hz to 20 kHz. This is roughly -1 dB/8ve, meaning for every doubling of frequency, the SPL goes down by 1 dB. If you want to read more about it, Floyd Toole has written a whole book about that topic. While this has basically ended the discussion as to "what is the target for a loudspeaker", it has not ended the discussion on how to get there. Note that this only applies to music. Entirely different standards apply when you start talking about movies / movie theaters. /EndQuote So Diracs standard curve is actually the most correct according to Floyd Toole. For music as the writer mentions. They all say that a fully flat frequency curve on loudspeakers in a normal room isn’t prefered. Nor is the Harman for that matter.. Although many ”like” more bass so that’s why it’s used. Not because it’s ”correct”. |
|
|
|
Post by marcl on Mar 18, 2022 7:57:44 GMT -5
Yes I got that response from Dirac support more than once. Fact is, I have actually read the Harman AES papers so I know the origin of the curves and they are irrelevant. If someone wants to boost their bass they can just boost their bass. I actually have my Fronts ramped up 3db below 100Hz but that's to balance some other irregularities in response above that point, as well as mitigating some of the BM bug issues. I end up with a flat curve when the curtain issue is sorted out. Found this online.. Talking about Harman curves.. /Quote Is there such curve for speakers? target curves for loudspeakers are pretty well researched. That's in fact what the Harman Target is based on (measure a loudspeaker that performs close to the loudspeaker target curve, but measure it with a measurement head instead of a regular microphone, and then make a headphone that produces the same result when measured with that measurement head). While the exact details are subject to preference, the overall goal for a loudspeaker is: have a flat, linear frequency response when measured on-axis in anechoic conditions (="only the sound projected directly to the front") have a smooth sound power output (="sound in all directions, not just to the front") in total, when you measure the loudspeaker not in anechoic conditions but in-situ (meaning in your listening room, including all reflections and reverb), you should end up with a frequency response resembling a straight line that goes down 10 dB from 20 Hz to 20 kHz. This is roughly -1 dB/8ve, meaning for every doubling of frequency, the SPL goes down by 1 dB. If you want to read more about it, Floyd Toole has written a whole book about that topic. While this has basically ended the discussion as to "what is the target for a loudspeaker", it has not ended the discussion on how to get there. Note that this only applies to music. Entirely different standards apply when you start talking about movies / movie theaters. /EndQuote So Diracs standard curve is actually the most correct according to Floyd Toole. For music as the writer mentions. They all say that a fully flat frequency curve on loudspeakers in a normal room isn’t prefered. Nor is the Harman for that matter.. Although many ”like” more bass so that’s why it’s used. Not because it’s ”correct”. There are two aspects to this. The aspect of flat linear anechoic response with uniform dispersion correlating to the mildly downward sloping in-room response ... yes that's all discussed in Toole's book, but not as an EQ target curve. It's the response one would expect - presumably in a room without significant modal resonances ... and who has THAT!? - due to the fact that higher frequencies fall off while every room has some bass gain. That tilted response is NOT the "Harman Curves" that we see referenced as the +6, +8, +10. The curves with the large bass boost come from the listening tests mostly done by Olive and associates to determine listener preferred bass and treble settings for speakers and headphones. In those tests a fairly broad demographic of both trained and untrained participants listened to a very small sampling of music, which notably did not include any classical or acoustic instrumental jazz. They were asked to adjust the bass and treble to their taste and that's where the curves come from ... some people who liked more bass and less treble. And trained listeners preferred flatter response than untrained listeners. And I believe the more extreme curves came from the headphone tests, but some people use them for speakers. But I have heard some justify these curves as being related to the Fletcher-Munson type curves to compensate for our hearing at low levels ... which they are not. Note also that when Harman did the tests, they first EQ'ed the room response to a flat curve. Here's the music they used for the test:  |
|
|
|
Post by fbczar on Mar 18, 2022 13:09:13 GMT -5
Found this online.. Talking about Harman curves.. /Quote Is there such curve for speakers? target curves for loudspeakers are pretty well researched. That's in fact what the Harman Target is based on (measure a loudspeaker that performs close to the loudspeaker target curve, but measure it with a measurement head instead of a regular microphone, and then make a headphone that produces the same result when measured with that measurement head). While the exact details are subject to preference, the overall goal for a loudspeaker is: have a flat, linear frequency response when measured on-axis in anechoic conditions (="only the sound projected directly to the front") have a smooth sound power output (="sound in all directions, not just to the front") in total, when you measure the loudspeaker not in anechoic conditions but in-situ (meaning in your listening room, including all reflections and reverb), you should end up with a frequency response resembling a straight line that goes down 10 dB from 20 Hz to 20 kHz. This is roughly -1 dB/8ve, meaning for every doubling of frequency, the SPL goes down by 1 dB. If you want to read more about it, Floyd Toole has written a whole book about that topic. While this has basically ended the discussion as to "what is the target for a loudspeaker", it has not ended the discussion on how to get there. Note that this only applies to music. Entirely different standards apply when you start talking about movies / movie theaters. /EndQuote So Diracs standard curve is actually the most correct according to Floyd Toole. For music as the writer mentions. They all say that a fully flat frequency curve on loudspeakers in a normal room isn’t prefered. Nor is the Harman for that matter.. Although many ”like” more bass so that’s why it’s used. Not because it’s ”correct”. There are two aspects to this. The aspect of flat linear anechoic response with uniform dispersion correlating to the mildly downward sloping in-room response ... yes that's all discussed in Toole's book, but not as an EQ target curve. It's the response one would expect - presumably in a room without significant modal resonances ... and who has THAT!? - due to the fact that higher frequencies fall off while every room has some bass gain. That tilted response is NOT the "Harman Curves" that we see referenced as the +6, +8, +10. The curves with the large bass boost come from the listening tests mostly done by Olive and associates to determine listener preferred bass and treble settings for speakers and headphones. In those tests a fairly broad demographic of both trained and untrained participants listened to a very small sampling of music, which notably did not include any classical or acoustic instrumental jazz. They were asked to adjust the bass and treble to their taste and that's where the curves come from ... some people who liked more bass and less treble. And trained listeners preferred flatter response than untrained listeners. And I believe the more extreme curves came from the headphone tests, but some people use them for speakers. But I have heard some justify these curves as being related to the Fletcher-Munson type curves to compensate for our hearing at low levels ... which they are not. Note also that when Harman did the tests, they first EQ'ed the room response to a flat curve. Here's the music they used for the test: <button disabled="" class="c-attachment-insert--linked o-btn--sm">Attachment Deleted</button> Marc, Do you think the results from Olive hold true for dipole speakers? |
|
|
|
Post by marcl on Mar 18, 2022 13:16:44 GMT -5
There are two aspects to this. The aspect of flat linear anechoic response with uniform dispersion correlating to the mildly downward sloping in-room response ... yes that's all discussed in Toole's book, but not as an EQ target curve. It's the response one would expect - presumably in a room without significant modal resonances ... and who has THAT!? - due to the fact that higher frequencies fall off while every room has some bass gain. That tilted response is NOT the "Harman Curves" that we see referenced as the +6, +8, +10. The curves with the large bass boost come from the listening tests mostly done by Olive and associates to determine listener preferred bass and treble settings for speakers and headphones. In those tests a fairly broad demographic of both trained and untrained participants listened to a very small sampling of music, which notably did not include any classical or acoustic instrumental jazz. They were asked to adjust the bass and treble to their taste and that's where the curves come from ... some people who liked more bass and less treble. And trained listeners preferred flatter response than untrained listeners. And I believe the more extreme curves came from the headphone tests, but some people use them for speakers. But I have heard some justify these curves as being related to the Fletcher-Munson type curves to compensate for our hearing at low levels ... which they are not. Note also that when Harman did the tests, they first EQ'ed the room response to a flat curve. Here's the music they used for the test: <button disabled="" class="c-attachment-insert--linked o-btn--sm">Attachment Deleted</button> Marc, Do you think the results from Olive hold true for dipole speakers? Toole specifically says his results apply only to the conventional box speakers that he used for testing and it is unknown what would happen with dipoles, etc. Olive also used only conventional box speakers. Olive's tests were done in the Harman new listening room. Neither has tested dipoles as far as I could find in the literature. We know that dipoles interact with the room differently. So maybe that's why I am very satisfied with a flat measured response and any time I elevate the bass it sounds heavy and colored. |
|
|
|
Post by fbczar on Mar 18, 2022 13:34:39 GMT -5
Marc, Do you think the results from Olive hold true for dipole speakers? Toole specifically says his results apply only to the conventional box speakers that he used for testing and it is unknown what would happen with dipoles, etc. Olive also used only conventional box speakers. Olive's tests were done in the Harman new listening room. Neither has tested dipoles as far as I could find in the literature. We know that dipoles interact with the room differently. So maybe that's why I am very satisfied with a flat measured response and any time I elevate the bass it sounds heavy and colored. When you say “elevate the bass” do you actually elevate the bass from the Maggies or do you boost the output of the subwoofers? |
|
|
|
Post by marcl on Mar 18, 2022 14:44:45 GMT -5
Toole specifically says his results apply only to the conventional box speakers that he used for testing and it is unknown what would happen with dipoles, etc. Olive also used only conventional box speakers. Olive's tests were done in the Harman new listening room. Neither has tested dipoles as far as I could find in the literature. We know that dipoles interact with the room differently. So maybe that's why I am very satisfied with a flat measured response and any time I elevate the bass it sounds heavy and colored. When you say “elevate the bass” do you actually elevate the bass from the Maggies or do you boost the output of the subwoofers? If I elevate the bass for the Front L/R I do it with the Dirac target curve. In my configuration the Maggies are set to Large and they play full range to just below 40Hz. The room has a huge peak at 40Hz, like +15db. The L/R outputs are split so that one side goes directly to the Maggies and the other goes to the miniDSP that feeds the subs. The miniDSP applies HPF 10Hz and LPF 50Hz, 48db/octave. For nearly all music, the subs don't come into play ... or they only slightly overlap the Maggie response at the bottom end around 40Hz. When Dirac measures the L/R, it sees the combined response of the Maggies playing full range plus the subs playing below 50Hz. I apply the target curve to that measurement which extends from about 16Hz on up. So, if I were to apply a bass boost as many do below 100Hz, it would apply to the L/R output which feeds both the Maggies and the subs. What I find is the system sounds best when the measured response after applying Dirac filters is as flat as Dirac is able to make it. |
|
|
|
Post by atomic4877 on Mar 18, 2022 19:49:03 GMT -5
I ran Dirac today using some of the recommended methods that I been able to rather from here such as following the mehlau instructions posted here with the speakers 20-30 db the noise floor and the tip of having the sub 5-10 db louder on the volume calibration page. I am running a dspeaker dual core 2.0 which I ran first then did Dirac. For some reason Dirac would not produce test tones for the right sub when the dual core was connected. I ended up running the dual core then disconnecting and running Dirac. I bumped the subs up in the dual core about -5 db to get it to my overall liking for the subs. The overall results seem good, however one thing I noticed is that after running calibration is that I have to have the volume level set to -20 to get to my normal listening volume vs. -30 as normal pre-Dirac. Is this normal to have to turn the volume up? Also, what level are you supposed to use for the pink noise generator to level match the speakers? Low, medium, or high? I want to make verify the levels Dirac has set, but the default low volume even turned up to 0 seems too soft.
|
|
|
|
Post by hsamwel on Mar 18, 2022 19:51:04 GMT -5
Found this online.. Talking about Harman curves.. /Quote Is there such curve for speakers? target curves for loudspeakers are pretty well researched. That's in fact what the Harman Target is based on (measure a loudspeaker that performs close to the loudspeaker target curve, but measure it with a measurement head instead of a regular microphone, and then make a headphone that produces the same result when measured with that measurement head). While the exact details are subject to preference, the overall goal for a loudspeaker is: have a flat, linear frequency response when measured on-axis in anechoic conditions (="only the sound projected directly to the front") have a smooth sound power output (="sound in all directions, not just to the front") in total, when you measure the loudspeaker not in anechoic conditions but in-situ (meaning in your listening room, including all reflections and reverb), you should end up with a frequency response resembling a straight line that goes down 10 dB from 20 Hz to 20 kHz. This is roughly -1 dB/8ve, meaning for every doubling of frequency, the SPL goes down by 1 dB. If you want to read more about it, Floyd Toole has written a whole book about that topic. While this has basically ended the discussion as to "what is the target for a loudspeaker", it has not ended the discussion on how to get there. Note that this only applies to music. Entirely different standards apply when you start talking about movies / movie theaters. /EndQuote So Diracs standard curve is actually the most correct according to Floyd Toole. For music as the writer mentions. They all say that a fully flat frequency curve on loudspeakers in a normal room isn’t prefered. Nor is the Harman for that matter.. Although many ”like” more bass so that’s why it’s used. Not because it’s ”correct”. There are two aspects to this. The aspect of flat linear anechoic response with uniform dispersion correlating to the mildly downward sloping in-room response ... yes that's all discussed in Toole's book, but not as an EQ target curve. It's the response one would expect - presumably in a room without significant modal resonances ... and who has THAT!? - due to the fact that higher frequencies fall off while every room has some bass gain. That tilted response is NOT the "Harman Curves" that we see referenced as the +6, +8, +10. The curves with the large bass boost come from the listening tests mostly done by Olive and associates to determine listener preferred bass and treble settings for speakers and headphones. In those tests a fairly broad demographic of both trained and untrained participants listened to a very small sampling of music, which notably did not include any classical or acoustic instrumental jazz. They were asked to adjust the bass and treble to their taste and that's where the curves come from ... some people who liked more bass and less treble. And trained listeners preferred flatter response than untrained listeners. And I believe the more extreme curves came from the headphone tests, but some people use them for speakers. But I have heard some justify these curves as being related to the Fletcher-Munson type curves to compensate for our hearing at low levels ... which they are not. Note also that when Harman did the tests, they first EQ'ed the room response to a flat curve. Here's the music they used for the test: <button disabled="" class="c-attachment-insert--linked o-btn--sm">Attachment Deleted</button> Trained ears usually don’t like bloated bass. What is too much or not is mostly preference. Some are really sensitive to low end sound which can make voices sound less clear. This could be the reason trained listners prefered less bass. I read that most people that prefered bass boosting were men below 40. Women and older men above 50 generally prefered flatter bass. I would put myself somewhere in the middle. I like some bass but it needs to be clean and fast. For movies I feel differently.. Still clean but I don’t mind a few extra dBs for rhe extra real life sense of explosions or gunshots. Rumbles and distortion I can live without though. But I don’t like when the bass is so high set that, if someone closes a door or drops some keys, the whole room shakes. IMO it has to feel real, not more or less. |
|
|
|
Post by leonski on Mar 18, 2022 20:00:23 GMT -5
Clean and 'fast' is important. Call it 'accurate' if you will.
Most can't tell the difference and think that mid-bass bloat is 'deep bass'.......
Music and HT are 2 very different things. 25hz for music is FINE while movies? I have no idea, but it is lower.
the good news is that nobody knows what most 'movie effects' are supposed to sound like.......
|
|
|
|
Post by hsamwel on Mar 20, 2022 18:20:34 GMT -5
Clean and 'fast' is important. Call it 'accurate' if you will. Most can't tell the difference and think that mid-bass bloat is 'deep bass'....... Music and HT are 2 very different things. 25hz for music is FINE while movies? I have no idea, but it is lower. the good news is that nobody knows what most 'movie effects' are supposed to sound like....... Haha of course not.. We the viewers can’t know how the bass should sound in a specific movie. But I know how things sound when you drop them, even heavy things, and how deep voices sound.. I don’t want these to sound unnatural and too heavy. I just today were in a show room with a REALLY expensive setup. A Sony VPL-GTZ380 4k lazer projector viewed on a 175” sound transparent screen. The setup was a 9.5.4 on a Trinnov. .Bass was handled by 2x dual 18”, 2x single 18” and a wopping 32” subwoofer. The guy from the speaker company explained that NO 16-18” sub will do real 10-12hz even though the manufacturer tells you it will. Not in the SPL it should anyway. This 32” did it to 7hz in full SPL. It was setup to handle all low bass below 35hz. Crossed to all the 18”s which handled above 35hz.. The amplification was even more ridiculous. The smallest being the amps for atmos and surrounds at 750W each. The subs getting 2500W each if I remember correctly. The funny thing was he told us the 32” was the small model in the series. The biggest being an incredible 80”. This system sounded amazing! It had really deep and clean bass. No bloated mid bass or midrange. It was the most natural sounding system I have ever heard. |
|
|
|
Post by leonski on Mar 20, 2022 20:33:21 GMT -5
Music, is in no case is below 16hz (a low pedal tone on pipe organ) I don't kwow of any stringed instrument or human powered
instrument to go that low......or lower.....except Tim Storms.
In movie effects?
We ALL have ideas what stuff is 'supposed' to sound like.....using your example, 'when dropped'
But ultimately we are at the mercy of Foley / Sound designer / and whoever does the 'mix'.......
A deep voice is another matter......maybe. FAR more reference exists to Judge. And that won't go anywhere NEAR 30hz.
I wonder how Many 20 amp circuits it takes to make your 'reference' system work well.....?
There is only ONE Tim Storms, the lowest note not even your 32" sub will touch.....
Just for curiosity? I wonder what the real upper limits are for subs 32" and UP.
I would LOVE to hear a recording of a pipe organ sound even CLOSE to the way I heard it LIVE in
Brugge Belgium when I walked in on a practice session......
|
|
|
|
Post by marcl on Mar 22, 2022 7:28:14 GMT -5
Dirac Live 3.2.32022-03-21FixesUpdate URLs for server region: China.When asked in the AVSForum thread if they changed anything else, Flavio replied "not that I know of". I'm likely to install it because chances are every release has SOME kind of fix in it that a company either won't admit to ... or a developer won't admit to. 
|
|
|
|
Post by marcl on Mar 22, 2022 9:21:39 GMT -5
Dirac Live 3.2.32022-03-21FixesUpdate URLs for server region: China.When asked in the AVSForum thread if they changed anything else, Flavio replied "not that I know of". I'm likely to install it because chances are every release has SOME kind of fix in it that a company either won't admit to ... or a developer won't admit to.  I installed 3.2.3 and loaded a previous 3.2.2 project. It loaded quickly and did not recalculate filters. |
|
|
|
Post by leonski on Mar 22, 2022 17:18:36 GMT -5
Any of you Bass Guys even BOTHER to look up Tim Storms?
Didn't think so.....
|
|
|
|
Post by marcl on Mar 23, 2022 7:55:30 GMT -5
Any of you Bass Guys even BOTHER to look up Tim Storms? Didn't think so..... You can't "BOTHER" to look up someone unless you have ever heard of them ... but having done so, I did, and so ... |
|
|
|
Post by atomic4877 on Mar 23, 2022 13:04:56 GMT -5
Just wondering, how many points of measurement does everyone use?
|
|
|
|
Post by marcl on Mar 23, 2022 13:16:42 GMT -5
Just wondering, how many points of measurement does everyone use? At least 9 ... usually 11 or 13. The key is that the algorithm doesn't literally average over the measurements just to cover a broad area. It analyzes the measurements to see which peaks and dips are minimum phase vs non-minimum phase. And that determines how the correction filters are configured. |
|
|
|
Post by atomic4877 on Mar 23, 2022 13:20:56 GMT -5
Just wondering, how many points of measurement does everyone use? At least 9 ... usually 11 or 13. The key is that the algorithm doesn't literally average over the measurements just to cover a broad area. It analyzes the measurements to see which peaks and dips are minimum phase vs non-minimum phase. And that determines how the correction filters are configured. I have a 3 seat sofa and have run calibration twice using the 13 point, but unsure if this is overkill for such a centralized listening area. |
|
|
|
Post by marcl on Mar 23, 2022 13:23:52 GMT -5
At least 9 ... usually 11 or 13. The key is that the algorithm doesn't literally average over the measurements just to cover a broad area. It analyzes the measurements to see which peaks and dips are minimum phase vs non-minimum phase. And that determines how the correction filters are configured. I have a 3 seat sofa and have run calibration twice using the 13 point, but unsure if this is overkill for such a centralized listening area. If the sofa back is high you might skip the two lower rear measurements. But overall giving Dirac more data makes for a more accurate result. |
|
|
|
Post by leonski on Mar 23, 2022 20:54:30 GMT -5
Are you REQUIRED to do the measurements in a certain order.......? And 'tell' the software in advance how many measurement points you will use?
|
|
