Eq'ing for Bass
Hi, about eq'ing "head in a vise" problem.... and a better solution DOES exist....
a few years ago the guys at Harman produced a white paper showing the effects of using several subs in a room, averages out to just ONE main node that you then eq very easily.
The simplest implementation is to use 2 subs, one at each end of the room, then measure the resulting peak from the main node. A good parametric like Behringer DEQ2496 works perfectly for this (I use mine in digital in and out with a good outboard DAC) and it works really well.
So, to add my opinions to those already expressed:
1) Yes, the better your bass drivers are, the less you will notice in the way of room coloration.
But you will still notice it you have one or two room nodes with peaks of +15dB... and such peaks are a LOT more common than you may think.
2) Two subs are better than one. This is not just my opinion, it is based on some pretty solid science. Check the HK white papers. Hmm, it's nice to see some actual research and extensive listening to validate the theory.
3) If you want to gild the lily, go with 4 subs. Four are better than two, especially if you like to keep stereo channels into the bass.
I do, as some recordings actually have stereo low frequency information.... and beat frequencies DO exist in nature... even if many commercial audio engineers continue to foster the myth that low bass is just fine with mono.
4) And, even with 5 way speakers, in some rooms you are going to run into ferociously big peaks. Adjusting your speakers to "fit" the room can work well, but try a pair of good subs with a parametric eq and I'll bet that you will be much happier with the results.
Best regards and may the Good Bass be with you.
Hi, about eq'ing "head in a vise" problem.... and a better solution DOES exist....
a few years ago the guys at Harman produced a white paper showing the effects of using several subs in a room, averages out to just ONE main node that you then eq very easily.
The simplest implementation is to use 2 subs, one at each end of the room, then measure the resulting peak from the main node. A good parametric like Behringer DEQ2496 works perfectly for this (I use mine in digital in and out with a good outboard DAC) and it works really well.
So, to add my opinions to those already expressed:
1) Yes, the better your bass drivers are, the less you will notice in the way of room coloration.
But you will still notice it you have one or two room nodes with peaks of +15dB... and such peaks are a LOT more common than you may think.
2) Two subs are better than one. This is not just my opinion, it is based on some pretty solid science. Check the HK white papers. Hmm, it's nice to see some actual research and extensive listening to validate the theory.
3) If you want to gild the lily, go with 4 subs. Four are better than two, especially if you like to keep stereo channels into the bass.
I do, as some recordings actually have stereo low frequency information.... and beat frequencies DO exist in nature... even if many commercial audio engineers continue to foster the myth that low bass is just fine with mono.
4) And, even with 5 way speakers, in some rooms you are going to run into ferociously big peaks. Adjusting your speakers to "fit" the room can work well, but try a pair of good subs with a parametric eq and I'll bet that you will be much happier with the results.
Best regards and may the Good Bass be with you.
The distortion issue
Oh, forgot to mention that the dual sub w. main node eq approach is a subtractive approach which does not usually require any boost at all, only a notch filter specifically tuned to the amplitude and Q of the main node.
As such, this practice generally REDUCES the higher order distortion that would be found when pushing a driver with more power.
So you get the best of both worlds. Flatter bass, less distortion. More power, more output. Less room sound. All positives.
And yes, higher cost. When $$ permits, i will be putting in 4 servo subs like the Rythmik.com units
And that's all for today. Sleep well.
Oh, forgot to mention that the dual sub w. main node eq approach is a subtractive approach which does not usually require any boost at all, only a notch filter specifically tuned to the amplitude and Q of the main node.
As such, this practice generally REDUCES the higher order distortion that would be found when pushing a driver with more power.
So you get the best of both worlds. Flatter bass, less distortion. More power, more output. Less room sound. All positives.
And yes, higher cost. When $$ permits, i will be putting in 4 servo subs like the Rythmik.com units
And that's all for today. Sleep well.
Hi Shawn,
I read my last post to you and wanted to add that you have a great homepage and some great products. There is even nothing wrong with your digial out board and it is an improvement for your application connecting an additional processor to the DCX. The only thing I wanted to say that for those of us who have no further processor it's better to stay with the existing DACs because of jitter.
Merry Christmas,
Frank
I read my last post to you and wanted to add that you have a great homepage and some great products. There is even nothing wrong with your digial out board and it is an improvement for your application connecting an additional processor to the DCX. The only thing I wanted to say that for those of us who have no further processor it's better to stay with the existing DACs because of jitter.
Merry Christmas,
Frank
No, the drivers having to excurse further because of EQ to provide the desired response.janneman said:
Well, colour me different Jan, I have never heard an implementation that was well worthwhile wrt room EQ.
Brett said:
That's possible, of course. My own room really needs it because of its shape and large height. As I said, YMMV.
Jan Didden
It comes back to that old recording studio design rule, you can't fix time domain problems in the frequency domain. However, as home users, most of us are not lucky enough to be able to spend tens of thousands on room design and treatment, but with judicious use of what you can get away with any EQing can be minimised, and that's my route... 😉
"It comes back to that old recording studio design rule, you can't fix time domain problems in the frequency domain."
Unless it is a minimum phase problem... like room resonances and such in the bass.
Shawn
Unless it is a minimum phase problem... like room resonances and such in the bass.
Shawn
I have completed Frank's mod and have some details up on my website
Frank's mod in Ergo's DCX
I also made a history page with only the biggest milestone tweaks that I have dome to my DCX. Should be pretty interesting to check. I was even suprised myself how many things I have tried. Not all of them got to stay and many have changed over time.
Ergo's DCX history
Regards,
Ergo
Frank's mod in Ergo's DCX
I also made a history page with only the biggest milestone tweaks that I have dome to my DCX. Should be pretty interesting to check. I was even suprised myself how many things I have tried. Not all of them got to stay and many have changed over time.
Ergo's DCX history
Regards,
Ergo
Still I didn't understand the benefit of a 5-way speaker. With 3 drivers and crossing points at about 100-200Hz and 1.5-2.5kHz it's not a problem to support the full sonical range from 20Hz to 20kHz without increased distortion and 1-2 EQs per driver only. There is even no need for a dedicated sub.
Looking at the list of 5 drivers I have the vague idea that sound pressure is more important than really low distortion?
BTW, how do you get 20Hz from a 3015LF without intensively EQing if possible at all? Could it be that you are starting with the wrong part?
Frank
Looking at the list of 5 drivers I have the vague idea that sound pressure is more important than really low distortion?
BTW, how do you get 20Hz from a 3015LF without intensively EQing if possible at all? Could it be that you are starting with the wrong part?
Frank
Hi Ergo,
Always when I've submitted a post there is another one who has been faster. Probably I should improve my writing speed.
I had a look at your listening room http://esken.net/ergo/blog/?cat=4 That's really, really very impressive!!! Discussions about EQing bass most probably could be stopped totally with such a room.
But also your measurements are very impressive. I assume the pre mod measurements are based on a Tent clock and CS8420? Very interesting would be a comparison to the original Behringer crystal. I assume that causes the main part of the sonical difference.
The used low jitter clock oscillators (Tent and mine) should be similar (~0.5ps RMS 12kHz-20MHz). Also low noise power supplies for the oscillators should be similar. A certain improvement should be caused by better cabling.
Regarding the different SRCs (CS8420 and AD1896) I think it’s difficult to measure real differences although dynamic range and THD of the CS420 is (128 and 117dB) and that of the AD1896 is (132 and 123 dB). That’s because there is at least a DAC (AK4393) (116 and 97 dB) in the row of measurement.
But obviously at certain input frequencies (at least at 48 kHz) there is a real improvement caused by SRC. Quality of conversion is based on conversion rate. The AD1896 seem to have less problems with different rates.
What should have gone completely is this intermittent dull sound problem caused by the CS8420.
Regarding the measurement setup I do not understand so far how jitter and THD of the measurement AD board are compensated? To my understanding the jitter diagram shows jitter and THD of both that of the DCX and that of the AD board all together? Regarding jitter I think you are comparing the two clock oscillators (and add DAC error)? But which one is the better clock oscillator?
Frank
Always when I've submitted a post there is another one who has been faster. Probably I should improve my writing speed.
I had a look at your listening room http://esken.net/ergo/blog/?cat=4 That's really, really very impressive!!! Discussions about EQing bass most probably could be stopped totally with such a room.
But also your measurements are very impressive. I assume the pre mod measurements are based on a Tent clock and CS8420? Very interesting would be a comparison to the original Behringer crystal. I assume that causes the main part of the sonical difference.
The used low jitter clock oscillators (Tent and mine) should be similar (~0.5ps RMS 12kHz-20MHz). Also low noise power supplies for the oscillators should be similar. A certain improvement should be caused by better cabling.
Regarding the different SRCs (CS8420 and AD1896) I think it’s difficult to measure real differences although dynamic range and THD of the CS420 is (128 and 117dB) and that of the AD1896 is (132 and 123 dB). That’s because there is at least a DAC (AK4393) (116 and 97 dB) in the row of measurement.
But obviously at certain input frequencies (at least at 48 kHz) there is a real improvement caused by SRC. Quality of conversion is based on conversion rate. The AD1896 seem to have less problems with different rates.
What should have gone completely is this intermittent dull sound problem caused by the CS8420.
Regarding the measurement setup I do not understand so far how jitter and THD of the measurement AD board are compensated? To my understanding the jitter diagram shows jitter and THD of both that of the DCX and that of the AD board all together? Regarding jitter I think you are comparing the two clock oscillators (and add DAC error)? But which one is the better clock oscillator?
Frank
Great work Ergo!
It would be really great if you could post your measurement set up - pictures and explanation. I think that will help in understanding, as well as give a chance to be duplicated by someone else.
I was checking your new web site and is really nice. Just as a side note, I wasn't able to open room measurement from your site. I am on the Mac, and I have checked Firefox and Safari.
It would be really great if you could post your measurement set up - pictures and explanation. I think that will help in understanding, as well as give a chance to be duplicated by someone else.
I was checking your new web site and is really nice. Just as a side note, I wasn't able to open room measurement from your site. I am on the Mac, and I have checked Firefox and Safari.
I updated my page with a graph of measurement setup.
I used ARTA software for analysis and either ARTA or Foobar (for jitter signal) to play the source signal.
The signal was feed to DCX through the Coax SPDIF out from the Echo Layla24 soundcard.
The analog signal from DCX was feed to analog input of the Layla24 soundcard.
Thus both the source signal (SPDIF) and the ADC process had the same master clock - the clock inside the Layla24.
**************
I have a stock DCX2496 at work - I just forgot to bring it home for holidays. So I can do the comparison of Frank's mod against stock unit in two weeks or so.
Ergo
I used ARTA software for analysis and either ARTA or Foobar (for jitter signal) to play the source signal.
The signal was feed to DCX through the Coax SPDIF out from the Echo Layla24 soundcard.
The analog signal from DCX was feed to analog input of the Layla24 soundcard.
Thus both the source signal (SPDIF) and the ADC process had the same master clock - the clock inside the Layla24.
**************
I have a stock DCX2496 at work - I just forgot to bring it home for holidays. So I can do the comparison of Frank's mod against stock unit in two weeks or so.
Ergo
Attachments
AR2 - I know the problem with some of my pages. These older pages have some mistakes in HTML code even though those were auto created by WordPress - I just have not had time to fix those 🙁 I will try to do it soon though.
Ergo
Ergo
ergo said:
No worries, just wanted to give you feedback. By the way have you received package?
I just had to comment on the below statement... A good clock has low jitter in the "audible" range - that would be 10hz-30khz approx. Mhz is of cource compleatly unintresting. The range indicated below is not enough! Its the low freq range say 20-500hz that is the tricky one!!
oettle said:
Hi Ergo,
Looking at the measurement setup I have no idea to compensate AD errors of the soundcard. Normaly you shortcircuit analog out and analog in of the soundcard and compensate the errors to a certain amount with a calibration routine. But now there is digital output used.
Frank
Looking at the measurement setup I have no idea to compensate AD errors of the soundcard. Normaly you shortcircuit analog out and analog in of the soundcard and compensate the errors to a certain amount with a calibration routine. But now there is digital output used.
Frank
Hi TNT,
My writing has been too slow again.
The jitter performance for the clock oscillators is 3 ps for the Tent one and 2.6 ps for the one I'm using RMS in the frequency range from 10 Hz to 20 MHz. So there is no real difference.
Frank
My writing has been too slow again.
The jitter performance for the clock oscillators is 3 ps for the Tent one and 2.6 ps for the one I'm using RMS in the frequency range from 10 Hz to 20 MHz. So there is no real difference.
Frank
Did you even bother to look at the 3015LF TH design? I'd bet no. If you can get 100-2.5k from a single driver at the SPL and distortion limits I want, then show me. I prefer not to use any EQ if I can. Expecting it means you're using the wrong part.oettle said: