Jump to content

Recommended Posts

Posted

ozmillsy,

I would be interested to see a comparison in the output of the RTA and a frequency decomposition (using cooledit etc) of 0.5 to 1 sec of the music that created the RTA output.

What is the frequency bin size of the RTA? Maybe it's just filtering that causes a display at 20Hz

Otherwise my suspicion is that it is room (or some other) nonlinearities that create the 20Hz output that you measure. Since for classical music there are no instruments that can produce 20Hz - you need an approx 8.3m tube (neglecting end effect, as I can't remember the correction factor).

Typical bass guitars can produce 41Hz with the lowest E1 string and 30Hz if they have a B1 string. Of course electronics can be used to make lower notes. But this wouldn't be typical for Jazz or most pop.

  • Replies 115
  • Created
  • Last Reply

Top Posters In This Topic

Posted (edited)
  OzMillsy said:
Drizt, as you know, I've been chastised on DTV over this subject in the past - and by some of the names in your OP I might point out. So I've been quietly reading this thread with interest.

I've experimented with various sub integration techniques, and it just isnt easy to do it properly on a 2ch analog system. It takes time and patience. Actually, it's not that hard to do the traditional technique of passing the signal to a sub off a second set of outputs, and dial up the sub to the roll off point of your mains. But that isnt doing it properly IMO. The active approach of removing the sub freq's from the mains, has benefits already stated in this thread. For me, it's also about Mid-bass performance, which is very much improved when the mains are not trying to deliver sub-bass. But to walk the active sub integration path, can take commitment. It's funny when some people want to improve bass and implement a sub, but are unwilling to even buy a SPL meter and run some test tones.

Anyway, there is a common theme emerging in recent times, that room gain can compensate for speaker roll off and provide adequate bass without the need for a sub. I dont subscribe to this theory. Room gain wont deliver 20hz bass if the speakers are not delivering 20hz bass. There is also a longer term theme, that instruments just dont go to 20hz. I've had a Behringer RTA on my system for some time now (out of loop), and just about everything I play back (virtually all styles), has bass right down to 20hz - whether it's got pipe organs or not.

It definetly does make a difference to my listening experience, to properly playback music in full range. The hard part is to get the system setup well. EQ helps, but isnt always necessary. Here is the in-room response of my left speaker/sub system, no room compensation EQ has been applied.....

Lch Freq response

Such a smooth response is unusual. Unfortunately the right channel doesnt look so pretty (it's too close to a wall, that I cant do anything about), but I still dont feel the need to apply any EQ.

Rch Freq response

Oz,

What settings and equipment are you using to produce those graphs? They do seem quite unusual. A more normal result AFTER EQ would look more like this

Edited by Gior
Posted (edited)
  Drizt said:
That Lch plot looks way too good to be true. Is it with 1/1 smoothing? Is it from the listening position?

as you know mate, the electronics in the sgrs have an eq circuit. The speakers and the sgr subs have been smoothed during their build.

My left channel is isolated from all walls (and I have cathedral ceilings). My right channel isn't.

Measured from the listening position (ear height).

Edited by OzMillsy
Posted
  OzMillsy said:
as you know mate, the electronics in the sgrs have an eq circuit. The speakers and the sgr subs have been smoothed at the during their build.

My left channel is isolated from all walls (and I have cathedral ceilings). My right channel isn't.

Measured from the listening position (ear height).

Doesn't matter what speakers are used your at the mercy of the room. I wouldn't ever expect to see a graph like the one you posted if its a raw response from the listening position. You be lucky to get a graph like that from 10cm away from the speaker in the middle of a paddock.

Posted (edited)
  Spearmint said:
So this is about the use of an AVR as a 2ch pre, and not so much about sub integration into a 2ch setup?

fark no, it has ALWAYS been about sub integration!!

My position has always been that 2ch analog pres deliver superior SQ, but that sub integration becomes an important aspect to consider and solve (if bass is important to you, as it is to me).

The fact is, that active sub integration is just so extremely simple and convenient on avr's. Modern systems have come forward in leaps and bounds for bass management and other features. That it really is a doddle to setup your system well, and integrate your sub in.

It's not as easy to do it in 2ch analog space.

I totally understand why some would choose to simplify their systems and drop down to 1 volume control device - rather than run 2.

Edited by OzMillsy
Posted
  Drizt said:
Doesn't matter what speakers are used your at the mercy of the room. I wouldn't ever expect to see a graph like the one you posted if its a raw response from the listening position. You be lucky to get a graph like that from 10cm away from the speaker in the middle of a paddock.

My R channel inadequacies support the validity of the L channel. You are welcome to come an measure it yourself, next time you are in my area.

Posted (edited)
  OzMillsy said:
My R channel inadequacies support the validity of the L channel. You are welcome to come an measure it yourself, next time you are in my area.

Oz, what level of smoothing are you running? Can you show us a full 20hz-20khz sweep with no smoothing by any chance or have you deleted those measurements?

Edited by Gior
Posted
  Brad Cl said:
ozmillsy,

I would be interested to see a comparison in the output of the RTA and a frequency decomposition (using cooledit etc) of 0.5 to 1 sec of the music that created the RTA output.

What is the frequency bin size of the RTA? Maybe it's just filtering that causes a display at 20Hz

Otherwise my suspicion is that it is room (or some other) nonlinearities that create the 20Hz output that you measure. Since for classical music there are no instruments that can produce 20Hz - you need an approx 8.3m tube (neglecting end effect, as I can't remember the correction factor).

Typical bass guitars can produce 41Hz with the lowest E1 string and 30Hz if they have a B1 string. Of course electronics can be used to make lower notes. But this wouldn't be typical for Jazz or most pop.

Brad, my behringer deq2496 takes a digital feed from my transport and measures the audio raw.

My RTA is not measuring the audio output of my system via a mic.

I generally don't listen to Classical. I listen to pop, electronica, and jazz (most often). And I have plenty of jazz that is bass heavy, and is generating info down to 20hz.

Posted
  OzMillsy said:
Brad, my behringer deq2496 takes a digital feed from my transport and measures the audio raw.

My RTA is not measuring the audio output of my system via a mic.

I generally don't listen to Classical. I listen to pop, electronica, and jazz (most often). And I have plenty of jazz that is bass heavy, and is generating info down to 20hz.

Huh?

Posted
  Gior said:
Oz, what level of smoothing are you running? Can you show us a full 20hz-20khz sweep with no smoothing by any chance or have you deleted those measurements?

I don't apply any in room eq (smoothing).

Each channel is a 4 way active speaker/sub system, and the electronics built in have an eq circuit to ensure they peform perfectly flat (at the factory). I'm lucky that my L channel delivers a plot like that in my room. My R channel doesn't.

I didn't run a full freq scan, as I was only interested in that region. But will do a full freq scan, if you're asking. I can only run test tones when I'm home alone, as it drives the wife/kid nuts. So give me some time to get back to you on that.

Posted
  OzMillsy said:
I don't apply any in room eq (smoothing).

Each channel is a 4 way active speaker/sub system, and the electronics built in have an eq circuit to ensure they peform perfectly flat (at the factory). I'm lucky that my L channel delivers a plot like that in my room. My R channel doesn't.

I didn't run a full freq scan, as I was only interested in that region. But will do a full freq scan, if you're asking. I can only run test tones when I'm home alone, as it drives the wife/kid nuts. So give me some time to get back to you on that.

But you ARE using a mic of some sort to measure this at listening position right? Your last statement to Brad got me very confused lol

Posted
  :) said:
drew the velo sms-1 might fit the bill, but no it isnt cheap either - local rrp might be $1500 ? or cheaper dont know it went for as high as $1800 some years ago when our dollar was in the dumps. Cheaper overseas, $500 us or so from outlaw. and they do pop up now and then 2nd hand.
There's one here on the forum, owner just stuck it on Ebay for $699...........originally after $750.

So if Drew's after one, it's here:

Velodyne Sms-1 Parametric Equalizer For Sale

cheers

Posted
  Gior said:
Huh?

Are you not familiar with the features of the DEQ2496 ?

I run an optical cable from my transport into it, and it can be configured to perform real time analysis on the digital data.

My transport also has a coaxial cable running to a dac. So, the rta is performed out of loop. I like looking at the bouncing info. :)

Posted
  MACCA350 said:
There's one here on the forum, owner just stuck it on Ebay for $699...........originally after $750.

So if Drew's after one, it's here:

Velodyne Sms-1 Parametric Equalizer For Sale

cheers

The Dude's unit was in mint condition. When I was over calibrating the Integra 9.9 he bought off me, we disconnected it then as it really wasn't needed anymore as the Integra did a great job with smoothing his sub response. I only used 3 measuring points due to time, but the results were pretty impressive on his system. Especially since with the Audyssey Pro you have the option of demonstrating the difference between Audyssey and no Audyssey with a click of a mouse button. It's always amazing watching the soundstage collapse into the speakers when you switch Audyssey off but open up completely when you turn it back on. You cease to notice the speakers and just get immersed into the movie instead.

Posted
  Gior said:
But you ARE using a mic of some sort to measure this at listening position right? Your last statement to Brad got me very confused lol

Sorry. Real time analysis of music is not using a mic.

In room freq response is measured by REW using a behringer mic/pre rig hooked up to my notebook.

Posted
  OzMillsy said:
Sorry. Real time analysis of music is not using a mic.

In room freq response is measured by REW using a behringer mic/pre rig hooked up to my notebook.

Ah, that clears up the confusion a bit...we have a mic in the equation now lol. I was confused with the multiple posts I was looking at.

Anyway, your graphs should be displaying significantly more combing especially if you are not smoothing the graph data at all. I'd double check your setup again on next measurements as your graphs really don't look quite right (yes, even the RIGHT graph). If you followed my link, you'd see an unsmoothed response from REW measured with a calibrated Behringer ECM 8000 with an Art USB Preamp. Regardless of if I use REW other software such as Fuzzmeasure Pro, I will consistently get that graph as that is a typical room response. Even in a heavily treated room like my main HT, the blue graph STILL shows combing even though it is significantly less than that green graph. RT60 times on my main room are in the 0.2 second range to give you an idea of how much absorption I'm running there.

No matter how well EQ'd the SGR speakers are in the factory, by the time they get to your room, they will look like my REW graphs more or less as that is just what rooms ALWAYS do to the sound.

Posted
  Gior said:
Anyway, your graphs should be displaying significantly more combing especially if you are not smoothing the graph data at all. I'd double check your setup again on next measurements as your graphs really don't look quite right (yes, even the RIGHT graph). If you followed my link, you'd see an unsmoothed response from REW measured with a calibrated Behringer ECM 8000 with an Art USB Preamp. Regardless of if I use REW other software such as Fuzzmeasure Pro, I will consistently get that graph as that is a typical room response.

No worries, I'll double check it. Are you in Sydney? If so, I dont mind having someone over to double check what I'm doing and how I'm measuring.

Posted

OzMillsy,

would you mind generating a software frequency analysis of some bass heavy pop or jazz. As an example I just looked at 'We Cry' by the script, which has a reasonable bass line. It is approx 35dB down from the peak at 20Hz. to get reasonable info at 20Hz, you need to select over 30sec of music for analysis.

Gior,

0.2s RT60 time is really low,does it sound a bit dead?

At present the RT60 time in my room is 0.4-0.5 for freq higher than 300Hz. I am in the process of making some treatments using ultratel.

What I would like to know is your RT60 time at less than 100Hz.

Mine gets as high as 0.7s - hence I need some bass traps! After which I hope that the submersive will tighten up somewhat.

Posted
  Brad Cl said:
OzMillsy,

would you mind generating a software frequency analysis of some bass heavy pop or jazz. As an example I just looked at 'We Cry' by the script, which has a reasonable bass line. It is approx 35dB down from the peak at 20Hz. to get reasonable info at 20Hz, you need to select over 30sec of music for analysis.

Gior,

0.2s RT60 time is really low,does it sound a bit dead?

At present the RT60 time in my room is 0.4-0.5 for freq higher than 300Hz. I am in the process of making some treatments using ultratel.

What I would like to know is your RT60 time at less than 100Hz.

Mine gets as high as 0.7s - hence I need some bass traps! After which I hope that the submersive will tighten up somewhat.

Hi Brad,

RT60 times at the best of times aren't the most useful figure in small rooms. I'd ignore them below 200hz as decay times would be the more accurate there. If using RT60 in small rooms, look at the figures above 200hz. For HT, 0.2 - 0.3 seconds is about right as you have a multitude of speakers. I run 7.1/2 now but am going to 9.2. I REALLY don't need the room adding any life to the sound. 0.4-0.5 is relatively common in an untreated room that is acoustically decent. If you have a look at the green graph of my second room that has no acoustic treatment, you'll see it averaging about the 0.5 mark. All that stuff about rooms sounding dead really is just talk. My main HT room sounds fine.

If you want to reducing 'ringing' with your subs, bass trapping can definitely help. However the lower the frequency, the hard to trap simply due to the size of the traps required. I've found good positioning can help minimise that 'ringing' a lot too and tighten up your impulse response heaps.

Posted

Gior,

I'm a little confused, RT60 is simply the time it takes the sound to decay 60dB. So it is a decay time.

However, it does raise an interesting point. I don't think my noise floor is 60dB below the 75dBC measurement level (who would have a 15dBC noise floor!).

SO I might have to look into how REW calculates RT60 from the impulse response, but unless something tricky is going on, I would imagine that the spectral decay plot is more reliable than the RT60 plot on REW.

So far I have moved the sub half way around the room, observing the RTA plot as I do so. I now have a reasonable freq response with the sub at the front of the room. I had to cross the sub over at 120Hz to smooth out a 110Hz dip from the mains.

What do you cross your submersive at? I am worried that I am getting to the point where I can audibly localise the sub.

I won't put the plot on the forum until I get a calibrated mic, as I don't trust it below 40Hz and above 5kHz.

Posted
  Brad Cl said:
Gior,

I'm a little confused, RT60 is simply the time it takes the sound to decay 60dB. So it is a decay time.

However, it does raise an interesting point. I don't think my noise floor is 60dB below the 75dBC measurement level (who would have a 15dBC noise floor!).

SO I might have to look into how REW calculates RT60 from the impulse response, but unless something tricky is going on, I would imagine that the spectral decay plot is more reliable than the RT60 plot on REW.

So far I have moved the sub half way around the room, observing the RTA plot as I do so. I now have a reasonable freq response with the sub at the front of the room. I had to cross the sub over at 120Hz to smooth out a 110Hz dip from the mains.

What do you cross your submersive at? I am worried that I am getting to the point where I can audibly localise the sub.

I won't put the plot on the forum until I get a calibrated mic, as I don't trust it below 40Hz and above 5kHz.

Yup you are correct with REW - it's the Spectral Decay plot that is much more useful than RT60.

Why don't you just move the sub to your primary listening position and walk around with the mic instead? It's much easier than moving the sub all over the room.

I cross all my speakers at 80hz now as I can't personally localise that. All my speakers extend below that level so Audyssey filters have been employed on all of them till their -3db in room cutoff points. The sub handles the sub 80 frequencies just fine :)

Posted (edited)
  Brad Cl said:
OzMillsy,

would you mind generating a software frequency analysis of some bass heavy pop or jazz. As an example I just looked at 'We Cry' by the script, which has a reasonable bass line. It is approx 35dB down from the peak at 20Hz. to get reasonable info at 20Hz, you need to select over 30sec of music for analysis.

30's? I can do better than that. On the DEQ2496 I can tell it to preserve the peaks.

So, I ran an album through the DEQ in full, while it was real time analysing the digital input info and preserving the peaks of all freq's.

The album: The Best of Andrea Bocelli - Vivere.

At the end of this album, this is what the captured peaks looked like-

Freq Analysis (peaks)

As you can see, on this album there are peaks occuring all the way down to 20hz and all the way up to 20khz. This album is NOT what I call bass heavy at all.

Choose to believe it or not, but the info is there in the digital source, dude! Most systems cant produce it, many enthusiasts choose to dismiss it.

Edited by OzMillsy
Posted
  OzMillsy said:
No worries, I'll double check it. Are you in Sydney? If so, I dont mind having someone over to double check what I'm doing and how I'm measuring.

I might do that when I get some time after I get back into the country.

Posted
  OzMillsy said:
30's? I can do better than that. On the DEQ2496 I can tell it to preserve the peaks.

So, I ran an album through the DEQ in full, while it was real time analysing the digital input info and preserving the peaks of all freq's.

The album: The Best of Andrea Bocelli - Vivere.

At the end of this album, this is what the captured peaks looked like-

Freq Analysis (peaks)

As you can see, on this album there are peaks occuring all the way down to 20hz and all the way up to 20khz. This album is NOT what I call bass heavy at all.

Choose to believe it or not, but the info is there in the digital source, dude! Most systems cant produce it, many enthusiasts choose to dismiss it.

Just realised that we have a DEQ2496 at work. I will do some experiments.

The point that I was really ttrying to get to is that what people perceive as bass is not as low frequency as you think. For example kick drums are mainly over 100Hz. 70-80Hz is the rumbly sound used in movies a lot and 30-40Hz is really deep bass. 20Hz or less is felt, rather than heard.

The lower the bass, the louder it needs to be for us to perceive it as the same volume as higher frequencies. Apart from placement flexibility, I think that is the reason subs can produce better bass - they can do it louder than mains.

  • Recently Browsing   0 members

    • No registered users viewing this page.




×
×
  • Create New...
To Top