Using ETF5 -- Some fundamental Qs

Discussion in 'Archived Threads 2001-2004' started by Ken Woodrow, Jan 7, 2002.

  1. Ken Woodrow

    Ken Woodrow Stunt Coordinator

    Joined:
    Feb 12, 2001
    Messages:
    113
    Likes Received:
    0
    I recently purchased the ETF5 analysis software for use in calibrating my system. I've just started exploring its capabilities, but have some fundamental questions about how to best use it in my HT system. The tricky part is that I have a HTPC as a source and I have been trying to use its sound card for the ETF analysis. The sound card is a M-Audio Delta 410 with 8 analog outs, 2 analog ins, and S/PDIF in/out. See my sig for the full system configuration.

    I'd like to be able to measure the frequency response of each speaker in turn as well as one speaker and the sub at the same time. My problem is, how do I get the software-generated test-tone to output to both the speaker and sub at the same time? When the sound is output to the analog outs from the card, it goes only to a single speaker. Is there any way to have the software-generated test tone output through the digital output of the card to my receiver? That way, the receiver will handle the bass x-over and sound will be output to both the speaker and sub.

    The hookup I've been using for the tests is:

    L channel OUT to L channel IN

    R channel IN to microphone

    R channel OUT to receiver

    I've read through all of online tutorials, but it seems like I am missing something fundamental.

    Any suggestions? BruceD?

    As a side note, I started out using the Rat Shack SPL meter as a mic (w/o correction), but quickly realized that I would be better off with the calibrated mic and preamp sold by ETF. The SPL graphs of my front speakers rolled off sharply at 12,000 khz, which I believe is due to the microphone, not the speakers!
     
  2. BruceD

    BruceD Screenwriter

    Joined:
    Apr 12, 1999
    Messages:
    1,220
    Likes Received:
    0
    I will give you a quick idea on my setup and a suggestion.

    I use an external electronic xover between mains (set to Large) and sub (i.e. I say SUB=NO in my processor), so for me, getting the signal to go to one or both is no problem.

    Primarily I just plug the ETF output signal cable into the amp input or xover input for the speaker(s) I want to test.

    ETF doesn't support digital output SP/DIF.

    You could attach a Y cable to end of the ETF output signal cable and allow it to feed any two analog inputs on the receiver (main right and sub) for testing these at the same time.
     
  3. Ken Woodrow

    Ken Woodrow Stunt Coordinator

    Joined:
    Feb 12, 2001
    Messages:
    113
    Likes Received:
    0
    BruceD:

    Thanks for the suggestions -- that pretty well answers my questions, particularly about the digital out capabilities of ETF. I need to check the specs of my receiver (Sony DB930) to see whether it executes bass management on any of its analog inputs (for example, if I were to connect the sound card to the CD inputs instead of the 5.1 inputs). If not, then I'll have no way of graphing the response of both the sub and speaker unless I utilize a different method of bass management.

    Eventually, I anticipate that software will be written for the HTPC that will provide bass management. Until then, I'm stuck using the receiver.

    I do have a Berhinger Feedback Destroyer that I could use as an external crossover, but I had planned to use it for bass EQ on my LFE channel, with the receiver providing the crossover and bass management for each of the main channels. But perhaps I can use it as an outboard crossover for the front two channels . . .

    Bruce -- what are you using for an outboard crossover? Do you use it for all four main channels, or just the front R & L?

    Thanks,

    Ken
     
  4. BruceD

    BruceD Screenwriter

    Joined:
    Apr 12, 1999
    Messages:
    1,220
    Likes Received:
    0
    Here are some of my thoughts, although I have a rather different setup:
    1) ETF lets you stack multiple responses in the same graph, so you could measure both (main and sub) independently (best way actually) and see how they each interact with room boundaries (frequency bass peaks and nulls) for best overall frequency response. This may actually lead you to the best bass crossover frequency as well.
    2) I use a Marchand XM-9L electronic crossover here on just my main L&R speakers and sub. So mains=large, center=small, surrounds=small, sub=no. The xover also allows me to sum the L&R bass input into a single mono output bass signal. My mains have a -3dB of 32Hz, and I use a 60Hz crossover which is almost one octave higher.
    This also means no matter where a pre-pro (or receiver) crosses over the center or surrounds, their re-directed bass is sent to the main L&R, and ultimately to the sub.
    It also means my sub is available for 2-channel music, and I really enjoy the foundation this gives to lots of music.
    3) I also use a BFD 1100 Parametric EQ (only to the sub) to cut the room-mode bass peaks. I generally ignore most of my nulls as far as the Parametric EQ on the sub goes.
     
  5. Ken Woodrow

    Ken Woodrow Stunt Coordinator

    Joined:
    Feb 12, 2001
    Messages:
    113
    Likes Received:
    0
    Now you've got me really intrigued. The Marchand electronic crossover looks like a great unit. I assume that you decided on the crossover point after analyzing and comparing your in-room response for your speakers and sub as you've described. I've looked at a variety of other solutions, such as the AudioControl Diva, the TACT RCS 2.0 and TCS, and the Clarity EQ PD-6.6, all of which provide multi-channel DSP with adjustable crossovers and EQ for all channels. I've held off on all of these b/c of their cost -- ranging from $3000 up to nearly $10,000 -- and because I believe a similar solution will be available for the HTPC within the next year. For the money, the Marchand may be the best thing going. Putting a 3-channel Marchand between the pre-outs of the receiver and my outboard amp would give me the maximum flexibility. I could run the following configurations:

    1. S/PDIF out from HTPC: Receiver decodes, L/R/C = Large; RS/LS = Small; Marchand filters bass from L/R/C to sub.

    2. 5.1 analog out from HTPC: Receiver passes through full-range signal to all 4 main speakers; Marchand filters bass to sub from L/R/C.

    3. 2 Channel analog out from HTPC: Marchand filters bass to sub.

    So, my question for you is (yes, there is a question in this post), did you perceive any negative effect on your sound by including the Marchand in your signal chain? Also, after looking at the various configurations of the XM9, am I right in understanding that the 3-channel version can be used as I have described?

    Thanks for all the input!

    Ken
     
  6. ThomasW

    ThomasW Cinematographer

    Joined:
    Nov 6, 1999
    Messages:
    2,282
    Likes Received:
    0
    I use a Marchand XM9 with my HTPC and a Sony TA-E-9000ES pre/pro.

    The S/PDIF out from the HTPC goes to the Sony. It's set to "No" for sub. The L&R main outputs feed the Marchand.

    BTW, I have 4 Marchand XM9's in two different tri-amped systems. They are dead quiet and sonically neutral.
     
  7. BruceD

    BruceD Screenwriter

    Joined:
    Apr 12, 1999
    Messages:
    1,220
    Likes Received:
    0
    Yes, after much research and reading of books and forums I came to my solution slowly, but critically.

    I also didn't have the kind of money necessary to implement digital correction systems.

    I've tried to critically listen to music in various configurations with my system, my favorites are female vocalists, jazz, and some classical. I've found the noise floor or S.N ratio of the Marchand crossover to be completely transparent to me, or no negative effects. I know it gets great reviews over at the AA forum (which is where I found out about it). I'm using it with Dynaudio Contour speakers and Parasound amplification.

    As far as it's use, let me explain. It is a stereo based unit i.e. 2 channels. But those two channels can both be cut up into 2 or 3 segments (just like a 2way or 3way speaker) based on which unit you buy. It is essentially a substitute for the passive crossovers found in most speakers. I only use it as a stereo crossover for the lowest 2 octaves (20Hz-80Hz) between my L&R mains and subwoofer.

    So, you couldn't actually use it for L/C/R + sub speakers, that would require something like a 4 channel unit. For that configuration, you might want to look at a new product called the "ICBM" from Outlaw. You should be able to do a search on it and find some discussions in this forum. I know it's been getting great reviews.
     
  8. BruceD

    BruceD Screenwriter

    Joined:
    Apr 12, 1999
    Messages:
    1,220
    Likes Received:
    0
    Ken,

    If you know how to setup a MIDI connection on your computer, you can control your BFD with software from your computer at the same time you are running ETF software in another window.

    A sort of real-time adjustment, test, adjustment, etc.
     
  9. Ken Woodrow

    Ken Woodrow Stunt Coordinator

    Joined:
    Feb 12, 2001
    Messages:
    113
    Likes Received:
    0
    I have no midi connection on the soundcard in my HTPC, but I could use my wife's laptop for the analysis. It also may be possible to temporarily install an inexpensive soundcard with midi connections in the HTPC along with the M-Audio card -- I believe that M-Audio and Windows XP support multiple sound cards. The software configuration could be tricky, though.
    One question concerning the data I've generated with ETF -- what is the best graph for understanding in-room response of a speaker and sub? I've found the 1/3 octave logarithmic to be the most intuitive, but what about gate times? Is there a meaningful correlation between the octave slice and gate times? I realize that 1/3 octave is popular because so many parametric EQs have a 1/3 octave resolution, but are there other reasons for using this slice? I also realize that increasing the gate time is necessary in order to see the low frequency response, b/c the low frequency waves need longer to reach the microphone.
    Thanks for the tip on the ICBM. I had forgotten about that little unit. Although it may not be completely transparent (see this month's SGHT), it will definitely improve the sound from the stereo and analog outputs of the HTPC, at least until software for bass management is released for multichannel sound cards. It will also open up the possibility of installing a SACD or DVD-Audio player into the system at a later date. It could be combined with a multi-channel analog preamp, like the Sony TA-P9000ES, for about the same price as the Marchand unit. That combination would almost allow me to ditch the receiver, but for the need to decode DD5.1 from my DSS recevier. Oh yeah, I guess I could find a used DD/DTS processer on eBay...
    You may wonder why I am obsessed with using the HTPC as a preamp, rather than just buying a good surround processor or receiver. One motivation is that a HTPC-based solution will be perpetually upgradeable as new surround sound formats are developed, b/c the decoding occurs in software. Already, I am able to decode CDs and WAVs in DD Prologic II w/o having to buy any more equipment. Another possibility is that speaker and room correction could be fully automated in software, eliminating the need for expensive digital EQ units like TACT or DiVa. Finally, I guess I am hooked by the DIY bug, the most incurable reason of all. [​IMG]
     
  10. BruceD

    BruceD Screenwriter

    Joined:
    Apr 12, 1999
    Messages:
    1,220
    Likes Received:
    0
    I built my currrent PC with HTPC in mind in Dec 1999, but never really got a decent soundcard, so now I just use it in my office. I wanted to get a FP video unit and use the HTPC to scale the picture size, but ran short on $$.

    Sounds like your having fun, and I'll follow your progress as I haven't bought the pre-pro I want yet.

    Typically the MIDI port is part of the joystick port on most desktop PCs and only requires an adapter cable. For my notebook PC I bought a Midiman Parallel port adapter and a couple of MIDI cables.

    The software is on the Behringer website.
     
  11. Greg Lee

    Greg Lee Agent

    Joined:
    Dec 12, 2001
    Messages:
    30
    Likes Received:
    0
    I hope I don't pollute this informative thread with too-dumb questions regarding ETF, but I get my BFD tomorrow and want to be as ready as possible.

    First dumb question - is is A) necessary, or just B) desirable to use ETF with its own test tone run through my HT system? Or would one of the test disks be suitable.

    Second question:

    Quote: (BruceD)

    It also means my sub is available for 2-channel music, and I really enjoy the foundation this gives to lots of music

    When I set my reciever (Onkyo 696) to Stereo mode (e.g. for listening to CD's) I get subwoofer engagement; this seems to be the same result you describe. Is this something special or unique to this reciever or am I misunderstanding something?

    With this in mind, I assume that I can take the 1/8" out from my PC sound card into a Y on to phono plugs into CD-In on the reciever as test tone. This will drive speakers and sub utilizing bass management so as to give a relevant result, or?

    Regarding MIDI control of the BFD: my sound card has a 'Digital Out'. I assume this is the correct output that can control the BFD?

    Last dumb question (for now) : What is 'HTPC' ?
     
  12. Ken Woodrow

    Ken Woodrow Stunt Coordinator

    Joined:
    Feb 12, 2001
    Messages:
    113
    Likes Received:
    0
    I'll take a crack at a few of these questions.

     
  13. Greg Lee

    Greg Lee Agent

    Joined:
    Dec 12, 2001
    Messages:
    30
    Likes Received:
    0
    Thanks for the quick answer Ken.

    I definitely want to use the RTA method - not the sine wave; measure; graph method - so I'll rig up the PC to HT connection.

    Re your comment about 'integrating the sub'

    My reciever definitely engages the sub in CD mode - I have my DVD player connected via optical link AND also analog cables to the CD-In. The sound for CDs is identical regardless of the mode (Stereo, DPII) or the input (DVD vs CD)

    Actually this brings up another interesting point - your comment about the rcvr translating from analog to digital for processing might explain the somewhat surprising result regarding the 'identical' sound; I expected some slight difference between the DAC of the DVD player (low-end Sony) versus the receiver. This would also imply that the world's best CD player would not benefit this system,since its DAC would be 'overridden' (not the best word) by the monkey business in the reciever, except in Direct mode.

    My sound card is a Creative Sound Blaster Live! Value (WDM)-this according to the Windows ControlPanel. Its the 'stock' card that came in my 833 MHz Dell.
     
  14. Ken Woodrow

    Ken Woodrow Stunt Coordinator

    Joined:
    Feb 12, 2001
    Messages:
    113
    Likes Received:
    0
    Greg:

    I believe that your Soundblaster has a digital audio out -- definitely different from a MIDI port. Usually, the MIDI port is combined with the game port.

    It sounds like your Onkyo is in fact redigitizing the analog inputs from your CD player. I suspect "direct" mode bypasses the internal DACS (and, hence, the sub) and sends a full range signal to the L/R speakers. Which makes it hard to compare DACs, b/c you'd be comparing the sound with and w/o the sub, hardly a fair comparison. One way to make the comparison fair, though, would be to disable the sub in the receiver (tell it "no sub" in the setup). That way, you could make the following comparison:

    A: CD player -- digital out -- receiver (=no sub)

    B: CD player -- analog out -- receiver (direct mode)

    Both ways should send a full signal to the L/R speakers, with option A using the Onkyo's DAC and option B employing the CD player's DAC.

    But we digress. You should be able to generate a test tone in ETF, route it to your CD inputs, and produce sound from both your sub and main speaker. BUT -- as BruceD explained, this may not be the best way to decide on your crossover points or sub EQ. I believe that he recommends testing each independently and comparing the graphs to determine the optimal crossover point and then testing the sub with various EQ settings on the BFD.

    I plan to experiment with it this weekend -- we'll keep each other posted. Eventually, if I get ambitious, I'll do a FAQ for my website.

    - Ken
     
  15. BruceD

    BruceD Screenwriter

    Joined:
    Apr 12, 1999
    Messages:
    1,220
    Likes Received:
    0
    Greg,

    Just as a reference point, ETF can be used with the internal MLS test signal it produces, or you can download the CD image files they also provide in the download area, rip then to a CD and use the CD MLS test signals as a source for the ETF program instead.

    The CD is mostly for people who don't have a sound card that supports full duplex capability (record and play at the same time).
     
  16. Greg Lee

    Greg Lee Agent

    Joined:
    Dec 12, 2001
    Messages:
    30
    Likes Received:
    0
    Ken and Bruce,
    Thanks for the info so far.
    Regarding the use of ETF to measure response from the mains and sub separately in order to set crossover - is there still merit to this since I can't adjust the crossover with my reciever? The sub amp does have a variable crossover, but I had just assumed I should set this to its highest level (160 HZ); I suppose with these wonderful tools at our disposal there is no reason to assume anything - just test it![​IMG]
    Further, assuming that setting x-over is not a factor, I am not understanding why using the EQ on the sub w/o the mains running would be useful. It would seem that you need to EQ the sub in consideration of the energy also being put into the room by the mains.
     
  17. Ken Woodrow

    Ken Woodrow Stunt Coordinator

    Joined:
    Feb 12, 2001
    Messages:
    113
    Likes Received:
    0
    Greg, et al.

    Just an update and perhaps an answer to some of Greg's questions. First, with respect to the combined speaker/sub measurements, my point about running the tests separately was that doing so would (theoretically) help you to decide on a crossover point. If you pre/pro gives you no option, then it may be a moot point. But it might still be helpful to see the energy produced by the sub alone, b/c it is the only thing you can adjust when using the BFD as an equalizer for the sub.

    I investigated my receiver and discovered that it WILL activate the sub when receiving input from the analog CD inputs. I suspect that it does so by converting the signal to digital, applying DSP, and reconverting to analog. It also as a 2-channel direct mode that bypasses the DACs.

    I ran two measurements last night using a low frequency signal from ETF into my CD inputs: (1) Front speaker alone without sub (no extra digital-to-analog conversion and no crossover); and (2) Normal two-channel mode (with receiver's sub crossover engaged) but with the front speaker disconnected. My goal was to determine how much low frequency energy my mains could produce and see where it would overlap with the sub. My results were interesting -- my main speakers produce well to about 60hz and then they drop off quickly. The sub (subject to a 80 hz high-pass crossover in the receiver) produces energy well up to 80 hz. I have a lot more work to do, but this preliminary measurement tells me that I should try crossover points below the conventional 80 hz found in most receivers. Therefore, I'll probably investigate the use of an external crossover like the Marchand or Outlaw ICBM. (I could use the sub's crossover, but that would require additional cabling that is not feasible in my room and would work only on the front two channels).

    - Ken
     
  18. BruceD

    BruceD Screenwriter

    Joined:
    Apr 12, 1999
    Messages:
    1,220
    Likes Received:
    0
    Greg, yes it is preferred to use the receiver's digital crossover and either turn the sub's internal crossover off or to it's highest setting, like your 160Hz.

    Ken, I'm assuming your receiver's crossover is at 80Hz.

    ------------------------------

    Here is some of my experience with crossover options.

    Many receiver's use an 80Hz crossover with different slopes for the main speakers (high-pass) and the sub (low pass). Slope=how fast the crossover cuts the sound output below the crossover frequency (like how steep is that ski slope).

    A 24dB/octave slope is steeper than a 12dB/octave slope and cuts more of the sound output as well as cutting it faster.

    So this means the slope of the crossover for the low-pass or sub is actually used to cut the frequencies above the crossover frequency.

    The slope of the crossover for the high-pass or mains is actually used to cut the frequencies below the crossover frequency.

    Typically a receiver's crossover has a slope of 24dB/octave for the low pass and a slope of 12dB/octave for the high pass. The high-pass was designed for a "sealed" main speaker that has a natural bass rolloff of 12dB/octave @80Hz. So, the speakers 12dB + the receiver's 12dB was suppose to equal 24dB, except most of us don't have sealed speakers with a -3db of 40Hz (the optimal design for this crossover).

    I wasn't particularly satisfied with that kind of high-pass crossover (I could never get it to smoothly integrate mains and sub), which is the reason I bought the Marchand electronic crossover so I could roll my own crossover (since my pre-pro is not adjustable).

    The Marchand has symmentrical 24dB slope Linkwitz-Riley crossovers (high-pass and low-pass) and zero (0) degrees of phase offset. It also provides level controls for high-pass, low-pass, and the exact crossover point plus some more features.

    Anyway, my mains are spec'd at -3dB to 32Hz, so I tried a few different crossover values in my room before finding 60 Hz worked best. 60Hz is approximately one (1) octave above my -3dB low point of 32Hz.

    I've read that many recommend a full octave of bandwidth be used to ramp down the high-pass crossover for the mains with a 24dB slope.

    You can actually get filters with 36dB and 48dB slopes, but I think their phase variences are not zero.
     
  19. Greg Lee

    Greg Lee Agent

    Joined:
    Dec 12, 2001
    Messages:
    30
    Likes Received:
    0
    Could someone give an explanation of what is meant exactly by 'losing headroom' when you boost a frequency using the BFD. I assume that a boost is achieved essentially by applying a corresponding cut to all frequencies outside the filter. So to compensate, you would have to boost the 'overall' signal going to the sub, either at the reciever speaker setup or at the amp. Is this bad?

    I guess the degree of 'badness' depends on your preferred listening volume in proportion to the power of your system. For my listening in a large room (6500 cuft) with an SVS 25-31 driven by PE 250 W amp, I have no problem balancing the sub level to the mains at 80 db; but I guess if I'm listening at reference level, and the soundtrack calls for 115 db, the system will not have enough umph?
     
  20. BruceD

    BruceD Screenwriter

    Joined:
    Apr 12, 1999
    Messages:
    1,220
    Likes Received:
    0
    I'll give it a stab.

    And no, a boost doesn't institute a corresponding cut.

    Typically amp headroom is the instantaneous capability of the amp to drive speakers producing large dynamic swings in SPL (eg. soft to very loud) without compressing or clipping the music. Clipping produces distortion.

    Bass frequencies (especially loud ones) put higher loads on all amps, using up more of the amps capacity quickly.

    When you boost a bass frequency with the BFD you are putting even more strain on the amp, essentially pushing it to compress or clip even sooner by using up more of the amps capacity.

    This reduces the absolute compression or clipping level from that of an unboosted bass frequency. Normally you want this as high as it can get.

    Lowering of this compression or clipping level is called a loss of headroom. There is simply less and less amp power left to use.

    Some amps even provide a specification for headroom. The higher the better.
     

Share This Page