test tones

[normc]

Anybody had this problem which I'm finding a little confusing. After buying a new subwoofer (Velodyne SPL1000U), I've used sine wave test signals to experiment with speaker placement, bass traps and curtains to optimize everything. Having achieved the best compromise I can find, I then played the pink noise test tones from my Rotel SP to adjust the balance between the 5.1 channels. The result was that the center channel was 3db too high and the subwoofer seems to about 20db too high, with the pink noise.

My question is should I trust the test tones of the SP and adjust off them before checking the overall frequency response of the system, or the other way around? It seems to be a chicken and egg situation, or is it a matter of repeating it several times before arriving at the final adjustment? Thanks for the input.

 

[captiankirk28]

You will want to go with the test tones.

 

[Alex2507]

You will want to go with the test tones.

... of the Rotel SP (I think). How does it sound? A tad bass heavy?

 

[normc]

Can't really say the RSP-1069 (it's paired with the RMB-1075 and the Denon 1940 as the source) has any sonic signature that I've noticed. Bass is far more a function of the room, speakers, placement and adjustment I suspect.

 

[Alex2507]

It took me a minute to fiure out why you said the above but I was being cute in reference to you saying this:

the subwoofer seems to about 20db too high

I wasn't suggesting that your gear was coloring the sound. It just sounds like you need to turn your sub down ... way down.

Welcome to AH.

 

[AVRat]

Make the adjustments, get the speakers balanced first, then check overall response.

 

[Savant]

Level setting should be accomplished with broadband noise signals and listening tests. Tones should be used only to identify any anomalies in the room/system response.

 

[STRONGBADF1]

Level match first, then check frequencies, make adjustments, then level match again.

 

[normc]

Perhaps my question should have been a little clearer.
I've done some extensive adjusting using sine wave test tones and arrived at what I considered to be a pretty balanced sound. I'm an ex musician so I have a pretty decent ear. But as an additional check, I played the pink noise tones from the RSP and the subwoofer level showed as being +20db compared to the other channels. This was very surprising and I wondered if the pink noise generator might be out of calibration. Then I realized it's most likely the same pink noise signal being rerouted to the different channels.

So I'm confused as to why such a large difference and which method I should believe. I know I'll believe my ears in the end but that will take a lot of listening to different program material. I'm just trying to speed up the process a little.

Given the difficulty and expense of eliminating all the peaks and valleys, my current thinking to get the most balanced sound is to adjust the level of the sub so that its average output level matches that from the other speakers.

Any thoughts from the experts on that practice?

Norm

 

[MDS]

Perhaps my question should have been a little clearer.
I've done some extensive adjusting using sine wave test tones and arrived at what I considered to be a pretty balanced sound. I'm an ex musician so I have a pretty decent ear. But as an additional check, I played the pink noise tones from the RSP and the subwoofer level showed as being +20db compared to the other channels.

'Balanced sound' with regard to frequency response is much different than using the test tones to balance the levels of each channel and each requires a different tool.

Pink noise is noise that has equal amplitude at every octave and it is used solely to get the levels of each channel roughly equal. You cannot mess with the individual channel levels to correct a perceived error in any given frequency - you need an EQ.

 

[normc]

Happy to learn from those more qualified and the above statement is certainly true with reference to the other channels, but I don't see the problem with applying it to the sub woofer. If the frequencies going to the sub below the X over frequency are 'out of balance' with those above X over going to the other speakers, why can't this be corrected by altering the subs output level. What else is it for if not that. It may not correct for the individual peaks and dips that exist in a particular room, but it will certainly correct for the overall energy balance between the highs and lows. Meaning the sound would be neither overly bright nor too shaded.

I think my point about the practice of balancing this energy by averaging out the frequencies is valid because it balances the energy between the 2 spectrums. For instance if the frequencies above the Xover point were all flat at a reference of say 70db, then their average would be 70db. If the sub rolled off linearly below a X over of say 100Hz to 50db at 20Hz, then its average would be 60db. It would then require its average output to be brought up to 70 db to obtain a 'balanced' sound. Of course, just raising its output by 10db is not going to give the same results as another sub that was flat down to 20Hz because such an adjustment would raise the output at 20Hz to 60db and at 100Hz to 80db, so the mid bass would be unnaturally emphasized. But its going to sound more balanced than leaving it at -10db relative to the other speakers. I've applied this to my set up now and the overall sound is nicely balanced with transparent highs, good rendition of ambience, and the bass instruments have the proper extension and weight. I still have the peaks and dips but as I said, eliminating them is both difficult and expensive. I have reduced them with bass traps in both front corners but the room is close to square so it's not an ideal shape.

In theory the exact amount of energy in either spectrum would need to be calculated using calculus but the average gives a pretty close approximation as long as the peaks and dips are not too extreme.

Hope this all helps someone else as its been a learning experience for me.

 

[STRONGBADF1]

Happy to learn from those more qualified and the above statement is certainly true with reference to the other channels, but I don't see the problem with applying it to the sub woofer. If the frequencies going to the sub below the X over frequency are 'out of balance' with those above X over going to the other speakers, why can't this be corrected by altering the subs output level. What else is it for if not that. It may not correct for the individual peaks and dips that exist in a particular room, but it will certainly correct for the overall energy balance between the highs and lows. Meaning the sound would be neither overly bright nor too shaded.

I think my point about the practice of balancing this energy by averaging out the frequencies is valid because it balances the energy between the 2 spectrums. For instance if the frequencies above the Xover point were all flat at a reference of say 70db, then their average would be 70db. If the sub rolled off linearly below a X over of say 100Hz to 50db at 20Hz, then its average would be 60db. It would then require its average output to be brought up to 70 db to obtain a 'balanced' sound. Of course, just raising its output by 10db is not going to give the same results as another sub that was flat down to 20Hz because such an adjustment would raise the output at 20Hz to 60db and at 100Hz to 80db, so the mid bass would be unnaturally emphasized. But its going to sound more balanced than leaving it at -10db relative to the other speakers. I've applied this to my set up now and the overall sound is nicely balanced with transparent highs, good rendition of ambience, and the bass instruments have the proper extension and weight. I still have the peaks and dips but as I said, eliminating them is both difficult and expensive. I have reduced them with bass traps in both front corners but the room is close to square so it's not an ideal shape.

In theory the exact amount of energy in either spectrum would need to be calculated using calculus but the average gives a pretty close approximation as long as the peaks and dips are not too extreme.

Hope this all helps someone else as its been a learning experience for me.

This is one of those times that you wont be satisfied until you try it. I mean that in a nice way! Besides, it's fun to experiment and learn these things at times. Try it and then see if you like it. I would not but you may.

I would find the results more unbalanced because of the exaggerated peaks in a frequency(s) would stand out like a sore thumb on recordings that have an instrument that uses that frequency. See, you don't necessarily miss something you can't hear very much but if a sound is unnaturally loud it can be distracting.

-SBF1

 

[highfigh]

Anybody had this problem which I'm finding a little confusing. After buying a new subwoofer (Velodyne SPL1000U), I've used sine wave test signals to experiment with speaker placement, bass traps and curtains to optimize everything. Having achieved the best compromise I can find, I then played the pink noise test tones from my Rotel SP to adjust the balance between the 5.1 channels. The result was that the center channel was 3db too high and the subwoofer seems to about 20db too high, with the pink noise.

My question is should I trust the test tones of the SP and adjust off them before checking the overall frequency response of the system, or the other way around? It seems to be a chicken and egg situation, or is it a matter of repeating it several times before arriving at the final adjustment? Thanks for the input.

I would use pink noise for adjusting the sub, too. Frequency sweeps are useful, too but if you sweep too slowly, the energy from one frequency can still be affecting the others, causing comb filtering and ultimately, your results.

 

[Savant]

I think my point about the practice of balancing this energy by averaging out the frequencies is valid because it balances the energy between the 2 spectrums. For instance if the frequencies above the Xover point were all flat at a reference of say 70db, then their average would be 70db. If the sub rolled off linearly below a X over of say 100Hz to 50db at 20Hz, then its average would be 60db. It would then require its average output to be brought up to 70 db to obtain a 'balanced' sound. Of course, just raising its output by 10db is not going to give the same results as another sub that was flat down to 20Hz because such an adjustment would raise the output at 20Hz to 60db and at 100Hz to 80db, so the mid bass would be unnaturally emphasized. But its going to sound more balanced than leaving it at -10db relative to the other speakers.

Overall, there are various technical flaws in this process. However:

I've applied this to my set up now and the overall sound is nicely balanced with transparent highs, good rendition of ambience, and the bass instruments have the proper extension and weight.

This is the important part. We could go back and forth, hashing out explanations for why the averaging described doesn't make technical sense (though I understand why you're approaching it that way), or the use of tones versus broadband sound to set levels, or... In other words, be nitpicky.

I don't think that's necessary and, ultimately, it matters very little since you are obviously satisfied with the end result.

 

[STRONGBADF1]

This is one of those times that you wont be satisfied until you try it. I mean that in a nice way! Besides, it's fun to experiment and learn these things at times. Try it and then see if you like it.

-SBF1

Overall, there are various technical flaws in this process. However:


This is the important part. We could go back and forth, hashing out explanations for why the averaging described doesn't make technical sense (though I understand why you're approaching it that way), or the use of tones versus broadband sound to set levels, or... In other words, be nitpicky.

I don't think that's necessary and, ultimately, it matters very little since you are obviously satisfied with the end result.

Agreed, Always best to make yourself happy and not some graph!

 

[Alex2507]

Always best to make yourself happy and not some graph!

Consider this your notice of having your AH status put under official review.

Without graphs we would be like monkeys trying to f^%k a football.

 

[normc]

Thanks for your input and help guys. Obviously what we're trying to achieve in theory is a flat response but given the various difficulties and expense of doing that, I'm hoping I've found the best compromise.

In the interest of learning more though, I'd still like to understand the 'flaws' in my approach and what better approach I might take.

I've spent $3,000 on a new sub, bass traps and curtains, spent quite a bit of time on research for listening position, speaker and bass trap placement. I think I've achieved a pretty good result but would love to know if I've missed anything.

Bear in mind the following;

1. I don't have an unlimited budget and the HT (which is not symmetrical btw) has to double as a living room.

2. This means the listening and speaker positions, while done very carefully, may not be perfect.

3. The Radio Shack meters are not reference measuring devices so the readings are not 100% reliable.

So given the above, I'd be keen to know what better approach I might use, or what more I might be able to do that could yield better results.

 

[STRONGBADF1]

Without graphs we would be like monkeys trying to f^%k a football.

You sould know!

 

[STRONGBADF1]

Thanks for your input and help guys. Obviously what we're trying to achieve in theory is a flat response but given the various difficulties and expense of doing that, I'm hoping I've found the best compromise.

In the interest of learning more though, I'd still like to understand the 'flaws' in my approach and what better approach I might take.

I've spent $3,000 on a new sub, bass traps and curtains, spent quite a bit of time on research for listening position, speaker and bass trap placement. I think I've achieved a pretty good result but would love to know if I've missed anything.

Bear in mind the following;

1. I don't have an unlimited budget and the HT (which is not symmetrical btw) has to double as a living room.

2. This means the listening and speaker positions, while done very carefully, may not be perfect.

3. The Radio Shack meters are not reference measuring devices so the readings are not 100% reliable.

So given the above, I'd be keen to know what better approach I might use, or what more I might be able to do that could yield better results.

Down load and then burn a test tone CD for free. I believe Real Traps site has one. If not just search this and other audio forums. One should be relatively easy to find. On these same sites there will also be conversion tables for which ever Radio Shack meter you use. Graph your results from the listening positing. Try this with your sub at different locations. (even a little change in position could make significant differences.)

If the results are bad enough you'll be buying an EQ as your next upgrade.

Even if you don't you can have some fun just playing.

 

[Savant]

normc,

Perhaps a flawed approach isn't quite it. The math for the averaging you described simply doesn't work; 70 dB and 50 dB do not average 60dB - neither aurally, nor mathematically. But, again, it's nitpicky. You achieved a sound quality that works. Why nitpick it?

As for the tones versus broadband noise, I can tell you understand what's going on with the comment you made about calculus, which is indeed what you'd need to determine the "equivalence" of the noise from the two types of source signal. But to delve beyond that is getting, well, nitpicky...again.

If I might suggest a very simple LF vs. HF level-matching approach that can be done with the Radio Shack meter: With broadband pink noise running through the whole system, just make sure the SPL(A) and SPL(C) readings are within a few (1-2) dB of each other with SPL(C) > SPL(A). If the SPL(C) reading is much more than 2 dB above the SPL(A) reading, there might be too much low end. If they're equal or if the SPL(C) reading is lower than the SPL(A) reading, then there's probably an LF deficiency. This can be verified / cross-checked with listening tests.

Of course, tones should still be used to diagnose and resolve - as best as possible with the available tools & funds - LF anomalies. But to simply level-match low and high, I find this to be the easiest approach if you don't have an RTA or FFT at your disposal.

HTH.