Test-source recommendations?

[taudo]

Thought so! Thanks again!

taudo

[taudo]

I have received the "Alan Parsons Sound Check 2" CD and "audio analyser" I mentioned earlier in this thread. Early testing indicates to me that this very inexpensive product does exactly what it claims to do, namely to provide "... a quick and simple method of ascertaining frequency response where [the] clinical accuracy of expensive laboratory equipment is not needed."

The product is a CD full of test tones along with a simple calibrated SPL meter built into the CD jewel case. To make a frequency response plot you simply adjust your speaker volume to 0 dB with the 1000 Hz 1/3 octave pink noise band, then let the next 31 tracks play while recording the SPL as given by the LEDs on the back of the jewel case. The LED's are calibrated in 3 dB increments from -15 dB to +12 dB with the 0 dB point reported to be set at 80 dB SPL "C" weighted (which I found to be quite a comfortable level - neighbors won't complain).

I ran the test several times and found that the results were quite reliable in that I got the same reading with each test. Not having test equipment, I can't say if these levels are properly calibrated to some external standard. For that I must take Mr. Parsons' word.

There are lots of other tracks on this CD - sine waves, sweep tones, uncompressed speech and music, sound effects, and various "utlilty tracks" to identify channels, identify noise, and tweak your timecode equipment.

For those of us with no access to _real_ audio test gear, this product is highly recommended! You can find it at the links I included in this thread above.

It would be really great if someone on this board with a benchful of test equipment could give us his analysis.

taudo

[David Deratany]

Somewhere along the way, I picked up a batch of wav files, each a 10 second pure sine wave, starting at 20 hz and going in 5 hz increments up to 100 hz. At the very least, I learned my grill cloths really didn't care for them. The good news is that I was getting 25 hz out of my 416s.

I'll be glad to attach the files in a reply to anyone who sends me a PM.

I wish I remembered who to thank for them. Maybe he is from this board.

[taudo]

Earlier in this thread BobR wrote:

>Generally speaking, gated measurements are usually good down to ~200Hz. These are combined with extremely close-miked measurements (to swamp out reflections) below 200Hz to obtain a combined response plot.

RTAs are good for measuring environments (speaker + room) but cannot deliver accurate speaker-only measurements. Same goes for any test CD + SPL meter combination.
====

Continued playing around with the Sound Check 2 CD seems to bear this out. Above 200 Hz, the data seems to "make sense", below that it indicates a roll-off far too precipitous to believe - and one that does not accord with my ears or changes in porting.

Given that we're mostly interested in responses below 200 Hz, I guess I'll have to modify my recommendation a bit to reflect this. Nonetheless, I think this thing is well worth it's price.

Now to investigate Praxis and CLIO.

taudo

[taudo]

Placing the Sound Check 2 meter _inside_ the enclosure yielded more interpretable results. It showed that my 828 with a 100 sq in port resonated exactly where theory (and Altec graphs) predicted it would - about 50 Hz. The "height" of the resonance hump was about 3 dB. There was a secondary hump at about 125 Hz, also about 3 dB.

It is interesting to me that this "high" tuning _sounds_ so good. Many would want the enclosure tuned much lower. However, this discrepancy may be partially explained by a line in the instructions for this CD that says "...with music systems, a slight rise (3-6 dB) between 63 and 125 Hz and another at around 10 kHz will help to improve the apparent "fidelity" of the system by stressing those frequencies where much of the energy occurs in music...".

Given that the raw frequency response plot of a system seems to bear so little relationship to the perceived "fidelity" of the system, might it not make alot of sense to worry less about theoretical perfection of the speaker system's response and, instead, rely upon an RTA with a 1/3 octave active equalizer to tune both the system and the listening room? (Of course, I'm assuming that you start with a "good" even if not ideal system to begin with. I'll choose Altecs!)

I'll be interested to see what the pros have to say.

taudo
(this is just too much fun to be legal! :-) )

[joyspring]

Taudo,

"Above 200 Hz, the data seems to "make sense", below that it indicates a roll-off far too precipitous to believe - and one that does not accord with my ears or changes in porting."

The cutoff of the 828 horn is ~200hz and the precipitous drop in amplitude below that is definitely believable and indeed indicated in published Altec frequency response plots.

A front-loaded horn is much more efficient than the back-loaded box/vent so a `shelved' response is to be expected, particularly when the box/vent is driven by an overdamped (low-Q) driver and (too low) port tuning.

"It is interesting to me that this "high" tuning _sounds_ so good. Many would want the enclosure tuned much lower."

Considering the exceptionally low total driver Q of the 416/515, tuning lower would cause an even steeper response `droop' in the passband resulting in a subjectively weaker bass as well.

`Vintage' Altec (and JBL and Tannoy) systems utilising low-Q high-efficiency drivers tended to use `too large' enclosures with `too low' vent tuning which yielded the characteristic drooping response with bump at port tuning.

With most Altec low-Q 416/515-type drivers with Qts varying from 0.18-0.27 and fs ranging from 20-30hz, a maximally flat `B4' alignment would yield f3 (-3dB) in the 65-100hz range! A 50hz tuning on an 828 enclosure is not too far off the mark.

"Given that the raw frequency response plot of a system seems to bear so little relationship to the perceived "fidelity" of the system, might it not make alot of sense to worry less about theoretical perfection of the speaker system's response and, instead, rely upon an RTA with a 1/3 octave active equalizer to tune both the system and the listening room? (Of course, I'm assuming that you start with a "good" even if not ideal system to begin with. I'll choose Altecs!)"

You're correct in stating that an ideal `flat' frequency response plot *may* have little bearing on the perceived fidelity. HOWEVER - in the MAJORITY of cases - the flat response will definitely indicate that the system in question will sound good.

Why? If the response is `flat', it generally indicates that the system designer bothered to consider all the things that will yield a flat response including phase, crossover frequency/topology, box/driver interaction/tuning, etc.

There's more to this though... the total power response; that is, that a system will deliver the same total energy at every frequency will also have a larger effect on how a system sounds in a room than simply evaluating on-axis frequency response. The focus on even power response was the impetus behind constant-directivity horns.

Equalisation with 1/3-octave filters can correct for room anomalies (and is often a fact-of-life) but should never be considered to adjust for systems with poor on-axis or total power response; system replacement should be considered at that point ;-)

And contrary to popular belief and marketing hype, audio equipment (including loudspeakers) are evaluated using objective measurements *far* before subjective listening tests. That is why Altec/JBL/other loudspeakers that sound good (with a few exceptions i.e. - the 604) generally measure close to flat.

BobR

[taudo]

BobR,

Thanks a million for that analysis! It seems to me that there's just a ton of good sense in there.

A couple of follow up questions/comments, if I may:

>The cutoff of the 828 horn is ~200hz and the precipitous drop in amplitude below that is definitely believable and indeed indicated in published Altec frequency response plots.
====
Wow. I've actually never seen such a plot

>Considering the exceptionally low total driver Q of the 416/515, tuning lower would cause an even steeper response `droop' in the passband resulting in a subjectively weaker bass as well.

====
Absolutely what my experience has been! Nice to see that validated. I thought I was going nuts!


>`Vintage' Altec (and JBL and Tannoy) systems utilising low-Q high-efficiency drivers tended to use `too large' enclosures with `too low' vent tuning which yielded the characteristic drooping response with bump at port tuning.
====

I guess I don't understand the reason(s) for this - unless it was considered beneficial in the days of limited BW source material and very modest tube amps. Is there another reason?

>A 50hz tuning on an 828 enclosure is not too far off the mark.
====
Again - thanks for the validation!


>in the MAJORITY of cases - the flat response will definitely indicate that the system in question will sound good.
====

Given that 416/828's response is inherently not very flat, what would be your recommendations for "flattening" a playback _system_ using A-7's? For example, as counterintuitive as it initially seems to me (in a system with 4 15" woofers and 40 cu ft of reflex space) - it may be that my system is going to need one or more subwoofers to flatten out the response below 100 Hz. Would you agree?

>There's more to this though... the total power response...will also have a larger effect on how a system sounds in a room than simply evaluating on-axis frequency response.
====
Are there, in fact, systems that do a pretty good job of delivering flat power response? I was under the (perhaps mistaken) that most systems try to deliver most of the energy in the lower frequencies to offset the marked nonlinearity of our ears.

>system replacement should be considered at that point
====

Alas, Alon Grand Reference systems are a bit out of my hobby budget!

>That is why Altec/JBL/other loudspeakers that sound good (with a few exceptions i.e. - the 604) generally measure close to flat.
====

Again, I am probably being even more dense that normal - but A-7's _sound_ very good to me despite apparently having pretty nonlinear response. Can you help me a bit more to understand why this seems to be so?


Once again, thanks a million for that very informative post!

taudo