update on my "new to me" VSX26 - first time out

[Coaster]

i spent many hours in my shop setting up my 'new to me' VSX26 this week. i was supposed to set it up outside yesterday and have at it at my leisure, but it rained cats and dogs all day yesterday.

 

this was for a wedding over two days, rehearsal friday then the whole shebang on saturday; ceremony and dance. fridays rehearsal ended up in a garage with no PA. saturday was like herding cats, i expected as much and was not disappointed. i set up about 4 hours before the ceremony and had a brief window between tent setup and pictures; i didnt want to bother anyone with system testing during pictures, and the people setting up tables and place settings took their sweet time so all in all i got about 30 minutes to really crank up the rig and listen.

 

boy am i glad i had that 30 minutes; i was way off. i was under a tent, i had no test equipment, i dont know what was the tent and what was my rig, so i used my ears.

 

first off i had the HF too hot, after all was said and done i dropped them another 3db on the HF amp attenuators. i went a little at a time and walked around. it sounded great in my shop, and looked pretty good on the RTA but it wasnt right outside under this tent.

 

second; the 250hz -3db i had made got axed. must have been a mode in my shop, it really sounded a lot better with that cut gone. i kept the -2db at 120, -1db at 400, -3db at 800 and all the work on the HF section.

 

i did have my white eq's hooked up, and found after listening for a while that i made the following additional cuts; -1 at 120 (already cut on the vsx), -2 at 160, -1 at 400hz (already cut on the VSX), -0.5 at 680, -1 at 800 (already cut on the VSX), and i noticed a sharp sound sticking out at 1k so i cut -1.5 at 1k.

 

if you never used white eq's i'll just say that a -0.5 cut is not subtle. the scale between 0 and -1 is huge.

 

then i started to notice that while the tops sounded better than i have ever heard them sound, the subs sounded weird. i have them HP at 41hz at LR24 and i find i am always putting bumps at 55hz and cutting 100hz so i played with that in the VSX. i added a 3db bump at 55, and cut -3db at 100. it may not have been the right thing to do but it sounded good.

 

this sounded rockin for playback. standing right in front of the tops didnt hurt my ears, and no one complained (lot of blue hairs stayed for i like big butts and some kanye sumthin' or other. i didnt pick the tunes). very intelligible.

 

the ceremony was outrageous sounding. it sounded just like talking, only louder. almost like the PA wasnt even there but if i shut it off you couldnt have heard a word. the mic placement was really bad as well, i had the mains firing 90 off axis to the mics in the same plane (speakers pointed at the mics, mics turned 90 degrees. its just the way it had to be for the situation). no feedback, no ringing, nothing. just loud, clear, good sound.

 

you'd swear i didnt have pile of crap speakers up there.

 

so basically, it sounds awesome and i have a lot more work to do. i need to actually get this thing in a situation where i dont have a tent, have smaart up and running and can really see what the hell is going on. obviously a lot of the work i did in my shop wasnt quite right, but the HF sounded good right off the bat; just a little too hot overall.

 

this is really neat, i really like this.

[davidss1]

happy to hear u r getting results that are pleasing to u .

[davidss1]

i still urge u to setup for and test using transfer functions to see whats happening with the phase relationships.

if u view coherence,it will help u to know what are reflections and what are not.

having a "baseline" tuning will help u have a consistant starting point to take into differing environments

[hornplayer]

 

this is really neat, i really like this.

 

 

Yep! Coaster is hooked. DSP claims another victim. Told Ya...

 

There will be another analog cross over for sale shortly.

[Coaster]

anyone want to buy a 223xl?

 

seriously.

[agedhorse]

Coherence can also roughly be determined by how much the eq per band affects the measurement per band... boost a band by 3 dB and if the measurement doesn't increase by 3dB there is a measurement anomoly. It can either be time offset between sources or a reflection or being in a peak/valley node of a reflection or summed acoustic problem, etc.

 

This is something that can be acquired with experience as well as understanding what the measurements mean, and what kinds of errors can creep in.

[davidss1]

 

Coherence can also roughly be determined by how much the eq per band affects the measurement per band... boost a band by 3 dB and if the measurement doesn't increase by 3dB there is a measurement anomoly. It can either be time offset between sources or a reflection or being in a peak/valley node of a reflection or summed acoustic problem, etc. This is something that can be acquired with experience as well as understanding what the measurements mean, and what kinds of errors can creep in.

 

 

words of wisdom and experience as always from andy,,,but for those less experienced,,a coherence trace is very helpful.

[agedhorse]

 

words of wisdom and experience as always from andy,,,but for those less experienced,,a coherence trace is very helpful.

 

 

Generally, the ability to get an accurate coherance trace is easier said than done. Just because you get a trace doesn't mean thart it's right or accurate. This is one of the dangers of relying on data that you can not (or don't) cross-check to be sure it's valid data.

[davidss1]

 

Generally, the ability to get an accurate coherance trace is easier said than done. Just because you get a trace doesn't mean thart it's right or accurate. This is one of the dangers of relying on data that you can not (or don't) cross-check to be sure it's valid data.

 

 

very true,,one cannot blindly trust 1 methodoligy.far better to cross test if time allows.

it helps to try several test mic positions to asses validity of coherence traces.

[agedhorse]

 

very true,,one cannot blindly trust 1 methodoligy.far better to cross test if time allows. it helps to try several test mic positions to asses validity of coherence traces.

 

 

Absolutely. I have seen too many false assumptions of valid data (alomg with not understanding the principles behind the data which would immediately call suspect to the data itself) lead to poorly calibrated and poor sounding end results... all the while the "dude" is bragging about how perfect his work is with Smaart and his vastly superior skills. Too bad his entire data set falls apart when the mic is moved 3 feet (~1M). Ultimate "fail".

[Coaster]

what are some clues to help me get a good transfer measurement?

[agedhorse]

Moving the mic 3 feet or 10 feet gives you a significantly different data set.

 

Changing the eq doesn't give you the measured effect that is predicted.

[jasps]

 

the subs sounded weird. i have them HP at 41hz at LR24 and i find i am always putting bumps at 55hz and cutting 100hz so i played with that in the VSX. i added a 3db bump at 55, and cut -3db at 100. it may not have been the right thing to do but it sounded good.... so basically, it sounds awesome and i have a lot more work to do. i need to actually get this thing in a situation where i dont have a tent, have smaart up and running and can really see what the hell is going on. obviously a lot of the work i did in my shop wasnt quite right, but the HF sounded good right off the bat; just a little too hot overall. this is really neat, i really like this.

 

Coaster,

Try using a BW slope for the HPF on sub duty. Until you can set your delay time correctly from subs to tops, I would experiment with your polarity between bands. If you get a boost in LF from a polarity reversal, you are on the right track.

[jasps]

 

what are some clues to help me get a good transfer measurement?

 

 

Go outside away from buildings (easy in ND). Put the mic on the ground so you don't get confused by reflections.

[Coaster]

should i keep the tops on poles then?

 

if on grass, is this a situation where i need a 2x2 board for the mic?

[agedhorse]

Coaster, Try using a BW slope for the HPF on sub duty. Until you can set your delay time correctly from subs to tops, I would experiment with your polarity between bands. If you get a boost in LF from a polarity reversal, you are on the right track.

 

With a BW alignment, you can silde the HPF frequency down to roughly 35-38Hz and get the same protection effect but probably with better audio results.

 

Reversing the polarity isn't going to do what you think, you are talking about 2 different frequency bands so there's little to no reinforcement of overlapping LF bands. Delay smooths out (presumably for the better ) the response where the bands do overlap, and the practical wayto do this is by delaying the overalpping bands so that they do match up iun the time domain. Unfortunately, analysis and delay are required plus the respponse will change as you move the measurement mic so be prepared to average a few data sets.

[agedhorse]

 

should i keep the tops on poles then? if on grass, is this a situation where i need a 2x2 board for the mic?

 

 

Yet, the tops must be as you inted to use them. The mic near a grass boundery is "ideal" because reflections will be minimized.

[jasps]

With a BW alignment, you can silde the HPF frequency down to roughly 35-38Hz and get the same protection effect but probably with better audio results. Reversing the polarity isn't going to do what you think, you are talking about 2 different frequency bands so there's little to no reinforcement of overlapping LF bands. Delay smooths out (presumably for the better ) the response where the bands do overlap, and the practical wayto do this is by delaying the overalpping bands so that they do match up iun the time domain. Unfortunately, analysis and delay are required plus the respponse will change as you move the measurement mic so be prepared to average a few data sets.

 

I don't get what you're saying: If the energy through the actual acoustic crossover has over 120 degrees of phase shift, there will be cancellation from subs to tops. Alternatively, there could be a polarity reversal in cabling or in the subs themselves. So until Coaster can figure out how to align his signals, a polarity reversal on the sub signal may help . Trying to figure out the "subs sounded weird" comment.

[agedhorse]

Because the phase shifts happen through the crossover point only, and on a 24dB/octave filter set, the phase will shift on one band from 0 to -360 degrees and on the other filter set from 0 to +360 degrees. Where would you like them to match up? At that point, every other point will not. It's a continuous transfer function.

 

Within the passbands, the phase shift will be zero, outside ofthe passbands, the phase changes with frequency. The phase does not reach the full 360 degrees for example until maybe 1 octave outside of the passband where the amplitude approaches 24dB down.

 

Look up the phase response of a high pass and low pass filter, you will immediately see that there is no solution that reversing polarity will correct.

[jasps]

 

Because the phase shifts happen through the crossover point only, and on a 24dB/octave filter set, the phase will shift on one band from 0 to -360 degrees and on the other filter set from 0 to +360 degrees. Where would you like them to match up? At that point, every other point will not. It's a continuous transfer function. Within the passbands, the phase shift will be zero, outside ofthe passbands, the phase changes with frequency. The phase does not reach the full 360 degrees for example until maybe 1 octave outside of the passband where the amplitude approaches 24dB down. Look up the phase response of a high pass and low pass filter, you will immediately see that there is no solution that reversing polarity will correct.

 

OK. I understand what you are saying and agree. The crossover from tops to subs will benefit from a polarity reversal if energy in the crossover region is different in polarity. That is something delay cannot fix.

[agedhorse]

Polarity also matters when combining signals that are in the same passband. Ie. multiple cabinets or drivers of the same bandwidth.