QSC PLX v. Crest Audio CA Series

[tombowlus]

Setting up a shootout I would think would involve a decibel meter, one pre, two amps, one cab, and a blindfold. Set both amps for same sound level, plug and play thru the same cab leaving the pre set flat. Same bass, same player, same bass line, and see if you, blindfolded hear a difference.

But would anyone complain that this is not fair because the gain settings on the amps would be different (unless the amps were rated for - and delivered - exactly the same output)?

 

[tombowlus]

You'd need to do more than just set the volumes the same--I would think you'de need to know the imput sensitivity of each amp and set the preamps so they were seeing the same relative signal level

Exactly!

 

[tombowlus]

To metron and pbj: that's why you need a decibel meter or some toher scientific method to make sure both units are putting out the same volume.

I think...

But that's not the same thing as each amp's output stage seeing the same signal gain, which I think is what Bob is asking for.

Tom.

 

[PB+J]

right--I've used the plx 1202 with two preamps--the sansamp RBI and the Avalon U5. The amp performed completely differently with each--to reach the same volume level with the RBI, I had to have the gain knobs on the PLX at around 2-3 o'clock, compared to having them set at around 10-11 o'clock when I use the avalon. The sans amp has a very different "gain structure" (I think that's the right term?) to the output signal than the Avalon and the plx reacts differently to each--it worked much harder to get the RBI to a volume level I reach at three with the Avalon. You would need to take the input sensitivity of the power amp into account to get a reasonable test.

 

[Kindness]

If I were doing an informal test with the intent of being as “scientific” as reasonably possible I would follow this procedure:

In order to make a fair comparison between power amps every element other than the power amps would stay in the exact same configuration. This prevents the intrusion of unwanted variables. I would actually use prerecorded clips of me playing the bass to ensure my playing doesn’t change between amps. I would then judge the tone of the two amps at equal SPL (using the gain control of the power amps to match SPL) for various levels below clipping (conversation level, drums with brushes level, gig level). Due to the different gain structure and input sensitivity of the two power amps I would not expect the gain settings on the front of the amps to correspond to each other. In fact, I would expect one to be turned up “higher” than the other to get equal output. I might also test the maximum power capabilities of each amp by testing the max SPL prior to clipping, but I can’t imagine wanting to clip either of these amps. Both are too loud for my taste at full power.

In order to do the test “blind,” I would first determine the gain control settings required to match SPL output for each volume level I wanted to test and write those down. I would then cover the amp area (but not the speaker area) and have another person make the gain setting adjustments and switch amps without allowing me to see what was going on. I would do an ABX test by first listening to the QSC at a given level, then listening to the Crest at the same volume, then listening to one of them without knowing which one to see if I could actually tell which was which.

Anyone may correct my suggestion where I am wrong.

 

[Passinwind]

Setting up a shootout I would think would involve a decibel meter, one pre, two amps, one cab, and a blindfold. Set both amps for same sound level, plug and play thru the same cab leaving the pre set flat. Same bass, same player, same bass line, and see if you, blindfolded hear a difference.

If I were doing it, I'd use a standarized signal source, say a 1kHz sine wave recorded to CD, or generated by my computer soundcard. I'd use a dummy load, but using the same speaker for both amps as the test load would be acceptable if you keep the level reasonable. Some folks will claim this is a better load, and in some cases that may be so. Measure equal amplifier output voltage with a good RMS reading voltmeter, which is not all that expensive these days. I'd shoot for a voltage somewhere in the middle of the output capability of each amp, so you'd have room to test both upwards and downwards of that level. This would also bump up the accuracy of your voltage measurement. If you're using a speaker as your setup load, be careful, lest you let out the smoke! Again, I'd recommend a dummy load. A 25 watt one will only run you a few bucks or so.

For the rest of your tests, don't touch the input attenuators on either amp, unless you want to try different sensitivity levels for whatever reason. In that case, re-match as above. Specific gain staging could be an explanation for what many of you may be hearing , so personally I'd use an oscilloscope to ensure that both amps are completely linear at the levels I'm using as I change the preamp volume. A decent soundcard makes a fine o-scope for these purposes, just be careful what you feed it! A speaker level DI box would work, or you could build a simple voltage divider if you have the basic skills to do so. With the right software, you could then do distortion measurements as well, which might be useful.

A human bass player almost certainly will not be able to maintain the needed equal input as amps are switched. I'd suggest recording a few bass riffs and playing those back repeatedly.

I shudder to mention this, but I've been contemplating a little get together and speaker shootout sometime in June or July out here. If someone from the Portland or Seattle area wants to bring in some big iron amps, I might be coerced into setting up a double blind test. I have access to a 300 seat nightclub during off hours, and I could wire up some patch cables and switching to sufficiently randomize the comparisons. Someone else is gonna have to pony for the beer though.

I reserve the right to weasel out of this at any time, and I also welcome better suggestions for testing methodology.

------------------Charlie

 

[Bob Lee (QSC)]

Just curious, Bob, but without a means of measuring the signal being sent to the output stage of an amp (for as cited above, there seems to be some difference in the input sensitivity among various amps), how do you propose that we, the end users, make a "proper" comparison, and ensure that all of our gain settings are "equal?" I can send the exact same signal into each amp, but because of the differences in input sensitivity (which I don't know is the correct term, but it's how I am thinking of it, so please bear with me for now, and correct me as need be), in some ways this is not an "equal" comparison. I can set my gain stages such that the perceived volume on each amp is the same, but that "test" might be "unfair" if one amp is (properly) rated higher than the other.

I have tried listening to amps at a variety of levels, with a variety of gain stage settings. I have intentionally listened with one amp louder than the other, then reversed my volume settings and repeated. I have used mulitiple different instruments, preamps, and cabs with the amps in question. My thoughts were to avoid falling into the pitfall of preferring whichever amp was louder. But you tell me that this is flawed, too. I'm not really sure exactly why my approach is flawed, but I would welcome any suggestions on how to make a more thorough comparison.

I do concede, of course, that as I have been doing all of these tests on my own, I am certainly aware of which amp is playing at a given time, so in theory the placebo effect would apply. But again, I have no preference for any given amp. I own all three (QSC, Stewart, and Crest), I plan on keeping them all, and I am endorsed by none of the companies involved (nor will someone of my meager playing abilities ever be endorsed! ). So yes, I'm not even doing single blind tests, but I would like any instruction that you could give as to how I could make the most relevant comparison between various power amps.

The only two ways that I can think of to determine if I have equal gain settings is to either measure the SPLs coming out of the speaker (but as mentioned, this is not inherently equal if the amps are not rated identically as to power output), or to find some way to test (and equalize) the level of the signal coming out of the amp's intial gain stage, but before the amp's output stage. Is there another, better way that I am not thinking about?

If you could give me detailed recommendations that I can pull off on my own, I would be happy to do my best to implement them. If possible, I will try to get someone else to come over and swap amps for me (should I make them were a blindfold, too, so that we're double blind? ), but I guess first off, I'd like to know how to improve my methodology.

Thanks, Tom.

Hi Tom,

Those are fair questions. The reason most casual comparative listening tests are flawed is that it's relatively trivial to set up two items next to each other and alternately try them out, but it's not trivial to set up their gains to match, so that the comparison is of performance-based sound quality, not loudness-based sound quality.

The most accurate way is to put a fixed signal level into both, measure their outputs, and adjust their gains so that the output voltages match. When I set up tests with an ABX comparator, I use its 1 volt 1 kHz sine wave oscillator to put a calibration signal into both amps, then I adjust each channel gain to put out the same voltage. I shoot for 20.0 volts, ±0.1 volt, so they are no more than 0.08 dB apart if one ends up at 20.1 and another at 19.9. Sometimes because of gain control detents I can't get exactly 20.0 volts on one or both amps, so I set the channels to a voltage that I can get on all of them, maybe 19.8 volts or 22.4 volts, whatever. The important thing is to get them as exactly the same as possible. This usually takes me maybe two to five minutes to fully calibrate two 2-channel amps for an ABX test.

You might not have a sine wave generator and an audio AC voltmeter to do this same sort of thing. Most people don't unless they are true audio geeks. You might be able to get a decent approximation by using the gain control markings. Some QSC amps, like the PLX and PowerLight series, have markings that describe the actual voltage gain in dB. Other amps have markings labeled in dB below full gain, with "0" at full and maybe marks at -2, -4, -6, and so on. For something like this you need to know the amp's full gain. Some manufacturers standardize models or ranges of models on certain maximum gains; for example, the Crest Professional Series models had a standard maximum gain of 32 dB (40x); they might still have that, or maybe they might now have a switch for changing to another gain setting. Check the user documentation if you're not sure. Then you set the gain controls on the amps so that they match up; for example, -3 dB on an amp with a max gain of 32 dB, -7 dB on an amp whose max gain is 36 dB, and 29 dB on an amp that reads gain directly, are all the same thing: 29 dB. However, the precision of the gain controls and markings may be a little off from one amp to another, especially the further down you turn the controls. Imprecision will affect the objectivity of your comparison, but within reason it's still a lot better than a haphazard setup.

Ideally you would have some means of switching between one and the other so you don't have a long time interval between comparing each one to the other. Don't touch the amp gain controls, but instead turn up the source so that you can see how they compare when pushed; if B starts farting long before before A does, for example, it's good to know about so you can consider that in your evaluation. Turn down the sourece, too, so you can hear what low-level signals sound like; if A has noticeable hum or hiss or distortion that B doesn't, that's good to know about, too.

One thing about a double-blind test is that it's not for determining which one sounds better, but primarily for determining if they sound different. You have to first establish that they sound different before you can credibly state that one sounds better than the other. Maybe you can establish that they do sound different from each other, but one doesn't sound better than the other, or maybe you'll find that you can't distinguish one from the other by sound; that's not uncommon, especially for audio products that have essentially objective functions, like well-designed power amps.

A blindfold generally isn't necessary unless you want it. An ABX comparator (Link Removed) makes a double-blind test easy; the device randomly chooses X from A or B, and you can listen to X and A and B and compare as much as you want until you feel you know the identity of X, and you enter or record your decision. Then do it again with a new selection of X, and again and again until you have a series of tests that you can draw some statistically significant results from.

But even fewer people have ABX comparators than have audio generators and voltmeters. So without one of those, do a "same-different" test. Have someone else write down a random sequence of A and B, maybe 20 to 30 steps long, keep it an absolute secret from yourself or whoever is doing the listening, run the same signal into both amps, and have another person plugging and unplugging the speaker cables according to that sequence. So for example, if the first step is B, you listen to it and get a feel for how it sounds, but you don't know that you're listening to A or B. Ready? The person switching disconnects B and then connects the cables to either A or B, whichever is next. You listen; is it the same or is it different? Write down your decision. Then when you're ready, have the person unplug the speaker cables from that amp and plug into the next amp in the sequence, and so on. At the end, compare the same-different decisions with the actual sequence.

Whether an ABX or same-different test, a score of 50% means that you did as well statistically as if you'd tossed a coin to make your decisions. That would mean that it was not proven that there was an audible difference. That doesn't prove that there isn't an audible difference, but similar results from repeated tests can suggest that.

Yes, it's a lot to go through if you really want to make comparisons. But the bottom line is that you have to compare them at the same volume, and you have to be aware that in sighted tests even the most honest or savvy person can easily perceive sonic differences that may not actually exist. I'm not immune to it either, so I rely on double blind tests when I can.

 

[44me]

Tom, here’s a simple way to match gains using, inexpensive test gear. You’ll need a DVM, and a signal generator. First set up the rig with the first amp, and set the gain to a reasonable test volume level. Shut the amp down and disconnect the amp’s input and output.

Turn the signal generator on and set it to a 100mV, 100Hz sinewave. Check the amplitude with the DVM set to 200mVAC. Connect the DVM across the amp output, set to 200VAC. Connect the signal generator to the power amp input. Turn the amp on and monitor the DVM reading. Adjust the signal generator output amplitude until the amp output is at 30V. .

Shut the amp down and connect the second amp to the signal generator and DVM, being sure not to move the signal generator settings. Lower the second amp’s input gain all the way, then turn it on. Increase the gain until you read 30V across the output. Then shut the amp off.

After this procedure, the two amp’s gains will be matched. To A/B test, use the same preamp and speaker cab. Adjusting the preamp settings is OK as long as you use the same settings when you test both amps.

You can probably find a cheap used signal generator on ebay. The cheapest new one I am aware of is made by BK Precision ($200 at Invalid Link Removed). Or, you probably get a suitable signal from your computer sound card if you are good with that kind of thing. To do this with tube amps you will need to use a load resistor (typically 4 or 8 ohms). Use a bit lower output voltage so you don’t need such a high power resistor. 15V into 4 ohms is a little over 50W – a 100W resistor would be OK (careful, it will get hot).

- John

 

[tombowlus]

If I were doing an informal test with the intent of being as “scientific” as reasonably possible I would follow this procedure:

In order to make a fair comparison between power amps every element other than the power amps would stay in the exact same configuration. This prevents the intrusion of unwanted variables. I would actually use prerecorded clips of me playing the bass to ensure my playing doesn’t change between amps. I would then judge the tone of the two amps at equal SPL (using the gain control of the power amps to match SPL) for various levels below clipping (conversation level, drums with brushes level, gig level). Due to the different gain structure and input sensitivity of the two power amps I would not expect the gain settings on the front of the amps to correspond to each other. In fact, I would expect one to be turned up “higher” than the other to get equal output. I might also test the maximum power capabilities of each amp by testing the max SPL prior to clipping, but I can’t imagine wanting to clip either of these amps. Both are too loud for my taste at full power.

In order to do the test “blind,” I would first determine the gain control settings required to match SPL output for each volume level I wanted to test and write those down. I would then cover the amp area (but not the speaker area) and have another person make the gain setting adjustments and switch amps without allowing me to see what was going on. I would do an ABX test by first listening to the QSC at a given level, then listening to the Crest at the same volume, then listening to one of them without knowing which one to see if I could actually tell which was which.

Anyone may correct my suggestion where I am wrong.

I agree wholeheartedly. This is exactly (and I mean exactly!) what I was thinking is the best, most fair and accurate manner that I could think of to test in my own home. Unless I hear any improvements or additional suggestions, I plan on repeating my shootout using this procedure.

Thanks, Tom.

 

[tombowlus]

I shudder to mention this, but I've been contemplating a little get together and speaker shootout sometime in June or July out here. If someone from the Portland or Seattle area wants to bring in some big iron amps, I might be coerced into setting up a double blind test. I have access to a 300 seat nightclub during off hours, and I could wire up some patch cables and switching to sufficiently randomize the comparisons. Someone else is gonna have to pony for the beer though.

If I were in your neck of the woods, I'd be happy to come out (and bring beer!), but that's a bit of a drive for me!

Tom.

 

[tombowlus]

Hi Tom,

Those are fair questions. The reason most casual comparative listening tests are flawed is that it's relatively trivial to set up two items next to each other and alternately try them out, but it's not trivial to set up their gains to match, so that the comparison is of performance-based sound quality, not loudness-based sound quality.

The most accurate way is to put a fixed signal level into both, measure their outputs, and adjust their gains so that the output voltages match. When I set up tests with an ABX comparator, I use its 1 volt 1 kHz sine wave oscillator to put a calibration signal into both amps, then I adjust each channel gain to put out the same voltage. I shoot for 20.0 volts, ±0.1 volt, so they are no more than 0.08 dB apart if one ends up at 20.1 and another at 19.9. Sometimes because of gain control detents I can't get exactly 20.0 volts on one or both amps, so I set the channels to a voltage that I can get on all of them, maybe 19.8 volts or 22.4 volts, whatever. The important thing is to get them as exactly the same as possible. This usually takes me maybe two to five minutes to fully calibrate two 2-channel amps for an ABX test.

You might not have a sine wave generator and an audio AC voltmeter to do this same sort of thing. Most people don't unless they are true audio geeks. You might be able to get a decent approximation by using the gain control markings. Some QSC amps, like the PLX and PowerLight series, have markings that describe the actual voltage gain in dB. Other amps have markings labeled in dB below full gain, with "0" at full and maybe marks at -2, -4, -6, and so on. For something like this you need to know the amp's full gain. Some manufacturers standardize models or ranges of models on certain maximum gains; for example, the Crest Professional Series models had a standard maximum gain of 32 dB (40x); they might still have that, or maybe they might now have a switch for changing to another gain setting. Check the user documentation if you're not sure. Then you set the gain controls on the amps so that they match up; for example, -3 dB on an amp with a max gain of 32 dB, -7 dB on an amp whose max gain is 36 dB, and 29 dB on an amp that reads gain directly, are all the same thing: 29 dB. However, the precision of the gain controls and markings may be a little off from one amp to another, especially the further down you turn the controls. Imprecision will affect the objectivity of your comparison, but within reason it's still a lot better than a haphazard setup.

Ideally you would have some means of switching between one and the other so you don't have a long time interval between comparing each one to the other. Don't touch the amp gain controls, but instead turn up the source so that you can see how they compare when pushed; if B starts farting long before before A does, for example, it's good to know about so you can consider that in your evaluation. Turn down the sourece, too, so you can hear what low-level signals sound like; if A has noticeable hum or hiss or distortion that B doesn't, that's good to know about, too.

One thing about a double-blind test is that it's not for determining which one sounds better, but primarily for determining if they sound different. You have to first establish that they sound different before you can credibly state that one sounds better than the other. Maybe you can establish that they do sound different from each other, but one doesn't sound better than the other, or maybe you'll find that you can't distinguish one from the other by sound; that's not uncommon, especially for audio products that have essentially objective functions, like well-designed power amps.

A blindfold generally isn't necessary unless you want it. An ABX comparator (Link Removed) makes a double-blind test easy; the device randomly chooses X from A or B, and you can listen to X and A and B and compare as much as you want until you feel you know the identity of X, and you enter or record your decision. Then do it again with a new selection of X, and again and again until you have a series of tests that you can draw some statistically significant results from.

But even fewer people have ABX comparators than have audio generators and voltmeters. So without one of those, do a "same-different" test. Have someone else write down a random sequence of A and B, maybe 20 to 30 steps long, keep it an absolute secret from yourself or whoever is doing the listening, run the same signal into both amps, and have another person plugging and unplugging the speaker cables according to that sequence. So for example, if the first step is B, you listen to it and get a feel for how it sounds, but you don't know that you're listening to A or B. Ready? The person switching disconnects B and then connects the cables to either A or B, whichever is next. You listen; is it the same or is it different? Write down your decision. Then when you're ready, have the person unplug the speaker cables from that amp and plug into the next amp in the sequence, and so on. At the end, compare the same-different decisions with the actual sequence.

Whether an ABX or same-different test, a score of 50% means that you did as well statistically as if you'd tossed a coin to make your decisions. That would mean that it was not proven that there was an audible difference. That doesn't prove that there isn't an audible difference, but similar results from repeated tests can suggest that.

Yes, it's a lot to go through if you really want to make comparisons. But the bottom line is that you have to compare them at the same volume, and you have to be aware that in sighted tests even the most honest or savvy person can easily perceive sonic differences that may not actually exist. I'm not immune to it either, so I rely on double blind tests when I can.

Thanks, Bob! That is helpful. I think that Passinwind's idea of burning a signal onto a test CD makes sense. Do you think that would be okay?

As for a volt meter, I do have a multi-range volt meter from Radio Shack. Is this going to be up to the task, or do I need something more beefy?

Finally, would you anticipate that if I set the output voltages to match, then I would also be reading matching SPL's with each amp?

Thanks, Tom.

 

[Passinwind]

If I were in your neck of the woods, I'd be happy to come out (and bring beer!), but that's a bit of a drive for me!

Awww, we already have really good local beer here, so you can just bring a truckload of cabs instead, OK?

 

[tombowlus]

John:

Thank you as well for your suggestions (which seem to be mirroring what Bob and Passinwind are saying).

A big thank you to all who have responded to my request for assistance in developing as good of a testing procedure as possible. It is no secret that I am a gearaholic, and part of my affliction is that I really enjoy listening to different gear and appreciating their sonic qualities. I suppose I do have a "bias" in the sense that I enjoy listening for differences in sound, and to that extent, the placebo effect surely applies in that my brain is programmed to believe their is a difference (though I stand by the notion that I honestly have no preference for Brand A over Brand B).

I will definitely get one or two buddies involved to keep me honest!

Give me a couple of weeks to find time when the wife and kids are out of the house, and I'll see what I can discover!

Tom.

 

[tombowlus]

Awww, we already have really good local beer here, so you can just bring a truckload of cabs instead, OK?

Oh, you know it! I was going to just hit the nearest microbrewery on my way in! I have been to Oregon (my sister lives in Lebanon, and Bend before that) and Washington, so I know how good the beer (and wine!) is out there. Not to mention salmon and halibut!

 

[geshel]

It's not a dispersion issue, it's a projection issue. It is my finding that these switching amps are running out of current prematurely and not sustaining the note. This IMO causes the sound of the low end to fall of much sooner (i.e. closer to the cabinet),

Can you explain: "much sooner = closer to the cabinet". I don't understand how shortening the time the signal is sent to the cabinet shortens the distance the wave travels through the air.

 

[Bob Lee (QSC)]

Thanks, Bob! That is helpful. I think that Passinwind's idea of burning a signal onto a test CD makes sense. Do you think that would be okay?

As for a volt meter, I do have a multi-range volt meter from Radio Shack. Is this going to be up to the task, or do I need something more beefy?

Finally, would you anticipate that if I set the output voltages to match, then I would also be reading matching SPL's with each amp?

Thanks, Tom.

Sure! There are some useful test CDs out there. Michael "Bink" Knowles, the guy who organized last year's ProSoundWeb amp shootout, put out a pretty good one. With my marital status I can't buy all the test gear I'd like--occasionally I can get something through the Senate --so I use Bink's CD and an available CD player as my usual test signal source when I'm doing my own stuff at home.

I don't know about the Radio Shack voltmeter. Some are intended mainly for measuring power line AC and thus have a very limited frequency range. It doesn't hurt to try it and see if it gives useful readings at audio frequencies.

If you match the output voltages of solid state amps, then the SPLs should be the same through the same loudspeaker. Since most have a damping factor of at least a couple hundred, there's not going to be anything more than a very minuscule drop in voltage when you connect a loudspeaker load. Tube amps may be a different story because they often have higher output impedances, so you should measure them with the load connected.

 

[tombowlus]

Sure! There are some useful test CDs out there. Michael "Bink" Knowles, the guy who organized last year's ProSoundWeb amp shootout, put out a pretty good one. With my marital status I can't buy all the test gear I'd like--occasionally I can get something through the Senate --so I use Bink's CD and an available CD player as my usual test signal source when I'm doing my own stuff at home.

I don't know about the Radio Shack voltmeter. Some are intended mainly for measuring power line AC and thus have a very limited frequency range. It doesn't hurt to try it and see if it gives useful readings at audio frequencies.

If you match the output voltages of solid state amps, then the SPLs should be the same through the same loudspeaker. Since most have a damping factor of at least a couple hundred, there's not going to be anything more than a very minuscule drop in voltage when you connect a loudspeaker load. Tube amps may be a different story because they often have higher output impedances, so you should measure them with the load connected.

You know, you just reminded me that I think at least one of the Stereophile Test CDs have a variety of test tones on them. I could probably use that.

As for a volt meter, mine is substantially similar to this one:

[Invalid or Expired Link Removed]

Is this going to be adequate, or do I need something bigger? Also, any idea where I can buy a dummy load if I want to test tube amps?

Thanks, Tom.

 

[Passinwind]

Also, any idea where I can buy a dummy load if I want to test tube amps?

Digikey, Mouser, or Parts Express all can hook you up.

 

[Bob Lee (QSC)]

You know, you just reminded me that I think at least one of the Stereophile Test CDs have a variety of test tones on them. I could probably use that.

As for a volt meter, mine is substantially similar to this one:

[Invalid or Expired Link Removed]

Is this going to be adequate, or do I need something bigger? Also, any idea where I can buy a dummy load if I want to test tube amps?

Thanks, Tom.

I looked at the owner's manual, and it only mentions 50–60 Hz for the AC voltmeter. It doesn't specifically say the meter doesn't respond to higher audio frequencies, but I'm inclined to believe that if it did they would be proud enough of that to mention such in the manual. But like I said, there's no harm in trying it.

And I echo Passinwind's suggestions of Mouser, Digi-Key, and Parts Express. Parts Express (www.partsexpress.com) is audio-oriented and probably the most user friendly for non-engineers of the three.

 

[Mudbass]

If I were doing it, I'd use a standarized signal source, say a 1kHz sine wave recorded to CD, or generated by my computer soundcard.

But this won't tell you how an amp performs with a highly dynamic audio source like a bass guitar. A CD isn't very demanding and just about any amp is going to sound just about the same as any other amp. It seems the same as trying to compare the performance of two engines while at idle.