|
Post by homestudiosimplified on Jul 20, 2016 20:52:52 GMT
I like to do reviews on products that I use in my Home Studio. It is a service to those looking to buy that same piece of gear later or maybe just to geek out for a bit.
My question is, what would be the best method to give and accurate representation of a preamp? I am wanting to do a review on the Art Pro MPA II, however, I know it's going to be hard to give unbiased and accurate results. Any suggestions?
|
|
|
Post by Hexspa on Jul 21, 2016 23:36:49 GMT
I'm not the guy to ask but..
Null test?
|
|
|
Post by Ethan Winer on Jul 22, 2016 20:23:54 GMT
If I were to assess the quality of a preamp I'd measure as many details as possible: frequency response, distortion at soft and loud volumes, noise at low and high gain settings, etc. Most people don't realize it's actually pretty easy to measure these things with a good sound card and an audio editor program that can create test tones and display an FFT. (I use Sound Forge.)
I'm skeptical of listening tests because any preamp that's halfway decent won't have any sound of its own. Unless it's adding distortion intentionally, in which case all bets are off. Further, distortion you find pleasing others may love, or hate. Plus, every source responds to distortion differently anyway.
In your Facebook PM I think you asked about using re-amping, and that's a reasonable way to get Wave files for comparison. The only issue is matching the volume levels of your "reference" preamp to the preamp being compared. Of course, you also have to use the identical loudspeaker and music source, microphone, and mic position. But such comparison isn't really much of a test, because all you learn is how this preamp differs from another preamp. I think it's much more useful to know how the preamp you want to know about actually performs.
--Ethan
|
|
|
Post by Ethan Winer on Jul 22, 2016 20:25:22 GMT
Yes, a null test could be useful too. I'm actually working on designing a professional quality device that will do null tests in real time. It's a huge project! So it will take a few more months before I have a prototype.
|
|
|
Post by arnyk on Aug 8, 2016 8:51:21 GMT
I like to do reviews on products that I use in my Home Studio. It is a service to those looking to buy that same piece of gear later or maybe just to geek out for a bit. My question is, what would be the best method to give and accurate representation of a preamp? I am wanting to do a review on the Art Pro MPA II, however, I know it's going to be hard to give unbiased and accurate results. Any suggestions? As far as technical tests go, a very good starting point is the Audio Rightmark program. It does a fairly credible suite of tests and provides the results in several formats including completely marked up web pages. All you need is a Windows computer with a fairly good audio interface capable of doing full-duplex. A good cheap audio interface that I can use for this is the M-Audio AP 24192, now only available as used gear. With the right software it outperforms the original Audio Precision test set and is not that much worse than the current ones. Other audio testing software to look at: Arta Visual Analyzer
|
|
|
Post by arnyk on Aug 8, 2016 8:53:18 GMT
IME null testing is worse than useless because everything wrong gets lumped together into one number.
|
|
|
Post by Ethan Winer on Aug 8, 2016 16:57:08 GMT
I think null testing is great. Seriously, one great feature is it shows everything that's added (or removed), including stuff you might not think to look for. This dispels the common audiophile notion that "Science doesn't know everything people hear," or "Some aspects of audio can't be measured." Another feature is you can use music instead of test signals, again taking away ammunition from those who don't understand Fourier and believe music is somehow more complex than sine waves. Once you isolate the nulled residual you can then analyze it for level and spectrum with an FFT or even listen to it, or pass it through an A-Weighting filter. --Ethan
|
|
|
Post by Ethan Winer on Aug 8, 2016 16:58:43 GMT
It doesn't have to be one or the other either. So you can measure the standard stuff and also assess a nulled residual. Sort of like seeing a frequency response and also a waterfall for a room.
|
|
|
Post by arnyk on Aug 11, 2016 9:04:39 GMT
I think null testing is great. Seriously, one great feature is it shows everything that's added (or removed), including stuff you might not think to look for. This dispels the common audiophile notion that "Science doesn't know everything people hear," or "Some aspects of audio can't be measured." Another feature is you can use music instead of test signals, again taking away ammunition from those who don't understand Fourier and believe music is somehow more complex than sine waves. Once you isolate the nulled residual you can then analyze it for level and spectrum with an FFT or even listen to it, or pass it through an A-Weighting filter. --Ethan The biggest problem with nulling is that it is often impossible to null out differences that in fact cause no audible problems at all, particularly certain kinds of phase shift and timing errors. The better nulling software tries to address this, but it is far from perfect in that regard. For example nulling tests will usually puke all over perceptually coded files as compared to the wave file predecessors, but if you do DBT comparisons, the audible differences are vanishing or null. One can do frequency response, phase, IM and THD tests using music as the test signal without resorting to nulling. Some measurement software offers this as a standard feature. SMAART, for example.
|
|
|
Post by Ethan Winer on Aug 11, 2016 21:29:21 GMT
I agree that nulling doesn't works on some sources, and even when some sources can't null to even -20dB they can still sound identical. I'm not interested in trying to null things that drift, and I've tried and rejected DiffMaker as simply not working very well. But for some things null tests are great. It can prove that two wires sound identical, no matter what some delusional audiophile magazine writers believe. It can also work on a lot of electronic gear, though you might need to band-limit the audio if the device or devices being compared have phase shift near the frequency extremes. I'm halfway through designing a serious implementation of this concept, so in a few months we'll all know for sure what's possible. --Ethan
|
|
|
Post by arnyk on Aug 12, 2016 1:14:48 GMT
I agree that nulling doesn't works on some sources, and even when some sources can't null to even -20dB they can still sound identical. I'm not interested in trying to null things that drift, and I've tried and rejected DiffMaker as simply not working very well. But for some things null tests are great. It can prove that two wires sound identical, no matter what some delusional audiophile magazine writers believe. It can also work on a lot of electronic gear, though you might need to band-limit the audio if the device or devices being compared have phase shift near the frequency extremes. I'm halfway through designing a serious implementation of this concept, so in a few months we'll all know for sure what's possible. --Ethan The linear distortion and nonlinear distortion in wire is on the small side, but if you have really good measuring gear, you can measure it. The linear distortion is usually due to interactions with the signal sources and the loads. This can be fairly large and measured using standard measuring gear. It can also be audible. The nonlinear distortion is usually dominated by thermal modulation. This is more subtle, and requires some special gear and techniques to measure. This is generally so low that it can't possibly be heard.
|
|
|
Post by Ethan Winer on Aug 12, 2016 14:37:50 GMT
^^^ Good stuff Arny.
|
|
|
Post by homestudiosimplified on Aug 15, 2016 9:08:55 GMT
Well... My mind is blown (small explosion) as I am not quite as learned as you all on these topics. I have never performed null tests nor have I went into the depth that Mr. Winer has explained. I will do my best with the suggestions given to present a well rounded review. Thanks guys!
|
|