Its very easy. With Reaper, find the first transient waveform at the beginning of the sound file. Zoom in to see individual samples. Even if you had two files that were exactly the same except they were recorded at different times, the sample points would be located at different phases of the waveform. I can do that very very quickly as sometimes I need to time align files to as close to sample accurate as possible. I can get them to within a fraction of one sample period.
EDIT: Anyway, in case its not obvious by comparing the sample points to the where they lay on the waveform (the relative phase of the A/D clock with the waveform phase), it would be easy to match up which file is being played by foobar. I could see such a difference in the Toccata files, so not a problem. Even if everything was synced to one clock: file start time, D/A, A/D, any differences between files are easily seen on a sample by sample basis. It may be necessary to normalize them up to get enough magnification, but that is also quick and easy to do.
Last edited:
Sure, and when I report I can repeatably differentiate blind and at multiple sittings, I am doing so in good faith. If I wanted to fool you, you know I could have. No interest in doing that.
I often look at the spectrum through a phone when I listen to concerts, the higher SPL always are around 60Hz~40Hz, so it really makes a difference how the speaker handles this. There are lots of other issues that can cause not so optimum sound at high levels, so we cannot all attribute this to speakers unless we actually traced the problem to the source. This why I am interested in listening to Earl’s designs.
The only thing I understand about your questions is that it is causing you to lose sleep and having headaches. Not sure if anyone has a cure or not. Maybe you don’t understand the answers either.
Have you gotten a satisfactory answer to nonlinear vs linear distortion?
Also, while I certainly appreciate your frustration (I'm an engineer who works in medicine, so I totally understand being dropped into new lexicons on a daily basis), it's really going to reap benefits for you to understand what the "conventional" language is for whatever domain you're trying to step into.
First off, play around with two waves interacting, hopefully this gives you some feel for what we're discussing.
Simplest way to look at it is a perfect system (distortionless) would get an input and we get a scaled output. We're talking flat from DC to daylight (and on past). Output = Input * scaling_factor. Distortion is anything that deviates from perfection.
Linear distortion means that the scaling_factor is not the same for all input, and affects amplitude and phase of a signal -- so any dips/peaks in an amplitude response (from something like a very slowly swept tone where one only looks at the tested-frequency component), would be linear distortions. If we're talking speakers, then the bandwidth limitations of the speaker, and the wobbles in the output response are largely due to the linear distortions in the speaker.
Nonlinear distortion is the creation of new frequency products. There are multiple processes by which this happens, but using the link above, try this for me: place f1 at 20 hz and f2 at 40 hz. Now sweep "A2" from 0 to roughly 0.25. Looks a lot like a clipped signal, right? So, looking backwards, if we see an input that looks like the blue line and an output that looks like the purple line, we know that there is a very large 2nd harmonic being generated (the red line).
Clear as mud?
The typical test of harmonic distortion is just one of the forms. I believe that IMD is another and more important form of nonlinear distortion. There are others as well.
Thus researchers measuring harmonic distortion in isolation, a long with the difficulties and fatigue associated with such testing means that often harmonic distortion appears to be not a huge deal. Further its the spectrum of harmonic distortion that matters, which adds additional complexity.
Most people know that distortion goes up more with volume, less with horns and pro gear, but this is not easily quantifiable.
Not to mention that there are other factors like the elusive quality called dynamics in which something sounds larger regardless of spl. I've inquired about this, though not to much positive result.
Thus researchers measuring harmonic distortion in isolation, a long with the difficulties and fatigue associated with such testing means that often harmonic distortion appears to be not a huge deal. Further its the spectrum of harmonic distortion that matters, which adds additional complexity.
Most people know that distortion goes up more with volume, less with horns and pro gear, but this is not easily quantifiable.
Not to mention that there are other factors like the elusive quality called dynamics in which something sounds larger regardless of spl. I've inquired about this, though not to much positive result.
Last edited:
I have. Remember that Lidia is an audiologist so she has all the tools and has taken them to concerts. Live rock can get ridiculously loud.
That's an interesting comment, I just read a blog of a fairly serious audiophile recently that went out and borrowed an SPL meter out of curiosity. He has a lot of "audiophile" opinions but his broad tastes and generally sincere opinions interest me. He found that unless he wants to "blow" folks away the highest PEAK SPL at listening position with very dynamic music was 94dB, with most music 89db or so. Some of the averages for very comfortable listening were surprisingly low.
This highlights see the problem of the complexity of the subject because what you are saying is not true. IMD and HD and "others as well" are not different types of nonlinearity, they are just different results from differnt input signals to the exact same nonlinear system. The underlying nonlinear system is exactly the same in each case.
It amazes me how many times I have to make this correction.
Was that dB(A) or dB(C)? I'll bet "A" which is incorrect in this situation, because dB(A) is the default on most meters.
I avoid it like the plague. I have not measured it but I seriously doubt I find more than 70dBA or so average SPL at home comfortable. I find myself turning down every device that any family member has recently used.
Fair enough, but since I typically don't like loud music especially bass I suspect the difference is less for me. I should try it.
EDIT - The author did not specify but the meter had both A and C.
Last edited:
Its true that a nonlinearity will exhibit THD or IMD depending on the test signal used. But its also true that some things like an amp can have THD that varies as a function of frequency or with power levels. So, it would seem using different test signals could possibly provoke different responses if the nonlinearity is, shall we say, in itself nonlinear.