I still can't believe that you are not doing this deliberately. Let's look at the table for 2HD 3HD and 5HD and convert percentage into db.
2HD = 0.0849% = -61.4db. Chart confirms.
3HD = 0.0399%= -68.0 db. Chart says -56 db. Not even close.
5HD = 0.0122%= -78.3 db. Chart says -57 db!
Do you see? Your software or use of it has a problem. Now please stop being so obtuse and ultra defensive. My first post was quite clear.
I am not going to waste my time anymore. Read the manual - the displayed values are fractional percentages of the total power not same as picking off the graph peak value.
I am sorry but I completely miss your point. I stand by all that I said back then, it was not wrong.
The total power is almost identical to the graph peak value since the power sum of harmonics is extremely low.
Latest HD measurements of a SS 10F/8424 in a tractrix. 1.64volts drive, mic at 0.5m, produces 98.6dB with 1kHz excitation. RTA shows almost no distortion components above 2nd harmonic. This is equivalent to 92.6dB at 1m, with a little over 300mW of power. One of the cleanest sounding speakers I have ever heard. Again, the material used for the speaker is foam core with latex caulking sandwich CLD.
More info here:
http://www.diyaudio.com/forums/full...ll-range-tractrix-synergy-80.html#post4509142
More info here:
http://www.diyaudio.com/forums/full...ll-range-tractrix-synergy-80.html#post4509142
Last edited:
Does seem something is weird about how REW is doing its match for the %ages on the table, which leaves you wondering. That said, ~56 dB down on the 3rd/5th ain't bad.
Last edited:
Oh, yeah, I was talking about the previous set of traces, not the 10F/8424 measurements.
According to the manual, the values called out in the table are fractional power of harmonic order calculated by order of interest divided by total power (sum of all peaks including fundamental - so you can't just pull of the value from the ordinate of the graph).
Edit -The manual says:
Edit -The manual says:
Last edited:
Again, I don't think I'm trying to be argumentative, but, simply put, we're talking orders of magnitude, while the correction you're mentioning will be small, since the fundamental is so large. E.g sum of fundamental scaled to 0 dB and -60 db (0.001) = 1.001. Pretty much just the fundamental; even if you add in 9 harmonics at -55 dB, you're still looking at 1.016.
Makes me wonder if they're somehow hashing in the integral of the noise? But that still doesn't seem to account for the discrepancy. Either way, all these calcs should be done relative to the fundamental.
Makes me wonder if they're somehow hashing in the integral of the noise? But that still doesn't seem to account for the discrepancy. Either way, all these calcs should be done relative to the fundamental.
Last edited:
Agreed. There's an issue. The quotes from the manual (both times) say nothing to appease the concern. In fact, they support the concern. However, don't expect any further detail from the chart owner. I was very disappointed in his attitude. When Earl Geddes raises a concern he says "I am not going to argue with you anymore as you are set in your beliefs." When I raise a concern he says "I am not going to waste my time anymore." There appears to be some sort of learning difficulty on show, but it is not necessary to diss everyone else in the thread.
I show data and what I got was constant disbelief from Gedlee saying it could not be be so based on experience vs data. Your first reaction was pretty disrespectful in tone - and continued by accusing me of being obtuse. Then now accusing me of having a learning issue. Let's not kid ourselves who's got the nasty attitude and chip on his shoulder. If you want to have a healthy discussion it would start with asking how the values are calculated and let's look at them. Not the way you went about it. If there is a math issue I can extract values using a cursor for each peak and put it on a spreadsheet. Maybe we can make progress.
Last edited:
Not the 1st time your measures have been called into question. And if Earl is right, not the 1st time you have been shown to present iffy data.
I don't see eye-to-eye with Earl, but i'd tend to side with him.
Instead of dissing the constructive criticism, you should at least double check, you are still a rookie with much to learn (hopefully we areall still learning).
dave
You were the one who called into question my measurement of the CHN-70, claiming my mic was probably bad. It was this plot showing how the CHN-70 as measured by me in an OB and a sealed tapered TL was not the prettiest response, and not anything like the advertised response by the manufacturer.
Subsequent to this, my measurement was independently confirmed by others. As to my "bad" UMM-6 mic, I got another mic, a UMIK-1 that was independently calibrated by Cross Spectrum Labs. A measurement of the same speaker (not the CHN-70) shows that UMM-6 was pretty good and within 2dB of the calibrated UMIK-1 (graph vertical div is in 2dB/div):
There were some data I presented earlier that did not have the gate set to a shorter time to remove room reflections - but that was just a setting in the software. What "iffy" data are you referring to?
As to siding with Earl Geddes, you must have said this 1000 times on diyAudio:
Last edited:
One of many and most not by me (on the forum or by private correspondence).
And are you sure it was specifically about that? After all i was the one that posted a compare of yours vrs the factory, and they were broadly the same, yours lacked a lot of detail shown in the factory graph.
dave
Attachments
Bare,
Point well taken. Let's get back on topic of the best material, but the problem is "best" here means helping a speaker to achieve low HD. But mention of ultra low HD around here degrades to "there must be a problem with the software or the measurement" because it can't be that low. So maybe the best thing to do is to not talk about low HD anymore and talk about how easy it is to work with, how light it is, how inexpensive it is, etc.
Point well taken. Let's get back on topic of the best material, but the problem is "best" here means helping a speaker to achieve low HD. But mention of ultra low HD around here degrades to "there must be a problem with the software or the measurement" because it can't be that low. So maybe the best thing to do is to not talk about low HD anymore and talk about how easy it is to work with, how light it is, how inexpensive it is, etc.
xrk, in this case, my only comment would be to just talk about how many dB down the harmonics are, and that it'd be, at some point, nice to know that you're not getting anomalous mic/signal chain issues by cross referencing against your other mic. But no rush on that (and you don't really owe us that info).
REW's tabular calculation of distortions is wack, simple as that. 🙂 I mean, it makes sense from a calculation perspective (integrate over all bandwidth), but that doesn't make them right.
That doesn't take anything away from what you've offered. And the point about CLD (if as simplistic as your application) efficacy is valuable.
REW's tabular calculation of distortions is wack, simple as that. 🙂 I mean, it makes sense from a calculation perspective (integrate over all bandwidth), but that doesn't make them right.
That doesn't take anything away from what you've offered. And the point about CLD (if as simplistic as your application) efficacy is valuable.
Hey harmonic distortion has got to be more important then the "HOM" - at least it is so obvious it can be measured. 😉
Here is the PRV 5MR450NDY in the tractrix taken with a UMIK-1 (at 1m) vs UMM-6 (at 2m) shown previously. The HD still seems about -50dB (2nd H) to -70dB (3rd H) below the fundamental.
A few of the CLD panels delaminated - the paper facing separated from the foam core so there was a little more cabinet vibrations this time.
Attachments
Last edited: