Home Theater Forum and Systems banner
1 - 5 of 5 Posts

·
Registered
Joined
·
1 Posts
Discussion Starter · #1 ·
Hi all,

I did a search for this topic here on the forum, but didn't find anything. My apologies if there is something out there...please feel free to direct me to a thread if there is one.

I've attempted a few measurement captures myself using REW, while playing frequency sweeps with a test LP, but I'm not convinced the results are accurate...or else I don't understand how the software is processing results. One part that confuses me is how the software (after clicking 'Start Measuring') performs its sweep in just a few seconds (and I see some type of result), but then the sweep on the test LP lasts much longer, so are they working together somehow? I discovered how to perform smoothing on the graph (1/6 seems to work well), but again, I'm not confident in the raw data.

As far as interface, I going from my phono preamp output into a passive A/D converter (xitel Inport Deluxe), which has a USB output, and then into my PC's USB input. The 'soundcard' is an integrated Realtek on the motherboard. Obviously, this is not 'laboratory-grade' equipment :geek: (not to mention, "old"), so I'm not expecting laboratory results. But, I'm hoping to, at least, have some level of confidence that the results are usable,and consistent. I want to capture FR graphs for the same cartridge, but utilizing different load settings in the phonostage, so even if not 100% accurate in some respects, having consistent results is important, directionally. I hope that makes sense.

I'm guessing that calibration and setting appropriate dB levels will be critical, so any advice is much appreciated! TIA
 

·
Registered
Joined
·
154 Posts
You do not measure cartridge response with REW. What you measure with REW is in the NAME of REW... ROOM EQ Wizard. You measure your ROOM. And your ROOM with overwhelm anything your turntable does during the ROOM measurements. To measure the output of a phono cartridge, you connect an oscilloscope to the output of the cartridge (low level output) or to the output of a phono stage that includes RIAA EQ necessary to reproduce LPs so they sound right. The cartridge measurements will be ELECTRICAL measurements, not ROOM measurements.
 

·
Registered
Joined
·
11 Posts
You do not measure cartridge response with REW. What you measure with REW is in the NAME of REW... ROOM EQ Wizard. You measure your ROOM. And your ROOM with overwhelm anything your turntable does during the ROOM measurements. To measure the output of a phono cartridge, you connect an oscilloscope to the output of the cartridge (low level output) or to the output of a phono stage that includes RIAA EQ necessary to reproduce LPs so they sound right. The cartridge measurements will be ELECTRICAL measurements, not ROOM measurements.
Sorry you haven't been keeping up, Da Wiz. REW has been adding capabilities quite regularly since it was introduced. With the right interface it is now a powerful piece of bench test equipment, and widely used that way.

But back to the point...measuring cartridge response is tricky in general, for a couple of reasons. First, you are limited to test signals created by others--you can't feed a test signal from REW into the cartridge in most cases. You are stuck playing back test records. There are lots of different test records, and the ones that are made to test frequency response don't all agree with each other that closely. And the signals on the test record may or may not be formatted in a way that fits what REW is looking for. Another complication is noise pollution. The signal coming back from the TT/Cartridge/Phono stage will have a variety of noise signals mixed in. Probably the most troublesome one will be the random impulse noise. Most test equipment just doesn't know what to make of that.

In order to get a decent measurement of a frequency sweep, REW would have to be looking for the same length sweep as the one on your test record. And you would need to be able to synchronize the software to the start of the sweep somehow. Historically this has been done: there were test records with tones before the sweep which would trigger a chart recorder, or the operator could be asked to count so many seconds from the end of the cue tone & then start the chart recorder. This may sound primitive, but they were attempts to solve the sync challenge with the tech that was available then. The tech has changed but the challenge is the same. Time the sweep on your test record. Look for a setting in REW that lets you change the time of the sweep it's expecting. Figure out some way to tell when the sweep is about to start, and practice cueing REW till you can get it pretty close. Doesn't sound easy, does it? It isn't.
 

·
Registered
Joined
·
2 Posts
I've attempted a few measurement captures myself using REW, while playing frequency sweeps with a test LP, but I'm not convinced the results are accurate...or else I don't understand how the software is processing results.
Can you tell us some details about the frequency sweep that is recorded on your test LP? Are any specifications provided? How long does a frequency sweep last? What are the start and stop frequencies that are used?

As you have gathered, you need to capture the frequency sweep (chirp signal?) in its entirety. Is there more than one frequency sweep on the test LP? If so, you could capture all of them in one recorded block. Once captured, be it one or more sweeps, you will then need to go ahead and perform some FFT analysis in order to obtain the frequency response of your phono cartridge and phono preamplifier combination. If the sweeps are created in such a way that there is equal energy per hertz as the sweep progresses, this approach will work. I think that requires that the sweep frequency changes linearly.

This is the sort of response I got from FFT analysis of a linear frequency sweep covering 20Hz to 20kHz. There is probably a good mathematical reason for the roll off below 70Hz, but it escapes me at the moment. :(
Rectangle Slope Plot Font Parallel

One part that confuses me is how the software (after clicking 'Start Measuring') performs its sweep in just a few seconds (and I see some type of result), but then the sweep on the test LP lasts much longer, so are they working together somehow?
Those two sweeps are NOT working together at all. I don't have a lot of experience with REW, but I don't think that you can use it to simply capture an audio signal, but I could be wrong. The software Audacity can be used to do that, as can other DAW software.

I discovered how to perform smoothing on the graph (1/6 seems to work well), but again, I'm not confident in the raw data.
Yes, smoothing could be helpful, and 1/6-octave is a good compromise between maintaining resolution and averaging out the effects of noise from the LP itself. I'd tend to try 1/12-ocatve as well. Of course, all of that relies on you getting the captured sweep response file into REW for FFT-based spectral analysis. That's not too difficult, but you need to have the LP's sweep recorded as a .WAV file. I know you can use Audacity to accomplish that without too much trouble.

Obviously, this is not 'laboratory-grade' equipment :geek: (not to mention, "old"), so I'm not expecting laboratory results. But, I'm hoping to, at least, have some level of confidence that the results are usable, and consistent. I want to capture FR graphs for the same cartridge, but utilizing different load settings in the phonostage, so even if not 100% accurate in some respects, having consistent results is important, directionally.
As you are looking to see differences in response between settings, your equipment seems up to the task and should produce quite reasonable results.

I'm guessing that calibration and setting appropriate dB levels will be critical, so any advice is much appreciated!
I think that the main thing would be to get a reasonable signal level into Audacity or another audio recording package, but ensuring that there is no overloading/clipping occurring in the ADC. That's not hard to do, and there's usually enough signal level to get a good S/N ratio to ensure that your results are relatively free from noise (apart from what's on the LP, of course).
 
1 - 5 of 5 Posts
Top