REW repeatability?

**robertcharles123** · **10-03-09, 04:58 AM**

I tried another amp and no change in graphs, so amp seems to be ok.

Next, I ran the 10 Hz, 20 Hz and 40 Hz tones- I got 89.5 dB for 10 Hz
93.4 dB for 20 Hz
110.2 dB for 40 Hz

When that 40 Hz tone started, it was very intense.

Also, I ran a 100HZ to 1HZ sweep and it started at about 86 dB and stayed pretty much between 85 and 93 and got no lower than 86 on the meter before it stopped and the meter dropped off instantly. If correction values were able to be added, The numbers would be very high at the lower Hz readings. The 100Hz to 1Hz sweep seemed to be more level and much more intense at the end of the sweep than the 0-200 sweep is at the beginning.

When I do the 0-200 sweep, the sound seems to build and has a few nulls in the sweep, but the 100Hz to 1Hz sweep did not have that type of reaction. It actually felt like it built up as the Hz decreased and it showed it on the Galaxy CM140. If I were to be able to do a reverse sweep on REW, I would definitely see better results. Any way to do reverse sweep on REW?

**brucek** · **10-03-09, 08:46 AM**

Use the "Frequency Tracks Cursor" checkbox with the signal generator and you can drag the cursor around the REW graph plot and the frequency will follow. Turn on the signal generator and drag the cursor around to create any manual sweep you like.

You could also do a standard measurement first so you have a plot on the graph to compare the sound against while you use the 'Frequency tracks cursor' feature. If the plot shows a dip, drag the cursor there and move it around the dip and see if the sound agrees.......

brucek

**robertcharles123** · **10-04-09, 06:28 AM**

I just found this little bit.

[PDF]The Legend Lives On - The Dolphin Group - MK Sound ...
Vented subwoofers and subwoofers with passive radiators roll off too sharply (24 . dB/octave) to take advantage of Room Gain. ...

http://www.thedolphingroup.org/pdfs/...d_march_09.pdf - Similar pages

Would this have anything to do with the funny looking graphs I am getting?

I did not have a chance to do more measures, but I did see the generator and curser option.

Thanks,

Robert

By the way, I am going to test a sealed sub in my HT and see how that goes.

**robertcharles123** · **10-04-09, 11:11 PM**

Quote:

brucek wrote:

Use the "Frequency Tracks Cursor" checkbox with the signal generator and you can drag the cursor around the REW graph plot and the frequency will follow. Turn on the signal generator and drag the cursor around to create any manual sweep you like.

You could also do a standard measurement first so you have a plot on the graph to compare the sound against while you use the 'Frequency tracks cursor' feature. If the plot shows a dip, drag the cursor there and move it around the dip and see if the sound agrees.......

brucek

OK. I have a quick question I hope could make a difference. I used the cursor feature and just played around to get the feel for it. Before I get to involved with it, I need a little bit more info if possible. When I drag the cursor onto the frequency and let it play, I seem to be getting the SPL reading at the Galaxy at the listening position that corresponds to the number on the Galaxy. If there is a calibration file being added to the graphs, shouldn't the readings be of by the cal file number corresponding to the frequency being tested. For instance, when I got a reading of 64 dB on the meter, it read 64 on the graph for 10 Hz. If a cal file was used to create the graph, shouldn't the reading be 64+13.45 dB?

Also, I ran some sine 10 Hz waves and put the SPL where the mic was in a corner for some of the measures. I get a 15 dB higher reading than when I went right up to the cabinets. But when I run the sweeps, I do not get that higher reading on the graphs. It still shows a serious roll off. Any thoughts? I can actually feel the pressure increase tremendously when I walk from the cabinet to the corner of the room. So there is room gain down low, but I quess I won't be able to benefit from it because of the position of the cabinets.

Thanks,

Robert

**brucek** · **10-05-09, 07:27 AM**

Quote:

If a cal file was used to create the graph, shouldn't the reading be 64+13.45 dB?

The soundcard and meter calibration files are only added after a response sweep measurement is taken and applied to the graphs.
The SPL Calibrate routine is a rough calibration using a band limited pink noise to ensure the real SPL meter and REW SPL meters are within a ball park of each other.

Quote:

But when I run the sweeps, I do not get that higher reading on the graphs.

I would think you should.

brucek

**robertcharles123** · **10-05-09, 11:33 AM**

Quote:

brucek wrote:

The soundcard and meter calibration files are only added after a response sweep measurement is taken and applied to the graphs.
The SPL Calibrate routine is a rough calibration using a band limited pink noise to ensure the real SPL meter and REW SPL meters are within a ball park of each other.

I would think you should.

brucek

So, would it be possible to 1st take a graph and use that graph to see the effects of a cal file? I would like to look at one graph and edit that graph. Is that possible?

Or, do I just have to take a graph with the cal file and then take another without?

Thanks,

Robert

**brucek** · **10-05-09, 01:21 PM**

Quote:

So, would it be possible to 1st take a graph and use that graph to see the effects of a cal file? I would like to look at one graph and edit that graph. Is that possible?

Sure, although I can't imagine why since the mic being used requires some offset from the cal file to render it flat.

Either way, simply clear the mic cal file and take your measure. Then if you want to add the cal file to that measure, load the cal file and click Apply Windows in the IR Windows popup. This will apply the file.

You can in fact, take any measurement you've already taken and remove the cal file effect by clearing the file and clicking Apply Windows. You see, all your measurements with REW are taken without any soundcard or mic calibration files - they are simply added later. REW takes a raw measurement and creates the impulse response. When the frequency response plot is generated from that IR response, the cal files are then added. So, you can add or remove them from the response graphs whenever you want.

brucek

**robcart65** · **10-10-09, 01:11 PM**

Thanks a lot

**robertcharles123** · **10-15-09, 10:59 AM**

Quote:

brucek wrote:

Sure, although I can't imagine why since the mic being used requires some offset from the cal file to render it flat.

Either way, simply clear the mic cal file and take your measure. Then if you want to add the cal file to that measure, load the cal file and click Apply Windows in the IR Windows popup. This will apply the file.

You can in fact, take any measurement you've already taken and remove the cal file effect by clearing the file and clicking Apply Windows. You see, all your measurements with REW are taken without any soundcard or mic calibration files - they are simply added later. REW takes a raw measurement and creates the impulse response. When the frequency response plot is generated from that IR response, the cal files are then added. So, you can add or remove them from the response graphs whenever you want.

brucek

I see how that process is performed and I like thank you for the info.

Another ?

I have an older computer and way less than 1 GB RAM. I seem to be leaning toward that being the reason why I am getting some really oddball measures. Maybe not the complete reason, but maybe something to it???

Any thoughts?

Thanks,

Robert

**brucek** · **10-15-09, 11:43 AM**

Quote:

I have an older computer and way less than 1 GB RAM. I seem to be leaning toward that being the reason why I am getting some really oddball measures. Maybe not the complete reason, but maybe something to it???

I can't say for sure, but a computer with "way less than 1 GB" may be a bit challenged to run REW. If the operating system gets busy with some other duties and you get some hitching during the sweep, it can offer some poor results. You do require a clean sweep - if you hear any hitch, redo it.

brucek

**robertcharles123** · **10-15-09, 11:48 AM**

Quote:

brucek wrote:

I can't say for sure, but a computer with "way less than 1 GB" may be a bit challenged to run REW. If the operating system gets busy with some other duties and you get some hitching during the sweep, it can offer some poor results. You do require a clean sweep - if you hear any hitch, redo it.

brucek

I have definitely experienced "hitching".

A new computer should do it. They all have multi GB of RAM.

We will see. Either way, I will know the measures are close once another computer is used.

Thanks,

Robert

**glaufman** · **10-16-09, 10:57 AM**

Quote:

robertcharles123 wrote:

I have definitely experienced "hitching".

A new computer should do it. They all have multi GB of RAM.

We will see. Either way, I will know the measures are close once another computer is used.

Thanks,

Robert

If it really is just a memory issue, you usually can add memory to an old computer for a fraction of what it would've cost when the computer was purchased, and for a fraction of the price of the new computer... I just refurbed my wife's machine, maxing out the memory for $35 delivered where it would've been hundreds when we bought it. Not quite as fast as my machine with the modern processor, but almost...

Advertisement