Shiva Graphs to compare with SD-12

I figured out REW over the weekend, and here are 2 graphs I made with my 88.5liter vented shiva tuned to 25hz.
First, near field. The SPL was calibrated around 85:
first sweep shiva.jpg
Next is taken around 8 feet away, as close as the wires would allow to listening position. SPL at the proper 75db as I figured out what I was doing. There are 2 sweeps here to see if the outcome was repeatable:
far measurement shiva.jpg

Any advice on these? I am building a new sub to replace this one and will measure it in the same spot when completed.

Looks pretty good.

You could use some equalization to lower the peaks from 40Hz to 100Hz though.

Eliminating the peaks would bring out the low end better.


far%20measurement%20shiva.jpg


brucek

So I would use something like the BFD and set fairly wide notch filters centered at those frequencies? Of course I will wait until I get the new sub running and see if those are room effects first. Thanks Bruce!

The best software (free or otherwise) to use for setting up BFD filters is REW. Those peaks would come down easily with a BFD. The notches aren't particularly wide though. REW would find the correct filters to flatten them.

brucek

I guess there is a lot more to REW than just making sweeps... As soon as I get a deal on a BFD I will dig deeper.

Here are my SD12 graphs, for some reason they aren't scaled to just LFE. The first is near field, with the SPL reading like 94db, I forgot to adjust to 75 until the far field.




Where do I tell REW just to graph from say 10-200hz? Would make these easier to compare. I am still waiting to get a good deal on a BFD...

You set the area covered by the graph using the Graph Limits button, 2nd from right in the toolbar. You can also zoom in and out using the zoom buttons on the graph axes.
http://www.hometheatershack.com/room...panel.html#top

In the upper right hand corner of REW there is an icon called Graph Limits. Enter the values shown below in my pic. Every time the graph limits aren't set to these values you can return to the icon and it will have remembered your last entries..

Graph Axis Limits.jpg

brucek

Another issue I am having is that REW thinks there is clipping on the input signal. When I did the near field measurement at nearly 100db, it had no problems, but doing the far tests at the correct 75db, I get the error message. I am sure there is no clipping at this low level. Any suggestions?

REW will report clipping if input samples reach digital full range. You can see the actual captured input signal by looking at the scope tab, if the input reaches 100% anywhere it has clipped. That suggests the input volume is set too high or the SPL meter range is too low. Clipping may only happen at a particularly strong resonance, you can see what is happening to the input level by showing the VU meters and looking at the levels during the sweep.

So the actual range of the SPL meter effects the line output? I think I had it on the 70 scale and set to 75, but I guess I could try it on the 80 scale set 5 db down. Also, why is it not recommended to not use mic inputs? Mine seems to be working fine.

Yes.............

Well, the mic inputs are high gain and so are usually electrically noisy. They also usually supply a small phantom voltage (which you don't want). Their frequency response is usually limited too - not what you want.

brucek

would I be able to see this on the sound card cal?

See what?

brucek

If the mic input was not functioning properly as a line in, would I still get graphs like these? I am wondering if my results are valid basically. Thanks for all your help Ken.

Yeah, you can check it out by first seeing what the response of the mic input to the card is when you do the soundcard calibration routine. This test will show the response of the line-out connected to the mic-in and you'll know if it's worth using at all.

Then you can check with a DC meter to see if there is any DC present on the mic jack itself.

If it's a go after that you could do a response and see what the noise level looked like, and if OK, it will certainly produce regular graphs just like a line-in.

But why bother - just get a cheap soundcard.... it's less than the price of a pizza.

brucek

Here is the working SD-12 graph, my DSP-1100 will be here next week!

That's a very good response.

You'll have no problem equalizing that with the BFD.

brucek

Are the predictions made in REW with the filters fairly accurate to what happens in real life? Sadly I am the kind of person who will spend hours playing with Q's to get a response that sounds exactly like the automatic one.

The results that show in the 'corrected' graph in REW are amazingly close to what happens after you enter those filters into the BFD and remeasure.

I find the automatic filters don't get the 'corrected' plot as close as I'd like it, so I always tweak them while watching the 'corrected' plot until I'm satisfied and ready to load them into the BFD.

If you're a tweaker, then you're doomed for sure, because REW is a tweakers paradise (or sentence).

But seriously, the automatic suggestions are fairly close and if you entered them, they'd be fine...

Note that REW adds no gain to its filters and if you do that manually the 'corrected' plot is a crap shoot...

brucek