Phase characteristics of different software? (speakers only)
Jul 9, 2004 at 5:47 AM Thread Starter Post #1 of 5

ooheadsoo

Headphoneus Supremus
Joined
Dec 11, 2002
Posts
4,835
Likes
13
Anyone play with theirs just to figure it out? I'm always running test tones on my system to try to get the best integration between my subwoofer and my mains. For the longest time, I used WinISD's signal generator. I would run the signal down and try to match phase and amplitude.

A few weeks ago, I found a program named nti_minirator.exe that also has a signal generator. I tried to use it to match my sub and mains. It sounded slightly different for some reason.

When I got my emu1212m, I noticed that patchmix had a signal generator too so I tried to match on that. It was hard to use though at low frequencies, where you typically have the most problems in a room, so I didn't pay much attention to it.

Then I remembered that foobar has its own signal generator. I plugged in a bunch of values and was horrified. In foobar, my subwoofer was CLEARLY out of phase with my mains. I needed a big 45 degree shift (how you would do this without variable phase shift is *shudder*.) I figure I better follow foobar's results since I use it to play all my music, eh? I wonder why the results were SO FAR OFF of WinISD's signal generator? On a side note, I figured by ear that my room related midbass peak is about 8db tall.

Anyone else have any experience with this? I'm still reeling from how big the difference between winisd and foobar is. The good thing is that I feel I've improved the integration between my sub and mains...again
wink.gif


I can't wait til someone makes a better EQ for foobar now. I hope there's a lot more than 31 bands too. I could use a lot of definition at the bottom end.
 
Jul 9, 2004 at 6:17 AM Post #2 of 5
Are you just generating pure, constant tones (as opposed to frequency sweeps)?

I doubt it has anything to do with the phase behavior of the programs. The phase of the signal you're sending to your speaker rig is irrelevant if it's a pure tone and you're not doing the crossover in software. Rather, I suspect Foobar's tone generator may be introducing some distortion. You may want to try checking the "Oversample 32x" checkbox in Preferences -> Playback -> Input -> Standard inputs. Not sure that will fix it though -- Foobar's tone generator may be squirrely.
 
Jul 9, 2004 at 7:48 AM Post #3 of 5
I'm generating pure tones and running through them at 10 hz intervals. 32x oversampling doesn't change anything.

I just realized that the signal generator only generates a left channel signal. I thought something was funky with it before but I thought it was a phase cancellation issue. When the sub is going at low frequencies, it kinda emanates from most everywhere, it doesn't appear to sound like it's coming from a pin point location, so I suspected, but didn't make sure what was going on. Headphones confirmed that 100% real quick
biggrin.gif
Is that the reason? Should I not match the subwoofer to only the left channel? I think it should be ok, technically. The tone generated by foobar sounds "cleaner" than winisd's if that means anything.

Edit: actually, I now remember that I had noticed with winisd that a 45 degree shift in phase was better at some lower frequencies, but 100hz started to get some phase cancellation when I started shifting it. (I noticed this when I sealed my sub with the intention of using it that way.) It was also really hard to hear with WinISD. I felt like it didn't make that big a difference whether I shifted it or notWith foobar, everything was kinda fuzzy. At the time, I felt that the phase shift didn't help that much and I wanted the smoothest integration with my sub, so a dip at 100hz was out of the question. As a result I compromised by shifting it as little as possible and achieved, through what I heard in winisd as a good compromise. Running through a bunch of tones in winisd, it sounded pretty smooth going down. However, this same examination in foobar with only the left channel going revealed a huge phase cancellation, pretty much necessitating the 45 degree shift. The change caused by the phase shift when listening using foobar was huge.

I think for now, the best thing may be to up the level of my sub at 100hz just a bit, which makes the ~70-85 hz frequencies a bit too strong, and then bring that down via foobar's eq. That's why I'm eagerly waiting for an eq with better precision. Believe it or not, I want to pull down 40hz just a notch. What I've done so far with what I have sounds good to my ears. Of course, I've said this just about every time I adjusted my sub
wink.gif
 
Jul 9, 2004 at 8:07 AM Post #4 of 5
Yeah, that's one of the oddities of Foobar -- you have to use the Convert Mono to Stereo DSP if you want two channel tones out of the tone generator.

What you're noticing is in-room wave effects. These are phase, frequency, and room position dependent. Think of how waves in a pond cancel each other or reinforce each other depending on how they interact. When you're running both main speakers, the effects are more complex than with one speaker, and you'll notice the different effects that you're seeing.

You actually have two separate goals in optimizing your system. You should first strive to get the subwoofer properly matched with both the left and right speakers played individually (mono). Once you do that, you have the functional equivalent of two "perfect" floorstanding speakers capable of flat, deep bass. Your next goal is to position these perfectly matched speakers so that room effects result in balanced bass at your listening position. No matter what you do (this is unavoidable in any speaker system), there will be spots in your room where the bass is too heavy and spots where the bass is too light. These effects are also frequency-dependent so it can be aggravating to optimize. The goal is to have nice, balanced bass coming from both speakers at your listening position. While striving to meet this goal, if you move the speakers too much, you may wreck the individual subwoofer/speaker matching (i.e. the prior goal). If this happens, go back to step one. The best speaker positioning will be a compromise between all these factors.
 
Jul 9, 2004 at 8:15 AM Post #5 of 5
Yeah, I've tried the moving thing. It's extremely sensitive. Whenever I try new placement with my speakers, I have to turn off the sub because it sounds astoundingly awful out of phase. And yes, I know I was hearing room modes as far as the huge dips and peaks I get, though I'm fairly sure the peak I'm getting at 70-85 hz is actually a peak in the response of the woofer in conjunction with the 4th order variable crossover built into the amp. The huge peak at 120-170hz is definitely room related. My left speaker sits on top of my sub. I'm actually planning on treating my room next. Gonna buy, hopefully, 8lb./sqft fiberglass and panel up my front and left wall. If you've seen the pic of my system, you'll know exactly what I mean
tongue.gif
Putting the 8lb/sqft fiberglass in the corners of my room should act as a semi decent midbass trap down to 100hz, if the specs the manufacturer are giving out are accurate.

Btw, another quirk with different software: with nti_minirator.exe, I can't hear any tones above 16khz. With patchmix's signal generator I can clearly hear it but there's warbling. With foobar, I can clearly hear them and cleanly. WinISD doesn't go above 1khz. *sigh* Say goodbye to winisd except when I'm trying to piece together a new sub and nti_minirator is gone for good
rolleyes.gif
 

Users who are viewing this thread

Back
Top