Help me out here... what shows I have FDW active and where do I find that? Is that in a REW window?
EDIT: I think it's the analysis window but I left at the default which I didn't think had it active but I can check. And regardless, IIRC, I should have it set to about 6 cycles?
Last edited:
Just to be clear, in JRiver you choose the driver for your preferred audio output. I'd prefer an Asio driver or Wasapi over direct sound.
In REW you set it to JRiver. If by any chance this does not work properly it could be some JRiver settings need tweaking. While REW just works for me, I've ran into trouble using APL_TDA this way. After I changed a setting within JRiver all was well. I think it was: Play silence at startup for hardware synchronisation that I had to set to none. It was a looong time ago
.
Maybe the fact that it says: 5: Dec 1 23:09:13 [FDW] {1/48} in the legend?
Click on the IR Windows button on the top of the REW screen and see if there's a [v] in front of Add frequency dependent window and notice the number of cycles or width in octaves radio button.
Sometimes its better to cheat .
15 cycles on a sound stream means you're looking at 15 x the time it takes for one wave at a certain frequency to develop. An FDW is not like a gate set at a certain time frame for all frequencies. It is more like a sliding window, short at high frequencies and getting longer and longer for each lower frequency.
If you look at 6 cycles, the result will be smoother (that's not the goal why we pick 6 cycles) but it will hopefully be done before any mayor reflections happen (that is the goal of looking at a lower number of cycles). Mayor reflections from vertical line arrays happen due to parallel vertical planes or ridges. You have the arrays tucked away in the corners. Each vertical plane or screen or ridge along those walls which are parallel to the arrays (which in turn makes the distance from the drivers to that plane or ridge quite identical) is going to show up somehow.
For instance, if there is a door frame sticking out about a meter away from the corner you'll see it in the measurement. Or a screen hanging in front of the front wall. You're basically sitting in two giant waveguides. Together with all the stuff in that room. Small stuff compared to the array height gets filtered out. Parallel stuff to the arrays will show up in your measurements if it's tall enough.
That's the secret why it only took me 3 damping panels, placed just right, to absorb most of my offending reflections. It's also the reason why my arrays have a smooth outer shape, all rounded to avoid similarity in distances. Even in your corner, there is a ridge from baffle to wall. Round it off and it probably does better (or use absorption material in that corner). If it's early enough, DRC will try and make up for it. That's why the sliding window in DRC is important to get the best results. Also the reason why no standard template did what I wanted it to do.
Arrays do wonders to filter most stuff out, as long as each driver has a different distance to that object, from driver to that object to your ear (or your microphone). The more drivers and differences in distances, the better it smooth's out the reflection of that object. It gets averaged out.
What does the IR look like, when we zoom in on the first 10 to 15 ms? Preferably from a single channel. You may post both to see what's happening.
.15 cycles on a sound stream means you're looking at 15 x the time it takes for one wave at a certain frequency to develop. An FDW is not like a gate set at a certain time frame for all frequencies. It is more like a sliding window, short at high frequencies and getting longer and longer for each lower frequency.
If you look at 6 cycles, the result will be smoother (that's not the goal why we pick 6 cycles) but it will hopefully be done before any mayor reflections happen (that is the goal of looking at a lower number of cycles). Mayor reflections from vertical line arrays happen due to parallel vertical planes or ridges. You have the arrays tucked away in the corners. Each vertical plane or screen or ridge along those walls which are parallel to the arrays (which in turn makes the distance from the drivers to that plane or ridge quite identical) is going to show up somehow.
For instance, if there is a door frame sticking out about a meter away from the corner you'll see it in the measurement. Or a screen hanging in front of the front wall. You're basically sitting in two giant waveguides. Together with all the stuff in that room. Small stuff compared to the array height gets filtered out. Parallel stuff to the arrays will show up in your measurements if it's tall enough.
That's the secret why it only took me 3 damping panels, placed just right, to absorb most of my offending reflections. It's also the reason why my arrays have a smooth outer shape, all rounded to avoid similarity in distances. Even in your corner, there is a ridge from baffle to wall. Round it off and it probably does better (or use absorption material in that corner). If it's early enough, DRC will try and make up for it. That's why the sliding window in DRC is important to get the best results. Also the reason why no standard template did what I wanted it to do.
Arrays do wonders to filter most stuff out, as long as each driver has a different distance to that object, from driver to that object to your ear (or your microphone). The more drivers and differences in distances, the better it smooth's out the reflection of that object. It gets averaged out.
What does the IR look like, when we zoom in on the first 10 to 15 ms? Preferably from a single channel. You may post both to see what's happening.
Last edited:
Ok. That furthers my understanding of FDW. Very helpful, thanks! I may have some issues from the edges of my projector screen along with the walls. The screen edge protrudes from the wall about 1.5" and is about 10" from the center of the drivers. Likely similar to RA7's issues, but there's not much I can do about that without wife trouble.
I'm tempted to just temporarily throw all the rock wool I have ready for the panels against the first reflection points just to get the measurements done even though the panels themselves and the photo canvas that will cover them aren't ready yet. Or I guess I can just sit back and enjoy where I'm at for a while and measure when they're done. Nah!
It is 1AM now so my neighbors probably wouldn't be happy about sine sweeps at the moment. But tomorrow is another day! I'll do more sweeps to show the IR as you say and also do another iteration of the PEQ (both for JRiver and my XMC-1) so I can rerun DRC.
Speaking of bothering my neighbors at 1AM, I was listening to some bass heavy stuff just to see how it sounded. Pretty clean and not boomy at all. At least not compared to what you would expect of the source material. I did notice that the walls do vibrate a little. It was fairly loud and I would expect some movement of course. But I'm not sure if it is detrimental to the sound very much. I probably won't do anything about it. Especially since I don't make it a habit of listening to a lot of dubstep or whatever. But thought I would share the observation anyway.
Finally, one question before I call it a night. I know you can sync audio/video in JRiver. Is it possible to do the same with a video source from a browser (ie Netflix)? I tried making an adjustment in the regular way in the video options window tab, but it didn't seem to make a difference. Maybe it only works for video played from JRiver? And I don't suppose you can play Netflix from with JRiver?
I'm tempted to just temporarily throw all the rock wool I have ready for the panels against the first reflection points just to get the measurements done even though the panels themselves and the photo canvas that will cover them aren't ready yet. Or I guess I can just sit back and enjoy where I'm at for a while and measure when they're done. Nah!
It is 1AM now so my neighbors probably wouldn't be happy about sine sweeps at the moment. But tomorrow is another day! I'll do more sweeps to show the IR as you say and also do another iteration of the PEQ (both for JRiver and my XMC-1) so I can rerun DRC.
Speaking of bothering my neighbors at 1AM, I was listening to some bass heavy stuff just to see how it sounded. Pretty clean and not boomy at all. At least not compared to what you would expect of the source material. I did notice that the walls do vibrate a little. It was fairly loud and I would expect some movement of course. But I'm not sure if it is detrimental to the sound very much. I probably won't do anything about it. Especially since I don't make it a habit of listening to a lot of dubstep or whatever. But thought I would share the observation anyway.
Finally, one question before I call it a night. I know you can sync audio/video in JRiver. Is it possible to do the same with a video source from a browser (ie Netflix)? I tried making an adjustment in the regular way in the video options window tab, but it didn't seem to make a difference. Maybe it only works for video played from JRiver? And I don't suppose you can play Netflix from with JRiver?
I have found a way to edit sound/video sync in Netflix once, long ago. Let me look for it as it has been ages since I last played anything from Netflix.
In other words, it can be done. I had to set it within the video and had to redo it each time I played a different video. It would help to know the exact delay, which you could check with JRiver and a timing video. I'll see if I can find a good way to share the one I use.
See this video at about 1:32 into the video. Within the video hit "Shift + Alt" (if you're using Windows) and choose A/V Sync Compensation.
That way you can stream Netflix while keeping the sound flow trough JRiver. Just choose JRiver as your Windows preferred sound devise.
(JRiver must be running or open for the sound stream processing)
In other words, it can be done. I had to set it within the video and had to redo it each time I played a different video. It would help to know the exact delay, which you could check with JRiver and a timing video. I'll see if I can find a good way to share the one I use.
See this video at about 1:32 into the video. Within the video hit "Shift + Alt" (if you're using Windows) and choose A/V Sync Compensation.
That way you can stream Netflix while keeping the sound flow trough JRiver. Just choose JRiver as your Windows preferred sound devise.
(JRiver must be running or open for the sound stream processing)
Last edited:
Here's a link to a lip sync video I use: http://www.rsr-concepts.com/diyma/Lip-Sync.mkv
I'd also recommend using the tools in this post on the avsforums to setup and maximise your Home Theatre experience: AVS HD 709 - Blu-ray & MP4 Calibration - AVS Forum | Home Theater Discussions And Reviews
I'd also recommend using the tools in this post on the avsforums to setup and maximise your Home Theatre experience: AVS HD 709 - Blu-ray & MP4 Calibration - AVS Forum | Home Theater Discussions And Reviews
Last edited:
Be aware if you hang framed damping panels you're not hanging them "inside" your wave guide and add another ridge parallel to the array
.Can you use a (soft) frame that absorbs waves hitting the sides of the screen? Some kind of foam?
A little edit:
Apparently it does not work in Chrome. It did work in Internet Exploder
.
Last edited:
One more thing for you to ponder about when you wake up.
Whatever is true in front of the drivers in your array, is also true at the back side of them. If they share the same internal distances which are not damped enough, this too can cause dips in the frequency response.
If you measure the impedance curve, it should be without (seemingly small) wiggles. Each wiggle can cause a dip in the FR curve. Good damping inside the enclosure can prevent this. Materials like wool felt on flat surfaces and fiberglass insulation material to fill the empty space are excellent to absorb most culprits. Knowing is better than wondering about it. I would check the impedance of the array as is. It should be a gradual curve without wiggles, resembling the impedance curve of a single driver.
One known resistor and a few wires/connectors and you can measure impedance within REW.
Whatever is true in front of the drivers in your array, is also true at the back side of them. If they share the same internal distances which are not damped enough, this too can cause dips in the frequency response.
If you measure the impedance curve, it should be without (seemingly small) wiggles. Each wiggle can cause a dip in the FR curve. Good damping inside the enclosure can prevent this. Materials like wool felt on flat surfaces and fiberglass insulation material to fill the empty space are excellent to absorb most culprits. Knowing is better than wondering about it. I would check the impedance of the array as is. It should be a gradual curve without wiggles, resembling the impedance curve of a single driver.
One known resistor and a few wires/connectors and you can measure impedance within REW.
Last edited:
Yes, the impedance test was on my list of things to do. I don't have a sound card to input to though. I think my NUC only has SPDIF out and headphone out. I do have a cheapo USB line in converter that could work but I'm not confident in that unit being able to do a good in this application. Worth a shot?
I just use a very old (windows) laptop for that job, running it from the headphone out to mic input. A known value (non inductive) resistor and a couple of old mini jacks plundered to do the job. I didn't need it to determine T/S parameters, it's only about the smoothness of the plots. I also measured each driver separately to be sure there were no duds.
Hmm, I think my old MacBook Pro has audio inputs through the headphone jack. I can USB output to a DAC and then input through the headphone jack to make this work I think.
I don't quite understand the wiring scheme though. I've looked up a few different images but still not entirely clear. Anyone have a photo of their setup that might be helpful?
I don't quite understand the wiring scheme though. I've looked up a few different images but still not entirely clear. Anyone have a photo of their setup that might be helpful?
I opened my bag of roxul just to see how things might sound. I seem to notice better soundstage at minimum. And clarity if that makes sense.
Took a few sweeps in REW also and it’s obvious the reflections are much better tamed. The main room nodes are still there but not as pronounced. I redid my XMC PEQ with these new measurements while I was at it.
I tried running DRC again with the new EQ in place but levels are too low and I didn’t want to crank it up in the evening after my neighbors are probably already wondering what the previous 10 sweeps were for! Maybe tomorrow if I have time. I probably need to move all the clutter out of the room too. Things are rattling!
Took a few sweeps in REW also and it’s obvious the reflections are much better tamed. The main room nodes are still there but not as pronounced. I redid my XMC PEQ with these new measurements while I was at it.
I tried running DRC again with the new EQ in place but levels are too low and I didn’t want to crank it up in the evening after my neighbors are probably already wondering what the previous 10 sweeps were for! Maybe tomorrow if I have time. I probably need to move all the clutter out of the room too. Things are rattling!
Attachments
Be careful with a DAC and amp in these measurements. I'm using the headphone out as my amp for impedance. It's easy to fry a sound input, which is why I use that old laptop and not my soundcard.
Does this picture help?
See this thread for the calibration (the wires that go to the speaker for the impedance test are shorted (connected to each other) during calibration. See comments from JohnM, he should know
.
For serious measurements I always need to remove some random stuff my girl puts into our room for show
. Things like candle light holders etc.
Had a DATS device for several years and had been practical tool for measuring drivers TS data plus capacitors and inductors up until now, but lately started miss data numbers above 20kHz where its program cuts scale probably because hardware unit is a 44,1kHz USB sound card device, guess REW and ARTA can measure higher in frq if soundcard is supporting rate so will look into that scenario in future to also get data up above 20kHz.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.