Graphs from REW

[VonMagnum]

More graphs. I hope I made relevant ones. First up, Waterfall decay plots to 200Hz up to 450ms, without boost and below that with bass boost and a full bandwidth Spectrogram (bass boost one didn't look much different). I've also included a RT60 block plot. It looks like the decay range is a bit higher than recommended (300ms) at mostly 350ms, but not terrible (probably due to the active back wall; if I put up a tapestry on that wall I bet it'd improve somewhat). I don't know about that 60-70Hz bit that's longer?

 

[VonMagnum]

@PENG - Here's a 1/24 Octave smoothed Frequency Response graph.

 

[VonMagnum]

I upgraded my old Macbook Pro with a 512GB SSD and took it and my Presonus (for maximum accuracy) along with my UMIK-1 on a trip downstairs last night. I took 2ch and 6ch matrixed reference measurements for stereo mode from all six seats for both left and right channels. I also took reference, flat and off measurements for the mlp in both true 2ch (plus sub) and 6ch matrixed stereo (uses wides and front heights with active mixing, the heights to raise the soundstage for a dialog lift effect.

Finally, I also took stereo measurements (as in both channels at the same time) to see how they integrated at the MLP.

The readings weren't as good as the upstairs (might need some bass traps). The MLP wasn't terrible (closest to the natural upstairs results) , but it was pretty obvious that Audyssey isn't quite what it's cracked up to be and obvious why many prefer to limit it to the Schroeder region. It's biggest benefits were in the bass region. Moving the mic even a few inches could alter the response significantly at higher frequencies so I don't know how much stock I'd put into those readings anyway.

Otherwise, it pretty much confirmed what my ears heard in terms of the front left and second row center sounding pretty good (not too far off the map response). The front right sounds like mega bass and the graph confirms it's got quite a rise there (OTOH some like that seat best, probably because they dig extra bass).

The back seat was surprising. I thought I needed a second sub to even things out, but the graph suggests the strongest relative bass curve is back there and the overall response curve besides that isn't anywhere near as bad as I suspected, but the volume level other than bass is down 8dB by then, making it just sound weak. It's down 4dB in the second row which is more manageable (dialog increase or lower bass I have cranked anyway and just raise master volume a few dB, which also helps the back, but it's still off).

I'm not sure why the volume falls off so rapidly in just 20 feet, but I suspect the narrow room is doing some reflecting/canceling plus it would probably help to put the chairs on mini-risers as well. Clearly, one or even two rows (especially closer spaced as my second row is back more from the first than the third) is easier to get more consistent / even sound than three, but then that chair isn't used much (screen looks much smaller as well).

It's fine for football games, etc. And that's just 2-channel playback. It's closer to surround speakers and thus they're louder than the front. I don't know how you could even such a setup out. Speakers would need to be further away from all seats. If you could pass mono directional dialog, etc through side surrounds at a lower volume to increase volume for rows further back that would probably help, but you'd no longer be discrete. I've never heard of a home system capable of such a thing anyway.

I can post some graphs when I get home, but there are so many, I'd have to be pretty selective.. The waterfall/spectrographs show the room modes. The rest just show what I described.

 

[VonMagnum]

Left / Right Channel MLP 2-channel Stereo PSB T-45 Speakers in my home theater with the Def Tech PF-1500 Sub set to 80Hz and T-45 set to 'small'. Taken from MLP (front center seat).

Left Channel = Green ; Right Channel = Blue
Set to "FLAT" Response in Audyssey. 1/6 Scaling

Waterfall Plot PF-1500 Sub Crossed to PSB T-45 Speaker at 80Hz (Right Channel)

 

[VonMagnum]

This is the Rear Seat (3rd Row) Frequency Response Graph (Right Channel). As you can see, the bass is sky high from the sub, but then falls off to not great, but not horrible frequency response, but at a much lower level and the highs fall off above 10K as well (given the seats aren't elevated I'm not shocked).

With the dialog lift function enabled (active mixer effect) and the matrixed front wides on, the response in the back is considerably improved, at least for higher frequencies (I normally have these engaged at all times anyway)

As you can see the room modes start showing up a lot more in the rear of the room. This waterfall is from the rear seat (3rd row) with the right channel and sub going.

 

[William Lemmerhirt]

Wow, got a couple of pretty wide troughs at the MLP. 65-150hz looks like a doozy. I gonna try and look closer tonight if I get a chance. Can you set the vertical limits to be 45-105? Also, could you post a graph from 10-200 with the height at 45-105?

My last audyssey run by itself wasn’t beautiful either. I’m super glad I have a minidsp!!!

 

[VonMagnum]

Well, that was "true stereo" mode (2-channel). I don't actually use that and it wasn't actually calibrated for that so that may have something to do with it. The 6-channel matrix mode is what I calibrated it with. It looks considerably better (forgot to post it). It's actually +/- 2.5dB over most of the range save that dip at 54Hz and 658Hz (dip isn't like that at the top in "flat" mode).

Here's the graphs with the calibrated 6ch matrix stereo mode and the Subwoofer response at the settings you asked for (Reference Calibration I normally use; ultra high-end is lower, but it sounds better to my ears). Still one somewhat nasty trough at around 54Hz, but the really nasty ones at 80-150Hz are gone in this mode at the MLP. The microphone might be in a slightly different position as well (ears aren't actually in the middle, after all).

And I still think that "ancient" Def Tech PF-1500 sub does pretty damn well for a 1995 subwoofer. (and that's set to the "Flat" setting; it's normally +8dB above that (Flat at 22Hz with the cursor, down 5dB at 20Hz, which makes it +3dB at my normal bass setting). I'm also impressed with my Rat Shack SPL meter. It was right all along with the slow frequency sweeps that match what I see here exactly.

 

[VonMagnum]

Here's a graph showing 2-channel OFF (no correction) versus 6-channel OFF (no correction) versus 2-channel FLAT and 6-channel FLAT. You can see where the differences occur in the frequency response (slightly messy, but a lot of overlap in some regions). Again DefTech PF-1500 Sub + PSB T-45 Speaker (Right Channel Only)

GREEN = 2-channel Stereo CORRECTION = OFF
ORANGE = 6-channel Stereo CORRECTION = OFF

Purple = 2-channel Stereo CORRECTION = FLAT
BLUE = 6-channel Stereo CORRECTION = FLAT

And below that the second graph shows just 10-200Hz with all four again

6-channel Matrixed Stereo in "FLAT" produces by far the best overall curve (BLUE LINE)

 

[PENG]

@PENG - Here's a 1/24 Octave smoothed Frequency Response graph.

View attachment 29528

It looks good between 25 to 90 Hz, but I bet RC such as Dirac Live, or even plain Audyssey XT32 can dial your FL, FR, and Sub together much better between 25 to 130 Hz at least, if not higher.

 

[VonMagnum]

It looks good between 25 to 90 Hz, but I bet RC such as Dirac Live, or even plain Audyssey XT32 can dial your FL, FR, and Sub together much better between 25 to 130 Hz at least, if not higher.

Perhaps, but would I hear the difference rather than just see it on the graph? For untreated response, that room is rather good, I think. I'll probably eventually get a MiniDSP to play with, though just to see what it can do.

Meanwhile, downstairs in the home theater room, I thought about taking some large oversized pillows down to the home theater room with the laptop and see if they had any effect at all on that room mode at 54Hz (to see if it was worth it to bother with ugly bass traps). There's 4 corners I can try there, two where it would particularly ugly unless it blended in and was easy to move out of the way (so I can get into the cabinets) and not an option higher up on the wall where the bookshelves are. The other two are the right back corner (where my back lamp is, but only up part of the way or it'd block my displays and half bathroom left corner again, limited by the display posters as a partial. It's not a function only studio, so it has to look acceptable.

 

[PENG]

Perhaps, but would I hear the difference rather than just see it on the graph? For untreated response, that room is rather good, I think. I'll probably eventually get a MiniDSP to play with, though just to see what it can do.

Meanwhile, downstairs in the home theater room, I thought about taking some large oversized pillows down to the home theater room with the laptop and see if they had any effect at all on that room mode at 54Hz (to see if it was worth it to bother with ugly bass traps). There's 4 corners I can try there, two where it would particularly ugly unless it blended in and was easy to move out of the way (so I can get into the cabinets) and not an option higher up on the wall where the bookshelves are. The other two are the right back corner (where my back lamp is, but only up part of the way or it'd block my displays and half bathroom left corner again, limited by the display posters as a partial. It's not a function only studio, so it has to look acceptable.

Good point, it is not a sure thing that you would hear a difference let alone a better difference. Below was what I achieved when I had the AV8801. My AVR-X4400H is doing a better job now, especially with the Editor App. You can see that in my room the FL, FR and subs did not play nice at all in the low range so surely anyone could hear a difference in my case, not just on paper.

XT32 did a pretty good job on its own, minidsp+REW did help a little but I didn't feel it's worth with the additional hook up and cables dangling, ymmv..

Not that this is a 45 dB window, and 1/48 smoothing. If I use 1/6 and 60 dB window it would look more like a straight line lol..

 

[VonMagnum]

Good point, it is not a sure thing that you would hear a difference let alone a better difference. Below was what I achieved when I had the AV8801. My AVR-X4400H is doing a better job now, especially with the Editor App. You can see that in my room the FL, FR and subs did not play nice at all in the low range so surely anyone could hear a difference in my case, not just on paper.

XT32 did a pretty good job on its own, minidsp+REW did help a little but I didn't feel it's worth with the additional hook up and cables dangling, ymmv..

Not that this is a 45 dB window, and 1/48 smoothing. If I use 1/6 and 60 dB window it would look more like a straight line lol..

View attachment 29632

That looks pretty good, but keep in mind I don't use the 'green' graph on the Carvers above. I use the red one so that nasty green dip actually lines up well with the rest of the graph in reality (this is the Pyschoacoustic Smoothing --best guess to what the software thinks I'll actually 'hear' in the room) with just the actual used bass curve shown. Suddenly the dip at 2.7kHz and 10kHz look the worst (I wouldn't trust the 10kHz one at all as moving the mic an inch could change it entirely (and certainly turning down the top trim makes it just smoothly fall flat instead there rather than rise back up again after 10kHz) and I'm not so sure the 2.7kHz one is reliable either as it could vary within the distance to my ears even (taken in the middle). All Dirac or Audyssey would want to do is tame my bass curve when it's a nice smooth sounding descent to 'flat' at 200Hz right now.

Really, from what I've read my worst measurement is the time for the room to die down (450ms isn't desirable for a studio environment; 300ms or less is). I made my rock album in this room and while only vocals and acoustic guitar were recorded with a microphone (everything else was direct in to keep noise levels low, etc.), it sounded pretty damn good (Pink Floyd like sound quality on the album and at a tiny fraction of a studio cost even renting a commercial one would have cost). +/- 5dB is considered good for room freq. response in a studio. They want the other parameters pristine more than freq. response according to a forum I've been reading and getting some of the others in line is much harder (all room treatments, not room correction that can bring ringing down one iota, etc.)

 

[PENG]

That looks pretty good, but keep in mind I don't use the 'green' graph on the Carvers above. I use the red one so that nasty green dip actually lines up well with the rest of the graph in reality (this is the Pyschoacoustic Smoothing --best guess to what the software thinks I'll actually 'hear' in the room) with just the actual used bass curve shown. Suddenly the dip at 2.7kHz and 10kHz look the worst (I wouldn't trust the 10kHz one at all as moving the mic an inch could change it entirely (and certainly turning down the top trim makes it just smoothly fall flat instead there rather than rise back up again after 10kHz) and I'm not so sure the 2.7kHz one is reliable either as it could vary within the distance to my ears even (taken in the middle). All Dirac or Audyssey would want to do is tame my bass curve when it's a nice smooth sounding descent to 'flat' at 200Hz right now.

Really, from what I've read my worst measurement is the time for the room to die down (450ms isn't desirable for a studio environment; 300ms or less is). I made my rock album in this room and while only vocals and acoustic guitar were recorded with a microphone (everything else was direct in to keep noise levels low, etc.), it sounded pretty damn good (Pink Floyd like sound quality on the album and at a tiny fraction of a studio cost even renting a commercial one would have cost). +/- 5dB is considered good for room freq. response in a studio. They want the other parameters pristine more than freq. response according to a forum I've been reading and getting some of the others in line is much harder (all room treatments, not room correction that can bring ringing down one iota, etc.)


View attachment 29633

I thought you dialed things in manually and that's very good results in the 25-90 Hz region. Above Schroeder, we both know the results are not too reliable, manual, Dirac, or whatever, probably are not reliable regardless, but one can try it, and go by ears.