Audyssey Room EQ Review

[Robbo99999]

Even what might seem like small changes can have huge impact of what you hear. Time-aligning the distances and volume will for sure have a positive effect on imaging if that is off. Theoretically this would not be needed in a symmetrical, treated room with identical response from each speaker etc.. But no one has a perfect room with perfect placement so some correction is usually needed.

I've found that this is a good starting point for very good imaging if the room and speakers allow for anyone that is interested to play around with placement.

I haven't worked out the maths on your diagram, is that a 30 degree angle from listener position to each speaker, but figured I may as well ask to clarify for everyone anyway? (And what's the "grader" thing mean in brackets?)

 

[3ll3d00d]

Nice review @amirm The number one issue I have with (all) H/W solutions like this is the limited number of FIR filter taps to actually provide real low frequency room correction. Whether it is Audyssey, miniDSP, Trinnov, DEQX, or other H/W processors they all suffer from the same issue - not enough FIR filter taps below 100 Hz to really be called room correction.

As a maths example, the miniDSP 2x4 HD product datasheet talks about 4096 taps. But this is the total number. For 4 channels you have 1024 taps each channel available. The frequency resolution of a 1024 taps filter @ 48 kHz samplerate is 48000/1024 = 46.875 Hz. So e.g. below 100 Hz there are just 2 frequency bins at 46.875 and 93.75 Hz. This clearly means that you have no control over the lower frequency range.

Audyssey has issues (default target curves, mid range dip, overcorrection at high frequencies in some cases, bass/mains alignment) but adequate resolution for a decent job at modal correction has never been one of them (as far as I recall anyway). They use as multi rate filter to improve resolution IIRC, a quick Google finds the patent for this https://patents.google.com/patent/US8539012B2/

 

[audioBliss]

I haven't worked out the maths on your diagram, is that a 30 degree angle from listener position to each speaker, but figured I may as well ask to clarify for everyone anyway? (And what's the "grader" thing mean in brackets?)

That math was figured out by someone way smarter than me. I should have clarified somewhat- “grader” is Swedish for degrees. So speakers at about +-23 degrees listening window, not to be confused with toe-in.

So let's say you are sitting 4m away from the speakers, that would be y=4
4/x = 1.18
4=1.18x
x= 3.39m

And vice versa of course if solving for y. So this means that if you are sitting 4 meters away from the center point of the line drawn between the tweeters try and have your speakers 3.39 meters apart - again that distance is calculated from the midpoints of the tweeters.

Most of the time it's probably easier to think about it as
listening distance / 1.18 = distance between speakers
distance between speakers x 1.18 = listening distance

A lot of the time you are constrained by the width of the room. Also you'd need to take into consideration individual speaker manufacturer recommendations. Some speakers for instance are built in a way where toe-in of them is sort of built into the speaker. But a rule of thumb and a starting point is to toe them in so you can just see the outside of the speakers from the listening position.

Toe-in is more to broaden the sweet spot a bit. Once you start listening off axis a bit toe in makes quite a bit of difference imo. But it also helps aids with reflection issues. Some would say that their speakers are not sensitive here since their off axis response is the same as on axis and the reflections just help out. I say test it out and see what works best for your speakers and room setup.

Again it's a starting point but I have found it useful and I thought I'd share it. Typically I see speakers way too close apart with a long listener distance with very minimal toe-in. But speakers are different and most of the time the room only allows for so much flexibility.

This sort of thing can be a controversial topic and there are no hard rules. It all depends!

 

[Robbo99999]

That math was figured out by someone way smarter than me. I should have clarified somewhat- “grader” is Swedish for degrees. So speakers at about 23 degrees.

So let's say you are sitting 4m away from the speakers, that would be y=4
4/x = 1.18
4=1.18x
x= 3.39m

And vice versa of course if solving for y. So this means that if you are sitting 4 meters away from the center point of the line drawn between the tweeters try and have your speakers 3.39 meters apart - again that distance is calculated from the midpoints of the tweeters.

Most of the time it's probably easier to think about it as
listening distance / 1.18 = distance between speakers
distance between speakers x 1.18 = listening distance

A lot of the time you are constrained by the width of the room. Also you'd need to take into consideration individual speaker manufacturer recommendations. Some speakers for instance are built in a way where toe-in of them is sort of built into the speaker. But a rule of thumb and a starting point is to toe them in so you can just see the outside of the speakers from the listening position.

Toe-in is more to broaden the sweet spot a bit. Once you start listening off axis a bit toe in makes quite a bit of difference imo. But it also helps aids with reflection issues. Some would say that their speakers are not sensitive here since their off axis response is the same as on axis and the reflections just help out. I say test it out and see what works best for your speakers and room setup.

Again it's a starting point but I have found it useful and I thought I'd share it. Typically I see speakers way too close apart with a long listener distance with very minimal toe-in. But speakers are different and most of the time the room only allows for so much flexibility.

This sort of thing can be a controversial topic and there are no hard rules. It all depends!

Ah, interesting. Yeah, so that's not the often quoted optimal 30 degrees then, you got it at 23 degrees...but yes if you have a narrow room then what can you do! I've tried shallower angles than 30 degrees with these speakers (e.g. less than 30), and the imaging can still be there, but it still improves all the way up to 30 degrees, I haven't tried wider than 30 degrees because the way I have my room configured at the moment then this is not possible. Regarding toe in, I've not played with my setup that much in terms of toe in, but I intuitively toed the speakers in so that I'm generally just seeing the flat front face of the speaker (so directly facing me)...well actually I can see the inside wall of each speaker slightly, so they're pointing only very slightly behind my head, but pretty much directly face on if you don't wanna nit pick. (My listening position is 2m away from the speakers).

(JBL recommend 30 degrees as listening position, seems to tally with my experience).
EDIT: Ah, here's my manual for my JBL 308p Mkii's, it does indeed look like the speakers are toed in so that you can see the outside wall like you mention, although I don't know if they've purposefully drawn the diagram like that or whether it was unintentional, as they don't talk about the amount of toe in. Intuitively to me it makes no sense to toe them in so you can see the outside speaker wall because that suggests the sweet spot is in front of you and that the left & right speakers "have crossed over in front of you".

 

[audioBliss]

Ah, interesting. Yeah, so that's not the often quoted optimal 30 degrees then, you got it at 23 degrees...but yes if you have a narrow room then what can you do! I've tried shallower angles than 30 degrees with these speakers (e.g. less than 30), and the imaging can still be there, but it still improves all the way up to 30 degrees, I haven't tried wider than 30 degrees because the way I have my room configured at the moment then this is not possible. Regarding toe in, I've not played with my setup that much in terms of toe in, but I intuitively toed the speakers in so that I'm generally just seeing the flat front face of the speaker (so directly facing me)...well actually I can see the inside wall of each speaker slightly, so they're pointing only very slightly behind my head, but pretty much directly face on if you don't wanna nit pick. (My listening position is 2m away from the speakers).

(JBL recommend 30 degrees as listening position, seems to tally with my experience).

There are quite a few different recommendations out there THX, Dolby and ITU have their own and I'm sure people like Floyd Toole has his. For the front speakers it's usually 45-60 degrees as you see in many diagrams or another way of saying that is +- 22.5-30 degrees.

If your speaker manufacturer, in this case JBL, recommends a certain angle I'm sure it's good to start there like you have.

If speakers are placed too narrow together the image will be very small and not expanded, if placed too far apart the image will collapse. I'm sure different angular separation between different setups can be needed. I think the important thing is to actually think about this is the first place that it's a variable to consider. A lot of the time speakers and seating position is placed where it is convenient.

If one has a narrow room it can be worth trying to put the speakers along the long wall. At least in this case the speakers will be very far from the side walls and you can get really crazy wide imaging(if source material allows), but limited in other ways(usually bass response becomes quite difficult in this kind of setup in my experience).

 

[Robbo99999]

Yep, I can see all that.

 

[RayDunzl]

I was under the impression that dirac at least did a combination of IIR and FIR, is this tap limitation possibly one of the reasons why?

I have a miniDSP OpenDRC-DI, with 6144 taps.

It uses IIR to adjust low frequencies and broad swaths.

It uses FIR for details and general phase correction, but there aren't enough taps for detailed change at low frequencies.

Using AcourateDRC, with a combo of OOR and FIR filters, it creates this full-range correction (as a handy example)

The top traces represent the change in frequency response for the FIR filters, add to that the effect of the bottom traces, the IIR filter response:

The phase adjustment is not graphed by the AcourateDRC GUI.

---

Using the option to create an FIR filter only, you can see some difference in the detail of the correction in the bass:

 

[RayDunzl]

And, speaking a little above my Pay Scale:

The IIR filters are somehow created with "biquads"

Here is one of fourteen the OpenDRC-DI will accept:

biquad1,
b0=0.46504793292111776,
b1=-0.48663553600708934,
b2=0.12731300430290824,
a1=0.48208264189610995,
a2=0.40308616866499475

So, that creates one or more of the curves in the IIR graph above (somehow)

The FIR filter data is 6144 bytes that represent an Impulse Response with which the samples passing through the filter are modified.

The first few FIR samples, of 6144 per channel:


b0 = 0.0,
b1 = -7.447376050072701e-21,
b2 = 2.933111381813597e-13,
b3 = 1.7735010508779236e-12,
b4 = 5.957014829333085e-12,
b5 = 1.5005633888232062e-11,
b6 = 3.174875640166164e-11,
b7 = 5.970551136824653e-11,
b8 = 1.0310663434154321e-10,
b9 = 1.66916203081513e-10,
b10 = 2.5685306659362084e-10,

all tiny numbers at the fringes.

The interesting part is near the center of the list (if centering was chosen)

Graphing the whole list of values:

And zooming in on the middle:

My (perhaps flawed) understanding is it takes in 6144 music samples, multiplies each sample by the corresponding value in the 6144 values in the filter, and outputs one sample.

Then, the oldest sample is dropped, the next (new) one added, and all those samples are multiplied by the filter coefficients, to come up with the next sample to be output.

 

[capt.s]

That's QSound at work.

I didn't know that was a thing so I Googled it. I've long had Madonna's 'Immaculate Collection' in my Test Disc list because of the incredible imaging - now I know why! Unfortunately, seeing as this was the early 90's - it apparently didn't catch on - shame!
BTW - how is that 30 years ago???????????

 

[da Choge]

Amir - Just picking-up this thread now -- those look like the Mark Levinson No53 monaural amps that you have in your system - didn't notice those in previous photos - Very Nice !!! How long have you had them? I always thought the 53s, when they came out, were really visionary and ahead of their time. My dream preamp was always the No52 (not that I could ever afford it and anywhose now it's out-of-production), but it's such a beauty and I imagine specs very well.

 

[tsanguine]

Cheers @amirm. With respect to audio/video latency, there are a couple of options to consider:

If using JRiver for video playback, the convolution engine takes into account the linear phase FIR filter latency, so you get perfect lipsync.

If using the Win10 Netflix app that supports Dolby Digital, Atmos, etc. One can route the audio through Hifi cable & ASIO bridge and on the output, select Audiolense Convolver which then hosts a minimum phase version of the correction filters. @Juicehifi 's convolver is very low latency, so I do not notice any lipsync issues.

If using a Media player that supports DirectShow (like Windows Media Player or VLC for example), one can install audio/video codecs (scroll down to see features) and using the 2nd option above, pretty much decodes everything...

Last I looked JRiver couldn't process Atmos though so you had to bitstream to the receiver and therefor lose all of the audio processing of JRiver. Has something changed or is this a solution for everything but Atmos or patiently waiting for a PC solution to decode the Atmos stream?

 

[amirm]

Amir - Just picking-up this thread now -- those look like the Mark Levinson No53 monaural amps that you have in your system - didn't notice those in previous photos - Very Nice !!! How long have you had them? I always thought the 53s, when they came out, were really visionary and ahead of their time. My dream preamp was always the No52 (not that I could ever afford it and anywhose now it's out-of-production), but it's such a beauty and I imagine specs very well.

I have had them for probably 5 years. I know I had them before starting this forum.

 

[amirm]

Cheers @amirm. With respect to audio/video latency, there are a couple of options to consider:

If using JRiver for video playback, the convolution engine takes into account the linear phase FIR filter latency, so you get perfect lipsync.

If using the Win10 Netflix app that supports Dolby Digital, Atmos, etc. One can route the audio through Hifi cable & ASIO bridge and on the output, select Audiolense Convolver which then hosts a minimum phase version of the correction filters. @Juicehifi 's convolver is very low latency, so I do not notice any lipsync issues.

If using a Media player that supports DirectShow (like Windows Media Player or VLC for example), one can install audio/video codecs (scroll down to see features) and using the 2nd option above, pretty much decodes everything...

I was talking about AVRs implementing long filters. Since their main use case is a disc player or cable box as inputs, they have to be careful with latency. They get uncompressed video on input which makes buffering requirement much harder than a software player which can buffer compressed frames.

 

[da Choge]

I have had them for probably 5 years

Kudos !!! - my bad; NOT been paying close (enough) attention to your setup photos; but anyways -- when are you going to post your tests to the forum? My guess is that you've prolly already tested them and wouldn't have too much to be embarrassed about. I would hope that these should prove to be primo amps. Apologies for this OT post.

 

[Caligari]

XT32 runs at 48 kHz. Anything above that is downsampled.

 

[mansr]

XT32 runs at 48 kHz. Anything above that is downsampled.

With quite a bit of aliasing, I might add.

 

[Caligari]

With quite a bit of aliasing, I might add.

Yes.

 

[capt.s]

With quite a bit of aliasing, I might add.

So, is 44.1 up-converted to 48 when XT32 is enabled?

 

[jam]

XT32 runs at 48 kHz. Anything above that is downsampled.

That's not imposed by a design limitation of the Audyssey MultEQ XT32 algorithm. Rather, it's as a result of the limited processing power of the DSPs used in many of the AVRs/AVPs. Therefore, they downsample the higher sampling rate signals such as 96kHz or 192kHz down to 48kHz.

 

[jam]

So, is 44.1 up-converted to 48 when XT32 is enabled?

No.