*Official Thread* Anthem AVM 70 & 90 and MRX 540, 740 & 1140

[desray]

By happenstance, I took out a bluray copy of a rather crappy movie called Stealth (2005) for a spin after re-watching a 4K re-mastered Top Gun in Dolby Atmos mix. The studio at that time decided to have the audio mixed in LPCM 5.1 instead of the more common DTS-HD MSTR.

Technical details on the audio formats
L-PCM 5.1 @ 48kHz/24bit
DTS-HD Master Audio™ @ 48kHz/24bit

Back then, I have a Denon AVR and using Audyssey and I recalled I wasn’t impressed with the sound mix. I watched it 2 times and put it back into the case and shelf it for nearly a decade until last week. Now with Anthem, I decided to give it another spin and oh my god! It sounded great! The level of details and the bass are powerful enough to shake the room. For those with LPCM 5.1 sound mix or have this bluray version of Stealth, I urge you to re-watch it again but select native mode MPCM 5.1 instead of Dolby Surround Upmixer (DSU). You will be thrilled by what you’ve been missing all these times…

[Ronildoq]

continuing from previously, lets look at the Harmonic Distortion

This is by no means using CEA-2010 burst tones to measure the distortion from the subwoofers. that is used primarily outdoors to determine distortion levels from subs.

This measurement of distortion relates to amplitude response and how much distortion when you crank up the volume on the subs. the below comparison is done based on 90db target levels. even a 1db increase can cause significant increase in distortion numbers as the subs run out of headroom.

let see… below is the 4 dedicated sub approach

see 12hz Total Harmonic Distortion is coming in at 4.36%, thats just 90db levels only… imagine playing it at 115db, that will exceed the THD threshold of 10% easily…

Now look at 2 sub approach below

when you look at 12hz, its coming in at <1% THD

The rest of the freq picked randomly shows approx 1% max and lesser. that is good

So you can see why, blending the subs as 1 is the approach to take, and not let ARC EQ individual subs

[desray]

@Ronildoq

So you can see why, blending the subs as 1 is the approach to take, and not let ARC EQ individual subs

Apart from the premium hardware in the innards, now I wondering whether the marketing hype about 4 independent sub is hyperbole from Anthem. One certainly no need to spend the extra premium to get this feature…Having say that, I can appreciate why it is always better (and most probably the right way) to combine all subwoofers into one single output. Anyway, we should give Anthem some time to optimize this subwoofer and auto phase control. Hope it will early rather than later.

[desray]

So apart from bro @Ronildoq, who else has come into possession of the AVM 90? Pls make know here

[desray]

@Ronildoq

Soundmix engineer’s perspective when mixing below…

I’m using Apple Music hence there is a wide selection of Dolby Atmos music for me for playback but sadly, Dolby Atmos mix in music is a hit or miss (at least for me)…Some good ones are really really good - e.g. Tiesto’s Boom!, the whole album of Bille Eilish Happier than Ever, and Dojo Cat to name a few.

For those who want to play Dolby Atmos music, I recommend to setup a different profile and play around with the Room Gain. You can drop the Deep Bass boost if you want. Remember music and movie in Dolby Atmos requires a different profile, especially so if you are streaming 4K movies on Netflix and streaming Apple Music from one device - i.e. Apple TV 4K.

[Ronildoq]

This is where I find the ARC can improve further. If it blends all 4 subs first, then apply correction. Then it’s good. At the moment, it’s applying correction individually then blend them.

Because some peaks will be attenuated by other subs from other locations. The multi sub approach is different from single sub. A corresponding sub placed elsewhere in the room, works to cancel certain frequencies or sum up with certain frequencies, depending on where you place them. So a lot of eq applied to individual sub is unnecessary . If another sub B placed elsewhere in the room is cancelling the peak of sub A, then there is no need to cut away the peaks of sub A

So ARC’s approach to equalise multi sub is the problem. The approach. The corrections applied is not the issue. It is applying correction filters accurately

[desray]

@Ronildoq

This is where I find the ARC can improve further. If it blends all 4 subs first, then apply correction. Then it’s good. At the moment, it’s applying correction individually then blend them.

Good point.

[Wind]

Trinnov shows the pre calibration, the filters applied, and the post result.

Now based on the first (one and only point) calibration in a typical HDB 5 room flat, you can use that as a reference point to make any changes, this includes, but not limited to phase, amount of sound reflection, filters granularities, amplitudes, the specific frequency region you want to correct and how much db you want to correct etc. Basically like a in built Rew to do all the legwork which satisfy your curiosity within secs.

However Ronlidoq is correct, if you want any form of validation to compare the differences, another software like REW is inevitable. You can also save the changes into different presets and do the validation from software or ears. Think of it like a real time result from any changes you applied which make reference from that very one calibration done.

Hope this helps

[Ronildoq]

ok i did further listening this afternoon, and i found the levels to be extremely loud on the AVM90. a normal track that i would playback at -10mv, i was playing it at -17MV, still the intensity was great. So i was wondering if i had done anything wrong. Turns out the rest of the folks find it to be equally loud as well. i dont recall playing that loud on the denon or lyngdorf

So i took the opportunity to measure the absolute levels, using the atmos test disc below

i have been using this test disc for a number of years now with the SPL meter. the test tones are recorded at -20dbfs.

So i fired up the AVP, at master volume -20, C weighted slow, with all subwoofers turned off, the SPLs for LCR are all reading 81.5db. That is like 16.5db above reference!!! (Technically it should read 65db, ie reference 105db - avp mv 20db - test tones -20db )

OMG, no wonder it was so bloody loud at -15mv, i was actually testing the movie at 1.5db above reference this evening. Luckily no complains from next door.

in case you are wondering what is the trim levels in the AVP, it is showing +0.5 and 0 respectively for LR, and + 4db for centre, mainly because the centre is behind an AT screen.

apart from this, when i calibrated for the Stereo profile, all channels off except subs and front left/right, it seems to be louder than the rest of the profile, easily another +5db louder

One of the other member who bought the same unit, told me the same thing, he said on lyngdorf his familiar soundtrack is played back at -18mv, but on the AVM90, he needs to reduce it to -35! wow, he is like spot on super ears, can tell, about 17db difference. Solid Lau Jiao

I am not sure if the XLR is at 4 volts, it could be higher. not sure why the levels are so loud and not at reference… hmmmm

[Ronildoq]

ok, just verified today. Quick measure validation works.

specially for Ricky, who was asking how to validate. Looks like you can go back to quick measure, turn on ARC measure, then Turn off ARC measure again, and immediately you can see the correction filters being applied, that is if you trust the manufacturer’s software to work correctly

you can see corrections made up to 200hz only as i have selected mains to be corrected up to 200hz.

Wow! thats a nice feature there. Dont need to pull out UMIK-1 mic. But if you need to verify with 3rd party software, by all means you can do that with REW, for free.

You can see what ARC is doing after, and the before measurements on front left speakers. thats really cool!

[desray]

Anthem did not advertise this nifty feature (Quick Measure) enough. It is used mainly to balance the subwoofer levels to be in line with the rest of the speakers. As Anthem only measures one subwoofer and speaker at a time (ARC On/Off), it lacks one critical feature in the Quick Measure - i.e. summation of 2 or more subwoofers into one. As Roni rightly pointed out, we need to utilize a third-party measurement tool like REW to see the combined readings of the subwoofers to see whether the freq response is smooth and flat. I truly hope this simple, yet important feature can be added in future ARC Genesis s/w.

[Ronildoq]

Ok managed to figure out the entire gain optimisation stuff on the AVM90

These few are important .

Notice the profile above is +8.5db calibration level

The results were fronts 0db as below , post calibration circled in red

Now let’s look at music profile calibration

When the calibration levels were 0, the mains were registering +8.5db post calibration

So that’s one point to note, the calibration level changes the results of what’s stored in your final trim levels in speaker page.

Speaking of trim levels, I believe there are 2 stages. One of it being gain stage in the digital domain and another the analog domain. I sus r the one from the main page is in the digital domain before room correction.

See the above, this is how I manually optimised the gain structure on the AVM90

So last night I made one more attempt to re do the calibration, now the system is back to 0 level reference

But before I get to that, I adjusted the levels in the main page, reducing them by -6.5db on all channels

There are pros and cons with the said approach, ideally I would only want to reduce the gain in the subwoofer channels, because we have re-directed bass being sent to the sub channel. A -6.5db is good as it preserves headroom in the digital domain. However, for the rest of the mains and speakers, we would want to keep that at 0db for a better signal to noise ratio.

The solution, increase amplifier gain on the sub amplifier , it will not hurt to use more amplifier power, but keep distortion Low in the digital section with a negative trim on the AVM90 subs level and try keep the mains at 0db

Next, I opted not to boost the bass, or use room gain. They were all left at 0. Instead, I took another approach, using shelving tilt instead with a negative trim coming in at -5.125db like the below

I’m going to try this, because I noticed the speakers sounded a bit bright when I eq up to 200hz for HT.

I played “ A quiet place” scene, and when the fire works kicked in, wow ear bleed man… . For HT, I prefer a downward slope, but for music, I’ll retain the flat target. That’s my personal preference .

So why did I choose -5.125db correction ?

That’s using the calculator

It’s approximately 5-6db tilt at my place

Hope that helps

Oh yes! Just found out there is a calibration mic u need to use and download from this section , hope that helps with the calibration process

https://anthemarc.com/downloads/

Gosh! Nowhere it is mentioned in the manual u need to download and use the mic calibration file. If I’m a newbie, gone case, I’ll be doing it all wrong for most of the stuff… hmm

One final thing I’ve noticed, on room perfect, it takes into consideration the power response from the speakers, whereas ARC doesn’t.

Undisputed room correction champion, no doubt, still Room Perfect…

[desray]

@Ronildoq

Oh yes! Just found out there is a calibration mic u need to use and download from this section , hope that helps with the calibration process

ARC: Downloads

Gosh! Nowhere it is mentioned in the manual u need to download and use the mic calibration file. If I’m a newbie, gone case, I’ll be doing it all wrong for most of the stuff… hmm

Oh no…you don’t need to do that. That is for previous-gen Anthem products in the likes of MRX 1120 and AVM 60. This newest generation (Gen 4) - i.e. AVM 90 does not need any mic calibration file to use for ARC calibration.

[desray]

@Ronildoq

These few are important .

Notice the profile above is +8.5db calibration level

The results were fronts 0db as below , post calibration circled in red

Now let’s look at music profile calibration

When the calibration levels were 0, the mains were registering +8.5db post calibration

So that’s one point to note, the calibration level changes the results of what’s stored in your final trim levels in speaker page.

So what you did here is to decrease the overall calibration MV by 8.5dB and when you do that, you are pulling down the reference level of 75dB to somewhere 66.5dB to be your calibration MV. That is what’s shown in your screenshots. You are going through all these modifications just so you can play louder at 5dB MV?

[desray]

@Ronildoq

Next, I opted not to boost the bass, or use room gain. They were all left at 0. Instead, I took another approach, using shelving tilt instead with a negative trim coming in at -5.125db like the below

I’m going to try this, because I noticed the speakers sounded a bit bright when I eq up to 200hz for HT.

I played “ A quiet place” scene, and when the fire works kicked in, wow ear bleed man… . For HT, I prefer a downward slope, but for music, I’ll retain the flat target. That’s my personal preference .

So why did I choose -5.125db correction ?

That’s using the calculator

It’s approximately 5-6db tilt at my place

I’m quite lost with this calculator and what it actually does, care to share more what those terms and values mean? How does this solve the harsh (bright) trebles for your mains? It will be great if you can you post your ARC calibration file (pdf) here for everyone to take a look? I will be intrigued to see what’s the results for your mains and subwoofers.

[Ronildoq]

@desray

I’m quite lost with this calculator and what it actually does, care to share more what those terms and values mean? How does this solve the harsh (bright) trebles for your mains? It will be great if you can you post your ARC calibration file (pdf) here for everyone to take a look? I will be intrigued to see what’s the results for your mains and subwoofers.

Use this , room curve calculator

https://www.ohl.to/calculators/targetcurve.php

You can always switch between -4 to -6 tilt.

Input your room dimension. But u need to know the decay times from 10-100hz, if the room is treated well, it can range from 0.2ms to 0.4ms, that determines how dry or lively the room is. In my case it was around 0.35ms and upper frequencies will generally decay faster due to shorter wave lengths , so about half of that.

Speaker directivity index will differ from speakers to speakers. So one can try the one I’m using above.

Put in these numbers and it will work out the estimated tilt needed in a room

Hope that helps. I’ll try listening again this afternoon and revert my findings

Still tweaking and exploring. I’m sharing as much along the way. So hopefully it helps, it might work, it might not. No harm trying. Revert to original if not happy with the sound

It works the same as a house curve, the only difference between the other approach vs this is the use of a shelving filter for all speakers, so the levels of the speakers are lower at upper frequencies. Vs using a boosting EQ to boost the subwoofers. So in my case, there is no boosting EQ applied to the subwoofers. The only boosting EQ I have am plied is using the LF boost from the speaker power amplifiers on seaton. So the line arrays, which handle mostly midbass, doesn’t boost below 30hz

In short, to achieve a given target curve, there are 2 ways to do it.

1. You boost the Low frequencies +6db
2. You cut the the high frequencies -6db

The end targeted result is a rising curve at the Low end. I prefer the latter. The levels are lower now, but it is compensated with increasing the master volume on the AVP

This is the better approach for house curve IMO

[desray]

@Ronildoq

Still tweaking and exploring. I’m sharing as much along the way. So hopefully it helps, it might work, it might not. No harm trying. Revert to original if not happy with the sound

Yes, pls do…that’s the whole point

@Ronildoq

It works the same as a house curve, the only difference between the other approach vs this is the use of a shelving filter for all speakers, so the levels of the speakers are lower at upper frequencies. Vs using a boosting EQ to boost the subwoofers. So in my case, there is no boosting EQ applied to the subwoofers. The only boosting EQ I have am plied is using the LF boost from the speaker power amplifiers on seaton. So the line arrays, which handle mostly midbass, doesn’t boost below 30hz

In short, to achieve a given target curve, there are 2 ways to do it.

1. You boost the Low frequencies +6db
2. You cut the the high frequencies -6db

So this is only applicable in your situation since your seatons have that Low Freq boost function. Your method doesn’t apply to the general users here, I guess…

With all these modifications you’ve made, some of which I will not even attempt to do for reasons I will share with you when we meet up for the demo session. Do take note that some of the approach suggested by you is rather unconventional in the Anthem community at large but don’t let that deter you from what you are doing right now…continue to experiment and share in this forum but do make sure we convey the right advice and guide across to our best ability. One thing is for sure, it makes me all the more excited to hear how good you manage to tweak the AVM 90 when I pay you a visit. Hope it’s sooner than later.

[Ronildoq]

@desray

So what you did here is to decrease the overall calibration MV by 8.5dB and when you do that, you are pulling down the reference level of 75dB to somewhere 66.5dB to be your calibration MV. That is what’s shown in your screenshots. You are going through all these modifications just so you can play louder at 5dB MV?

Yes that is correct. So the idea is to bring down the calibration levels to 66.5db at night. I do calibrations at night in living room when everyone is asleep.

The levels are sufficient in my case because the room ambient noise levels are 20db, so there is sufficient signal to noise ratio when using a lower target. In this case it’s more than 40db above noise floor, so that’s sufficient.

The goal is not to play it 5db louder.

The goal was to change the target levels so ARC sets the mains as 0db , so u have all the speakers close to ±3db post ARC

If on the contrary, I set calibration levels to 0, u can see ARC is applying +8.5db to left right speakers. This leaves only +3.5db more max to correct trim on other speakers. So let’s say your centre needs to be at +4db louder than mains, u actually don’t have room for correcting levels anymore

This approach is for those using AT screens where screens attenuates a certain amount of SPLs

[Ronildoq]

@desray

So this is only applicable in your situation since your seatons have that Low Freq boost function. Your method doesn’t apply to the general users here, I guess…

Spot on! This method only applies to those who use an outboard EQ like the minidsp 2x4 or balanced, that has options to use high shelving or has the means to cut frequencies at the Low end instead of boosting . Also for those who use plate amp on subs that has this option

You will be surprised how educated our community in Singapore is when it comes to the use of subwoofers. A lot of the folks I know run more than 2 subs easily….

All 5 who purchased the AVM90 during group buy are all using minimum 4 subs…

Just that a lot of them less active on forums

[Ronildoq]

@desray

One thing is for sure, it makes me all the more excited to hear how good you manage to tweak the AVM 90 when I pay you a visit. Hope it’s sooner than later.

Definitely will host you….will be good to hear your point of view as well… looking forward….

Hopefully I can complete it satisfactorily, some minor touch up and tweaking to go… almost there

Once you have experienced it first hand, you can then give your honest reviews on the AVM90