Marantz AV10 AV Processor Review

[Dj7675]

Yesterday I concluded my ART support tweaking after going from 7.2.4 to 7.4.4, and I definitely hit the 96 cross-term wall. When optimizing for corrected spreads, my overall responses deteriorated compared to the 7.2.4. I spent all day tinkering by adding/removing support speakers to find the optimal support for each speaker, and I confirmed that the corrected spread would benefit with additional cross-terms.

But again, I still don't know if if the visual improvement reflected in the corrected spread would be audible. I tend to think it may be evident only at reference levels and momentarily when certain unoptimized frequencies are played, which depends on content.

I too have the AV10 and with a 9.4.5 sysytem ran out of cross terms on what I wanted to do. My solution was to split all groups into individual channels and with some trial and error use the support channels which were most effective. Worked well for me.

 

[Pizzaiolo]

I too have the AV10 and with a 9.4.5 sysytem ran out of cross terms on what I wanted to do. My solution was to split all groups into individual channels and with some trial and error use the support channels which were most effective. Worked well for me.

Wow, that's an enormous system, not surprised it smoked the AV10's DSP processing capabilities. Splitting all those groups must have taken hours, if not days, of trial and error. I split my subs into 2 groups and optimized corrected spread for all speakers, which took ~4 hours. I then split another group and began the same process, but in my room that additional effort yielded minimal improvement.

Based on their limited DSP processing power, I really can't recommend D/M processors for those with systems larger than 7.4.4 seeking to use ART.

 

[Dan1210]

This is a fascinating discussion, i currently have a 9.4.3 system running through my Avm70 in a treated room so this is quite interesting to me and whether or not i would move to the Av20

 

[dlaloum]

Wow, that's an enormous system, not surprised it smoked the AV10's DSP processing capabilities. Splitting all those groups must have taken hours, if not days, of trial and error. I split my subs into 2 groups and optimized corrected spread for all speakers, which took ~4 hours. I then split another group and began the same process, but in my room that additional effort yielded minimal improvement.

Based on their limited DSP processing power, I really can't recommend D/M processors for those with systems larger than 7.4.4 seeking to use ART.

The default all supporting all definitely requires more than 96 cross terms for 11 channels (let alone 7.4.4 - which is 15 channels)

However if you remove the height speakers (which normally have minimal bass capability) from the "supporting speakers" category - that should bring it down to 11 channels on a 7.4.4 setup from where some tweaking such as not supporting fronts from rears.... having fronts support FH's and rears supporting RH's etc... would bring things down to within the 96 cross terms limit.

With a 5.2.4 setup this approach led to my setup using 54 cross terms with subs in a single group or 63 cross terms with my two subs split into individual groups (the later configuration provided no improvement either objectively via REW or subjectively, so I reverted to single sub group)

So what I am saying is that the D&M 96 cross term limit does allow very very good configurations that are within its channel limitations - but for those using more than 9 channels - the allocation of those cross terms, ie: the detailed definition of the speaker configuration for support becomes critical, and performance can degrade if you choose a configuration option that requires more than the 96 cross terms - and sadly, Dirac won't report how many it "wants" it merely reports how many it is using - which maxes at 96.

It appears to me that Dirac calculates based on an "unlimited cross terms" schema, then uses the first 96 of those cross terms.... leaving gaps if it's calculated schema requires more than 96... resulting in less than optimal performance.

So if the limit is 96, the safest approach is to choose a schema that uses only 95 cross terms maximum, so that you know for sure that it is calculating a configuration that works in real life. - if your configuration uses all 96 cross terms, there is a substantial likelihood that the configuration is actually calculated for MORE than 96 - and therefore the actual setup with 96 terms in use, will provide sub-par performance.

Right now economical alternatives with substantially more cross terms are either much, much, more expensive (StormAudio), or AVP only (Monoprice) - no AVR options that I am aware of have more than 96 cross terms.

 

[andrewskaterrr]

No promises, but I'm now a dealer for some of the major companies and have already started asking how many cross term limits products have and if it can be public info. IMO this is important info that people need to know when choosing a device. I didn't know about this until you posted about it, and now I most likely won't be upgrading to a AV10 like I planned to as I'll be doing 9.4.6 at my next house.

 

[dlaloum]

No promises, but I'm now a dealer for some of the major companies and have already started asking how many cross term limits products have and if it can be public info. IMO this is important info that people need to know when choosing a device. I didn't know about this until you posted about it, and now I most likely won't be upgrading to a AV10 like I planned to as I'll be doing 9.4.6 at my next house.

I was also trying to point out that I have objectively measured exemplary results on my setup (5.2.4) using only 53 cross terms...

 

[andrewskaterrr]

I was also trying to point out that I have objectively measured exemplary results on my setup (5.2.4) using only 53 cross terms...

This would be for setups that need more help or setups that want to include all speakers with ART in a 9.4.6 system.

 

[Oddball]

This would be for setups that need more help or setups that want to include all speakers with ART in a 9.4.6 system.

In my experience with 9.4.4 system, most of the support even up to 100hz range comes from subs. This is even if your bed speakers are relatively large and dig to 30-50hz range with authority, and you have entered these parameters in the support range and assigned high support level such as -18. There are ways around it by playing with parameters, but if you honestly enter the ranges, this is the outcome. I leave it at this as sometimes listen to reference levels where sub support in sub 100hz range is smoother with 4 large-r subs than with big towers. My best bet (but entirely a guess) is that subs are preferred as easier to align phase and timing in that range with 4 subs than e.g. additional 6 supporting speakers.

So the question how much cross support does one need is inherently tied to the system and room. Would definitely be better that D&M have more filters for larger system as that would provide more options. Having options, even if not used is value of its own. I have not heard why D&M chose this approach and if this is their DSP limitation or just a decision to simplify the ART implementation. Would be good to know.

As @dlaloum pointed out, ART might consider different approach with 2 subs and relatively large bed speakers and use more of the bed speaker support.

 

[Sal1950]

As @dlaloum pointed out, ART might consider different approach with 2 subs and relatively large bed speakers and use more of the bed speaker support.

At this point maybe we could move further discussion on this particular ART issue over to the "Dirac ART on Marantz AV10" dedicated thread.
Thanks.

 

[Oddball]

At this point maybe we could move further discussion on this particular ART issue over to the "Dirac ART on Marantz AV10" dedicated thread.
Thanks.

True. But so much in there that one could easily get lost. Perhaps time for you Sal to go ART and start a new better thread with a summary of findings? After all what do we have senior members for (mean by posts and likes - I feel just like apprentice around here).

 

[Sal1950]

But so much in there that one could easily get lost. Perhaps time for you Sal to go ART and start a new better thread with a summary of findings?

So much in here that one could easily get lost. LOL
Another ART thread, do you know how many we already have?
Thanks but I'll pass.

 

[dlaloum]

This would be for setups that need more help or setups that want to include all speakers with ART in a 9.4.6 system.

It is not a matter of "including all speakers" - my setup does in fact include all speakers - my heights are supported by my subs and mains - but my heights are not used to support the other speakers as they have no output below 150Hz!!

So what I am saying is that it does not involve excluding speakers - but APPROPRIATE inclusion...

So having my rear/surround full range speakers support the subs makes sense - having them support the rear heights also makes sense, I do not have them supporting the front mains - which enhances localisation... the front heights are supported by front mains and subs, but they in turn do not provide support.

ALL speakers are included in the ART schema - but they are included, "intelligently" - whereas the ART default setup has every speaker supporting every other speakers - and that, demonstrably, is NOT an optimal configuration - it compromises immersion due to compromised localisation within the 50Hz to 150Hz range, and it makes inefficient use of cross-terms.

Once a configuration is set up logically, taking into account which speakers should be supporting which other speakers, and logically excluding those that should not be supporting.... the cross-term budget becomes much much less constraining.

But yes I agree that configurations with more speakers are going to need more cross-terms, and in a "maximal" 7.4.6 configuration, 96 cross terms may well be a constraint.

I will also make another observation - with 4 full range main speakers - a single sub ended up providing excellent results - adding a 2nd sub did make a marginal measurable difference, but one that was sufficiently slight to be completely inaudible.

Many of todays configurations are based on assumptions such as mains being limited to 80Hz, and the bass being provided by subs, which in turn drive a "norm" of having multiple subs.

If people opt for full range towers, then the need for multiple subs reduces dramatically - and the associated need for additional cross terms is also reduced.

I don't believe most people have fully internalised the paradigm shift that ART involves - the improvements available through the use of full range mains, and how to optimise accordingly are really not being talked about.

Subs are no longer absolutely needed in the 40Hz to 150Hz frequency range - full range mains can do that - subs can be used for the additional extension down to 20Hz AND to boost total SPL's (as full range mains are typically more SPL limited) - configuring in this manner, with fewer subs, also reduces the cross-term requirements - you effectively have 4 more subs, which use no additional cross terms as the 4 mains already have cross terms allocated (and this becomes even more the case if you have 6 mains/base layer).

 

[dlaloum]

In my experience with 9.4.4 system, most of the support even up to 100hz range comes from subs. This is even if your bed speakers are relatively large and dig to 30-50hz range with authority, and you have entered these parameters in the support range and assigned high support level such as -18. There are ways around it by playing with parameters, but if you honestly enter the ranges, this is the outcome. I leave it at this as sometimes listen to reference levels where sub support in sub 100hz range is smoother with 4 large-r subs than with big towers. My best bet (but entirely a guess) is that subs are preferred as easier to align phase and timing in that range with 4 subs than e.g. additional 6 supporting speakers.

So the question how much cross support does one need is inherently tied to the system and room. Would definitely be better that D&M have more filters for larger system as that would provide more options. Having options, even if not used is value of its own. I have not heard why D&M chose this approach and if this is their DSP limitation or just a decision to simplify the ART implementation. Would be good to know.

As @dlaloum pointed out, ART might consider different approach with 2 subs and relatively large bed speakers and use more of the bed speaker support.

My own experimentation (and with the strong proviso of my setup being based on 75db SPL continuous, and 95db peak...) I have had no issues using my base layer speakers (spec F3=24Hz and 35Hz) with additional support from 1 sub to achieve exemplary results. Adding a 2nd Sub, did very marginally reduce decay further - but made no noticeable audible change.

It is possible that were I to listen at a reference level of 85db with peaks at 105db or 115db for LFE) - my response MIGHT be different... maybe.

I will also note something that others have not remarked on - in the bass range - between 50Hz and 35Hz, my mains in their ART supported configuration (supported by mains and subs) have lower measured THD than in their non ART configuration - in other words the total SPL for any 1 channel is a result of the output of that channel PLUS the assistance received from the other channels. - They are working less hard, and as a result there is less distortion.

This also means that their individual channel max SPL has been increased - if the speaker spec says 105db SPL max - it is likely that in supported configuration that speaker would now be a 110db max speaker or more...

The system as a whole has substantial additional headroom due to it being a cooperative effort rather than an individual one.

Having said that, I have not had the time and energy to go through extensive measurements to demonstrate that...

However my current conclusion is that the use of full range mains has not only reduced the need for subs due to additional "subs" in additional geographical locations (resulting in improved smoothness of the F/R / reduced nulls/modes) - but also reduced the need for subs to provide higher SPL capabilities for the LFE.

 

[Sal1950]

However my current conclusion is that the use of full range mains has not only reduced the need for subs due to additional "subs" in additional geographical locations (resulting in improved smoothness of the F/R / reduced nulls/modes) - but also reduced the need for subs to provide higher SPL capabilities for the LFE.

That is a pretty extreme case statement and all depends on how "full range" the towers are.
Not many reasonably sized towers have strong response down to 20hz and below.
It is all pretty relative.

 

[peng]

People are too ocd, imo just let the lcr do their job naturally even if they are all Revel Salon2, and let the subs do the ART thing, that’s why D+M gives you 4 Subouts. Humans don’t need bass resolution to be super high, so I would’t worry about the limits either.

 

[Oddball]

That is a pretty extreme case statement and all depends on how "full range" the towers are.
Not many reasonably sized towers have strong response down to 20hz and below.
It is all pretty relative.

You are being too kind Sal. No passive towers, in fact, have a "strong" response to 20hz and I would say, none below that. Even at 50hz it really depends on definition of "strong".

Let's take an example. Arendal 1528 towers - the biggest, heaviest and really bad ass passive towers. They are down almost 20dB from 60hz to 20hz, per Audioholics measurements. And Gene swears that they hit low bass like no other thing in various videos. Then, on the other hand Gene has 4 active 12" subs in his towers, and 24" tectonic sub in his setup. Don't follow him any more but last I checked this was it. Would not be surprised if he got some more subs in the meantime.

So, one should form their own views as to what is lip service (to be very polite), what is where people put their money, and what the truth ultimately is.

If in doubt, stack the subs as @peng and many others (myself included) suggest. It is a safe and no-fail solution for low end.

Arendal Sound 1528 Tower 8, Center 8+ Loudspeaker Review

Arendal Sound 1528 Tower 8 and Center 8+ speakers deliver audiophile sound, advanced tech, exceptional build quality & value, but there are caveats. Our review dives deep with measurements & listening.

www.audioholics.com

 

[Sal1950]

You are being too kind Sal. No passive towers, in fact, have a "strong" response to 20hz and I would say, none below that. Even at 50hz it really depends on definition of "strong".

Well you know me, I'm always super careful not to step on anyone's toes or hurt their feelings.

 

[dlaloum]

That is a pretty extreme case statement and all depends on how "full range" the towers are.
Not many reasonably sized towers have strong response down to 20hz and below.
It is all pretty relative.

I'm not disagreeing with you...

But then a heck of a lot of smaller subs aren't capable of 20Hz either! (I've argued for years that they aren't subwoofers, but external woofers, to provide the extension lacking for typical bookshelf speakers)

My mains and rears/surrounds are capable of (from the spec sheet...) 24Hz and 35Hz respectively - my subs theoretically extend that range down to 22Hz (and as a sealed design, have performance down below that as their F/R slope is relatively gentle)

In terms of measurements - when 2 such subs are combined with those full range mains - the result is exemplary frequency response, and decay control down to around 40Hz... decay control degrades below that and in real terms frequency response starts to drop off gently from 28Hz onwards.

So if I wanted true 20Hz extension, I would need much beefier subs (not an option in our living space!)

And I'm not even hinting at the considerations required if I wanted my F/R to extend down below 20Hz - as the bass heads in our community like to do.... especially at the SPL's that some people aim for (130db bass!!)

But when you set aside such edge cases, and consider the vast majority of setups, along with WAF and living space considerations, I think my comments reflect a pragmatic reality....

Here is my current measured performance for the FL channel - 5.2.4 setup using Dirac ART:

Yes I would like to improve the bottom end - but honestly, I do not find it lacking in any way, and performance is now far better than it has ever been before...

Here is the base performance before Dirac ART...

Having 4 full range (24Hz / 35Hz) speakers has allowed for far better performance than would have been possible with just 2 subs
(interesting that the ART configuration tends to roll off the bottom end a bit... I could probably push my subs a little harder...)

 

[HiFi 5.2]

Does anybody know if the AV 20 measures as well as the AV 10?

 

[IamJF]

And if/how much AV30 takes a hit compared to these 2?