[audiohobbit]

What's the earliest generation IPad that will run Centro? Is there anything even cheaper than a 6th Gen? For me, running Centro would be the exclusive use for the Ipad, so no point in spending for any capability in excess or that.

I'm anything else than an Apple expert, but the appstore tells me that ios9 is the minimum needed for CentroControl:
https://apps.apple.com/us/app/centrocontrol/id789780943#?platform=ipad

And Wikipedia tells me that even an ipad 2 should work since the highest supported operating OS is ios9.3:
https://en.wikipedia.org/wiki/IPad#iPad_series
But I have no idea if this really works without problems or very slowly etc.

As I said I wanted to buy from Apple directly and no other refurbished dealer, in order to have no problems hopefully (and one year warranty, and a new battery as they claim) and they only sell 6th gen (32 GB) here for 249 eur. On Amazon the cheapest 6th gen Ipad from a dealer was about 220 eur, so not that big of a difference.
An ipad mini btw seems to be more expensive than the normal ipad.

I also only really need the ipad for CentroControl, but I can use it for Logitech and Oppo player app as well here in my home cinema room, and maybe some other apps like home automation etc in the future.

 

[GeorgeA]

There’s a parody photo posted at the facebook page of av-in.com.

The English text by google translate:
“Is this the piece that is missing? In any case, run to see the Eiffel film, preferably in a Dolby Atmos room (Dolby Cinema). The work of the sound team, led by Cyril Holtz, is fantastic. With moments of ultra-immersive auditory jubilation based only on sound effects; I think of these scenes underground, under the pillars, or their assembly, in height. Bravo for this intense Dolby Atmos mix that we owe to Polyson and Archipel, and where, for the anecdote, a processor for headphones SMYTH Realiser A8 as well as an A16 were used, in particular for the pre-mix (not with the HS model of this wink photomontage, but the professional A16 2U Bal). Eiffel has been in theatres since October 13.”

 

[castleofargh]

I see what you mean but I'm not sure if it's totally "true". I got the info directly from Stephen Smyth that the blocked ear canal often is a reason that people say the PRIR sounds too bright.

Good to know, and hard to argue, of course the Smyths know their system best. What it tells me is the blocked ear does not fully eliminate the effect of the ear canal and the residual effect is not fully compensated by auto EQ. If I had good HPEQ measurements, I could likely validate this too. I will check with the HD800 HPEQ that comes with A16.

To be perfectly unclear, I've read(won't claim I understood or remembered it all^_^) about everything related to HRTF models that was ever freely published and then some more where google could cheat the paywall crap(knowledge and science behind paywalls is the burning books of our time). It's very common to see objective approaches(simplified models or fully objective but only for a portion of the total acoustic system), end up with mentions that people found them too bright subjectively.
no matter how much it puzzles me, I have no idea why it happens. Measurements with the A16 might just be one more case. It seems to me that as long as we measure speakers and headphones with the mics at the exact same position(not in 2 different sessions, and not when we feel like the mics might have moved!!!) then the ear canal should not matter as the variable would cancel out, no matter what resonance is or isn't caused by it.
But maybe that's wrong? Maybe the brightness has nothing to do with the ear canal? Maybe it does? Maybe several independent causes just happen to make me see correlation?
And of course, it may be gremlins.

If someone knows and can explain it to me, that would be cool.

 

[audiohobbit]

I quote Stephen here (again) from one of my E-Mails from him:

"When you hear excessive reverb it is almost always due to poor SNR, ie the ambient noise levels are higher than desired. The benefit of fw1.8 is that it includes the de-noise routine.

Brightness is, as you say, due to the fact that the PRIR and HPEQ are measured with a blocked ear canal, whereas the listening is done unblocked, and therefore the ear canal resonance is not correctly equalised. Only a manual EQ can remove it."

 

[KMann]

It's very common to see objective approaches(simplified models or fully objective but only for a portion of the total acoustic system), end up with mentions that people found them too bright subjectively.

I have seen literature that suggests, making measurements at the ear drum instead of using blocked ears produce even better results and may make better BRIRs, but from my own experience using in ear probe mics, it is impractical for a consumer application, though it yielded very good HRTF measurements.

Brightness is, as you say, due to the fact that the PRIR and HPEQ are measured with a blocked ear canal, whereas the listening is done unblocked, and therefore the ear canal resonance is not correctly equalised. Only a manual EQ can remove it."

The measurements do show the ear canal resonance is not correctly equalized, part of the reason may be that the resonance is not the same when you have open air vs a headphone on and even though it is blocked ear and there is residual resonance as in spite of using the foam it is not completely occluded.

I have the measurements of the left channel of one of the speakers from 3DSoundShop. The green trace is with flat HPEQ (i.e, no EQ), the blue trace is auto HPEQ and the magenta trace is with the EQ I did with man Loud, which is my subjective flat. Below is a exploded view.



I have the measurements of the left channel of one of the speakers from 3DSoundShop. The green trace is with flat HPEQ (i.e, no EQ), the blue trace is auto HPEQ and the magenta trace is with the EQ I did with man Loud, which is my subjective flat. Below is a superimposed view. You can clearly see the effect of ear canal resonance between 3k - 4k even after auto EQ, only after I did man Loud did I get it to where I wanted. Having the ability to do the measurements helped because I was able to verify that the effect of ear canal resonance was almost fully gone and it did sound more natural.



What was also interesting was how close the BRIR got to my own personal PRIR (I did man Loud on both). Top 3DSoundShop BRIR, bottom my personal PRIR that I improted from A8. It was also nice to see better sub-bass with BRIR

 

[Dixter]

Nice... I always wondered how they would look charted out... interesting in the spread of the SPL values... ???

 

[KMann]

Nice... I always wondered how they would look charted out... interesting in the spread of the SPL values... ???

I have applied smoothing and speaker measurements in a room, even after EQ, don't generally look as smooth as HP measurements, on top of that we are adding HRTF/HP response. I widened the scale to so that the graphs don't get too ugly and distracting. when I did some tests, they do sound perceptually flat except perhaps couple of areas in bass where I could do some EQ (and I wish man Loud had that option). While it is very hard to fix room response with EQ especially in the bass, it is relatively easy to do it on a HRTF.

 

[KMann]

The quality of BRIR from 3DSoundshop is objectively better (in terms of noise, reverb and clarity) compared to my own PRIR, though my own PRIRs have better depth and sense of space, the BRIR was cleaner and more focused. Below I have plotted the impulse response of the Genlec BRIR from 3DSoundShop vs my PRIR (a planar line array, in a largish carpeted room). Top Green is the Genlec BRIR, bottom Blue is my own PRIR. Both Left channels.

If you are doing your own PRIR measurements, I hope these could help as some sort of reference to check where your PRIRs fall.

 

[KMann]

In case this helps anyone trying to do man Loud:

To do the man Loud HPEQ with a flat curve instead of auto EQ as starting point, I removed the ear gain first, approximately by dialing levels similar to below in man Loud before I started comparing loudness levels in adjacent bands. Even if you are using a auto EQ, based on the measurements I saw, it may be good to start with -5dB instead of -10dB.


Then I first went through all the bands and and made sure the noise bursts were perfectly centered in each band. In my case I found the image shifted slightly to the right. So I adjusted the levels using A and B volume buttons to get them centered.

then I moved through the bands to judge the loudness levels and tried to match them. This can be difficult without prior experience, but measurements can come handy. The idea would be to make the region between 1kHz - 4kHz flattish. As you can see below, I messed up some between 1 - 2khz on one channel



So knowing the measurements, I went back and tweaked the man Loud some more, and got what I now use below:

 

[VandyMan]

That is really useful information KMann, thank you. I'm going to re-attempt a manLoud correction based on it.

 

[audiohobbit]

part of the reason may be that the resonance is not the same when you have open air vs a headphone on and even though it is blocked ear and there is residual resonance as in spite of using the foam it is not completely occluded.

yes I think something like that could be it.

thanks for your graphs. although the scale on the y axis is abit too wide for my taste. Normally I use no more than 50-60 dB on the y axis.

the green trace is with flat HPEQ (i.e, no EQ), the blue trace is auto HPEQ and the magenta trace is with the EQ I did with man Loud, which is my subjective flat.

They all look similar below 500 Hz (at least at this scale), so looks like auto HPEQ doesn't do anything below 500 Hz and maybe I'm right?

Then I first went through all the bands and and made sure the noise bursts were perfectly centered in each band. In my case I found the image shifted slightly to the right. So I adjusted the levels using A and B volume buttons to get them centered.

then I moved through the bands to judge the loudness levels and tried to match them. This can be difficult without prior experience, but measurements can come handy. The idea would be to make the region between 1kHz - 4kHz flattish. As you can see below, I messed up some between 1 - 2khz on one channel

I forgot to say that there are two methods/target curves: Equal Loudness 80, where the 80 phon equal loudness curve is the target, so 1 kHz should be around 80 dB,
and Equal Loudness 20, where the 20 phon curve is the target.

With EQL20 Stephen says you should lower each band so that you can't hear it and then increase the volume so that you just barely can hear it. You need a quiet room for this.
So no comparison between bands but the hearing threshold is used.
I tried this once and found it far more difficult than bring all the bands to an equal volume level with EQL80.
With EQL80 only comparing the volume of neighboring bands can lead to errors, I always switch fast over all bands back and forth (and the A16 allows this forutunately), so I can hear (especially) volume peaks better.

 

[Dixter]

Would be super nice if the A16 would create its own graphs like these... visuals are much easier to see where the adjustments need to be fine tuned... what is required to create these...

 

[KMann]

They all look similar below 500 Hz (at least at this scale), so looks like auto HPEQ doesn't do anything below 500 Hz and maybe I'm right?

I agree, I won't rule out that possibility; but then if you have a headphone that measures really bad below 500Hz, that will be a better reference, the LCD-5 I use is just flat and there is nothing to EQ there.

thanks for your graphs. although the scale on the y axis is abit too wide for my taste. Normally I use no more than 50-60 dB on the y axis.

Yes, it just depends on the objective, a zoomed in graph will be too distracting and is useful if you need to do some precise adjustments, it is easier to see larger features this way and I did not want to alarm anyone (if they are not used to seeing similar measurements)

I forgot to say that there are two methods/target curves: Equal Loudness 80, where the 80 phon equal loudness curve is the target, so 1 kHz should be around 80 dB,
and Equal Loudness 20, where the 20 phon curve is the target.

With EQL20 Stephen says you should lower each band so that you can't hear it and then increase the volume so that you just barely can hear it. You need a quiet room for this.
So no comparison between bands but the hearing threshold is used.
I tried this once and found it far more difficult than bring all the bands to an equal volume level with EQL80.
With EQL80 only comparing the volume of neighboring bands can lead to errors, I always switch fast over all bands back and forth (and the A16 allows this forutunately), so I can hear (especially) volume peaks better.

All good points I used EQL80 because I am quite used to doing similar EQ. The method you suggest would benifit those who are new to EQing based on loudness. I agree, I do a fast pas over the bands too. But at the end, doing the measurement gives you a pretty good idea if you broke something.

 

[audiohobbit]

The problem remains with in ear headphones, even with measurements you have nothing to compare to.

I thought of the following procedure, comparing the sound of the in ear to the auto HPEQed sound of my LCD2 (or any other over ears):
First I have to ensure that all bands are centered when playing the C spkr over the LCD2 (and/or L+R or L+C+R).
Then I put one in ear in my left ear and one over ear over the right ear (with Audeze it would be easy, put the HP on normally and just unplug one side)
Then I play the noise bands and adjust the volume of the in ear in each band so that the sound will be centered again.
Then I do the same for the other combination: in earin right ear and over ear over left ear.

Could this be a way to get the HPEQ done to sound as equal as possible to that of an auto HPEQed over ear HP?

Of course you need 2 amps, but that's not a problem.
The problem is, afaik the HPEQ sounds are only output on the front HP A socket not simultaneously on the back RCA sockets.

 

[KMann]

The problem remains with in ear headphones, even with measurements you have nothing to compare to.

I thought of the following procedure, comparing the sound of the in ear to the auto HPEQed sound of my LCD2 (or any other over ears):
First I have to ensure that all bands are centered when playing the C spkr over the LCD2 (and/or L+R or L+C+R).
Then I put one in ear in my left ear and one over ear over the right ear (with Audeze it would be easy, put the HP on normally and just unplug one side)
Then I play the noise bands and adjust the volume of the in ear in each band so that the sound will be centered again.
Then I do the same for the other combination: in earin right ear and over ear over left ear.

Could this be a way to get the HPEQ done to sound as equal as possible to that of an auto HPEQed over ear HP?

Of course you need 2 amps, but that's not a problem.
The problem is, afaik the HPEQ sounds are only output on the front HP A socket not simultaneously on the back RCA sockets.

Yes that is is a nice idea and should work, yes you will need two amps (does A16 allow both digital out and HPA to be active simultaneously?), you will also need to level match may be at 1kHz for volume.