New loudspeakers & room digital correction service

Have been using HAF filters for over a month now in HQP with Roon front end. Very happy, particularly with the cross talk reduction filters - my horns now image as well as my previous electrostats.

The filters are better than anything I could create with REW ot Acourate (using Mitch's techniques).

@sixman

100% agree with you.  They are so good I don't even think about additional tweaks.  I just listen to music

Highly recomended. Got correction zip file implemented to Roon yesterday. Very quick service. 

I plan to maybe write a small guide of how to do this without using REV as your tone generator. But play direct from Roon and only measure using REV. 

However all all information can be found here. 

https://community.roonlabs.com/t/roon-home-audio-fidelity-room-correction-convolution-filter-creation/29389/111?u=r1200cl

Thierry and Group:

I have few questions before embarking on testing HAF.

First, my system usage:

Tidal/Library managed by Roon, then HQPlayer transcoding everything to DSD256, the max for my dac. Speakers are electrostatics. Music is mostly classical, a lot of opera since I work for an opera company, and folks like Neil Young, Leonard Cohen, etc.

Questions:

Would it be better to acquire the UMIK-1 from Cross-Spectrum Acoustics to get a premium mic calibration or does it not really make any difference? I ask because CSA is out of stock and I would have to wait some unknown time.

In my system, is it better to have the filters made for Roon or HQPlayer? Which option would sound better, if it makes a difference at all? Would either option increase cpu usage significantly to cause a problem? I’m currently at 30% cpu utilization using Poly-Sinc-Xtr.

Do you find the correction usually more or less effective when using dipole speakers or does it not make any difference?

For the free evaluation, do you supply both the Standard and Excellence corrected files so I can determine which I like better in my system?

Thanks! I’m really looking forward to this.

Hi and thanks for your interest ! 

Here are my answers:

Having a premium mic calibration is not mandatory as I am seeking minimal correction to protect the transparency & natural character of loudspeakers: this is why having a perfect absolute values for the measurements is not a must.

Convolution is a pure mathematical operation and it will give the same results with all players, what can make the difference is what happens before & after this operation. The most common way of performing convolution is using FFT so there should not be CPU usage difference b/w the different options. If you want to check the CPU usage I can provide you test filters from other customer.

Having a dipole is usually an advantage for crosstalk reduction, otherwise the correction is mostly dependent on the room environment: it is logical to get bigger improvements with difficult cases.

And yes you get both versions of the filters for the free trial.

Best,

Thierry

Thanks Thierry:

UMIK-1 ordered. I imagine it will be about 30 days before I send files for test.

Hi Thierry et al,

Curious about using your service, I have a few questions but first a little context, I'll try to keep it brief. 

I'm new to convolution, I've been led here to check if there's a cheaper alternative to buying a DEQX. I have a pretty simple system: https://www.myhifiprofile.com/system/1 and I want to:

a. integrate subwoofers (line level inputs) with a low/high pass crossover

b. time align subs with mains

c. linearize phase of the mains 

d. room correct 

e. keep on using my Chord DAC for the mains
f. not under any circumstance introduce an extra AD/DA cycle

g. lossless (64bit) volume control in the digital domain (would have to be in the player). 

It's amazing to me that this is so hard to do, I guess most people are using (old) passive crossover to integrate their subs, no wonder "common audiophile wisdom" indicates subs are a bad idea because they are never seamlessly integrated... 

A DEQX can do all that by using one if it's digital outputs (to mains via Chord DAC / power amp) and analog outputs (to subs).

Hopefully I can achieve all of that with something like this https://www.minidsp.com/products/usb-audio-interface/u-dio8 and another DAC for the subs (going to be interesting time aligning all of this). Ideally I'd like to use the convolution engine in Roon (or I could do HQPlayer too if need be, though I don't own a license currently). 

My questions are:

1. Do I need a single set of filter that achieve all of the above or can I create a pipeline of filters? 

2. If the latter, where would your filters fit in the pipeline? 

3. Also related, I assume I need to provide measurements for you with the subs already crossed over and integrated?

4. Are your filters double precision? wav or dbl format?

Thank you!

Peter

Hello Peter,

You can certainly go for an extra DAC to manage your subs : with that option you can handle everything within Roon with one set of filters per DAC.

Regarding my service you have 2 options: you can define by yourself the crossover + time alignment filters and perform full range measurements with that configuration. I will use those measurements for the correction and merge the corresponding filters with the crossovers ones so that you get only one filters set per DAC.

The other option is to let me handling as well the crossover & time alignment definition. The rest of the process will be the same.

Regarding your last question I generate filters in wav format with double precision.

Hope this is clear enough,

Thierry

After close to 90 years, we are finally seeing some good effort to address the XTC. Keep it up!

Thierry — I have a 5.2 system using Roon for two and multichannel audio. Currently, I’m using Audyssey MultXT32 in my Marantz prepro, which integrates the two subwoofers. However, I believe that Audyssey is at a 48k sampling rate. Do your filter sets work for both stereo and multichannel and if so, how does one do the multichannel measurements with a two subwoofer environment?  And are your filter sets at 192k?  I have a UMIK-1 mic and I run Roon on an i7 server which I presume has the adequate horsepower to do what your filters need. Kindly advise. Merci. JCR 

Hi

Even if my filters are more focused on stereo reproduction they can be used in a 5.2 system using the flexibility offered by the Roon convolution dsp.

And yes I generate filters up to 384khz as a basis. Don't hesitate to contact me using this e-mail [email protected] to exchange more in details about your configuration.

Thierry

Thierry 

Hi again. I asked a few questions on Devialetchat and appreciate your responses. Rather than go back and forth thought worth asking here, as above posters have asked about integration of Sub with mains.

I have a 2.1 set-up:

• Devialet 1000Pro
• Antipodes CORE music server
• Lenehan Audio ML5 Mains speakers (2 way floor stander)
• Paradigm Sub-1

Attached is my current configuration for Sub1 integration with my front speakers via my Devialets. So I have it currently set at High Pass Filter, 80hz cut-off, 2nd Order.

Do I need to do anything else other than do my new measurements via REW or your new tool. Also do you think my setting for 2nd Order Filter at 80Hz would be correct (thinking of dropping this down to 40Hz so the Sub only covers the low range).

I'll need to get new filters soon as having my current speakers upgraded (same enclosures, but new tweeter and crossovers). Will have to wait a month to let the new speakers burn in (need to get some hours into the new crossover caps), but want to get things in order before I do them (also gives me a chance to try your new measurement tool, or re-do with REW).

Cheers

John

Hello John

For the sub integration you have different alternatives : you can either make an arbritary choice for the crossover frequency (like 80 hz) or you can perform room transfer measurements of each speaker alone (sub, main left, main right) to better optimize the crossover strategy (including potential overlapping as having multiple sources in the bass region can be beneficial)

Once you have a crossover defined you need to perform the different full range room measurements.

Cheers

Thierry

@fresponse

Your website states:

• If you have the appropriate calibration file (90° for UMIK) point the mic to the ceiling otherwise point it to the midpoint b/w loudspeakers

I’m not sure I understand this 100%, as all UMIK has a calibration file. 

So what is your recommendations with a calibration file present ?

Ceiling or midpoint ?

• Set the output level as if you were listening to music

Have you abandon the use of pink noise at 75 dB and the sweep measurement file ?

When I did my measurements I played everything through Roon, (and even with HQPlayer in the chain, which probably was a mistake). 

Files I used was found here. I expect this is the most correct way to do it if using Roon. 

I will send you new measurements in a couple of weeks. Only using Roon. 

I like to do two measurements file. 

One with nothing added in Roon and one withe these settings:

I would like to hear your recommendations for Roon settings.

Should headroom management be enabled ?

Any other settings ?

On 28.3.2018 at 1:07 PM, fresponse said:

Last but not least, we propose a simple tool for room transfer measurements (inc. remote control with tablet or phone)

Click on that measurement file package in your website from an iPad was not successful. But maybe that’s not meant to be or ?

37 minutes ago, R1200CL said:

 

Click on that measurement file package in your website from an iPad was not successful. But maybe that’s not meant to be or ?

R1200CL

See attached (Hope Thierry doesn't mind me posting them on here, if not I'll delete them).

if not correct for me to load files here see link  HAF Measurement Tools

HAF_measurement.pdf

sweep_signal_IR.wav

HAF_measurement.pd

4 hours ago, fresponse said:

For the sub integration you can perform room transfer measurements of each speaker alone (sub, main left, main right) to better optimize the crossover strategy (including potential overlapping as having multiple sources in the bass region can be beneficial)

 

Thierry

Sorry to be such a square, but how do I go about doing this, as it's sounds from your explanation that this would be the better option. Also can I do this using your tools, or with REW?

Cheers

John

Hello,

@R1200CL

- If you want to go through Roon for the measurements you must use REW as I have not included this option in my tool

- I recommend keeping your usual Roon settings if you go through it for the measurements : if you hesitate b/w two different settings we can check if they have an impact or not on the acoustic measurement

- There are other measurement mics than UMIK and they usually come with only one calibration file, this is why I mention the 2 options for mic orientation : if you have UMIK it is easier to manage measurements by pointing the mic to the ceiling

- You can still use pink noise to set the volume, but I realized that it brought some confusion for REW newbies : using the normal listening level is a good & less complicated option

@Celts88

I can guide you if you'd like to fine tune your sub integration : you just need to perform acoustic measurements of each element separately at listening position and send me the files.  

Cheers,

Thierry

On 4/3/2018 at 2:55 PM, fresponse said:

@Celts88

I can guide you if you'd like to fine tune your sub integration : you just need to perform acoustic measurements of each element separately at listening position and send me the files.  

Thierry

Thanks for the offer, I'll take you up on that. Not in an immediate rush as I'm currently having my speakers upgraded (new tweeters, crossovers, caps, isolation footers, etc., etc.) so will have to wait until I get these back (in 2 weeks time) before thinking of doing new measurements.

As the Tweeters / Crossovers / Caps are new would also like to give the speakers a good burn in before doing the measurements, but not sure how much this will affect new Filters (looking at getting the Excellence Head filters). Thinking at least 1 month should be enough time, but some say at least 2 months minimum.

When doing the sub integration measurements do I do the same sweep measurements required for filters (Pink Noise, etc.), with Left Main, Right Main & Sub as 3 seperate group of readings. Also, I wouldn't have to do the multiple readings required for the filters (is one per speaker sufficient, sitting in the listening spot)

Cheers

John

Hi John,

You need to set the overall volume (either with pink noise or just using your standard listening level) and then make only one sweep measurement per loudspeaker (main left w/o sub, main right w/o sub, sub) putting the mic at the listening position.

Cheers,

Thierry

@fresponse

Ok, so just one sweep measurement per speaker, but how do I do just the subwoofer?

Do I play music (or Pink Noise) and disconnect my speakers cables leaving just the sub (assume I also stop all crossover points within my Devialet amps, and disable my HAF Filters within Roon)?

Thanks for your help.

Cheers

John

Hi,

You have to set the level using your standard listening configuration (sub active) and then you remove all filters (Devialet & Roon) and perform sweep measurements for each driver independently (either by muting if possible with the Devialet or by unplugging one speaker cable)

Cheers

Thierry

@fresponse

I'm trying to test out how Roon with HAF sounds VS HQPlayer embedded with HAF.

This is not working so far.

One fundamental difference is that it seems to me that HQPlayer embedded needs one file pr. speaker.

And I don't know if HQPlayer embedded even accepts ZIP files.

Can you please advice and maybe consult @Miska in this matter. Maybe update your website with a small user guide ?

Is is so simple that I have to unsip the HAF file and upload several times the different filters ?

Only wav ? What about the cfg files ?

Hi,

I usually explain the procedure by mail to the ones interested in using HQP instead of Roon. You can get in touch with me as some pre-processing is needed to convert the stereo files to mono ones.

Best,

Thierry