Re: Considering Audiolense

Mikkel

unread,

Aug 6, 2012, 6:27:56 AM8/6/12

to audio...@googlegroups.com

Hello Scott,

As a novice user of Audiolense I'll give an answer to those questions which I can answer. My setup is a PC connected to an Harman Kardon AVR245 receiver and a 7.1-setup.

I have used REW to generate filters. Moving to Audiolense I have experienced a significant difference in the audio quality. Most noticably is the time domain correction but even the frequency correction is much more even with audiolense.

My room is an untreated living room with plasterboard ceiling and painted brick walls, a thick carpet, big couch and a few other inventory. So a pretty bad environment seen from an acoustic perspective. Even given these conditions Audiolense makes a huge difference.

Ad.1: As you can see above, I run a similar setup. Nothing fancy in my setup.

Ad.2: Can't really say. Haven't made much comparison.

Ad.3: Speakers are mounted on the wall, left front in a corner, the room is somewhat reflective... yet I hear no pre-ringing. I don't even believe the impulse response shows any preringing (but I would have to check)

Ad.4: Bernt or others would have to give you a qualified answer... it depends on the filter in your AVR but my guess is Audiolense is on par/superior. I let Audiolense handle it and it sounds terrific.

Ad.5: I'll let others answer... but I would prefer having as few instances handle the audio as possible. So I would let Audiolense handle the correction and bass management

Ad.6: Bern't will have to answer... but TTD is one significant feature I beleive

Sorry I cannot answer all your questions but I can assure you that you will find a huge difference using Audiolense compared to REW. I have never hear audyssey so I cannot comment. There seems to be a tendency to discard audysseys room correction many places on the net... but don't take my word for it as I have never experienced it.

Subjectively, and having played classical music in many different environments I'm very critical when it comes to sound quality. With Audiolense Bernt have managed to improve my very mid-range system in ways that made me think I was in the concert hall listening to Tchaikovsky's symphony no. 6. It is that good.

Btw, I use a ruler flat target curve. At least for classical music I find too much information being lost using e.g. the Brüel and Kjaer house curve.

Best regards,

Mikkel

Den mandag den 6. august 2012 05.07.34 UTC+2 skrev Scott:

Hello,

After playing with REW and Jriver's Parametric EQ, I have noticed some very noticeable improvements in my system. But like most hobbyists I've been wondering if there's more to be gained. I plan to play around with the demo when I get some time, but I have some questions because my setup seems to be much simpler than the average user here.

First off, I currently have a Denon 2312 that has Audyssey multeqXT and 5.1 surround (Ascend LCR, HSU sub, JBL surrounds), in a small and untreated room. All bass management and amplification comes from the AVR, which is connected to my computer via HDMI. I currently use Audyssey with slight adjustments in Jrivers parametric EQ to give me a house curve. To my untrained ears, it already sounds pretty good, but these posts about room correction rivaling expensive studios has me very intrigued. I also like the ability to be able to tweak the target curves and the filters themselves. I've already seen what a difference minor corrections can make while playing with REW.

So my questions are:
1) Does anyone have run a similar setup? I see quite a few people use Audiolense for it's active XO capabilities with expensive DACs and power amplifiers, but does anyone use a more consumer oriented setup with Audiolense in favor of AudysseyXT or XT32?
2) I've read Audyssey is also a convolution solution and uses "minimum phase with some tweaks" while Audiolense can do both minimum phase and linear phase corrections. What are some subjective audible differences between the methods?
3) It seems the #1 issue with time domain correction is audible pre-ringing. Just how much of an issue is this a typical setup? Does a typical user live with a small amount of pre-ringing artifacts in favor of the other benefits or can it be removed completely with enough tweaking?
4) Is there an advantage to letting Audiolense do the bass management from mains and surrounds to the subwoofer instead of using the AVR? How is that accomplished? Setting the speakers to large and letting audiolense set high pass filters?
5) related to 4), is it better to just let the AVR do the bass management and EQ the separate channels with the sub active?
6) What are the main differences between the "surround" and "XO" versions that I should be concerned with considering my current equipment?

Thanks for your time.

Bernt Ronningsbakk

unread,

Aug 6, 2012, 8:06:44 AM8/6/12

to audio...@googlegroups.com

Hello Scott,

As every setup is different to every other setup, there will also be difference with regards to how well Audiolense works, what to expect of improvements and what problems that may appear that needs trouble shooting. But I think Mikkel’s response is quite representative of what to expect.

Ad 2: Audyssey creates minium phase corrections. Audiolense Surround produces minimum phase corrections with linear phase crossovers, if you use bass management in Audiolense. Audyssey & Audiolense are in principle quite similar here in how they work and the issues they attempt to correct. You get time alignment of the speakers, and a level matched frequency correction across the speaker setup.

I don’t know your surround receiver and the Audyssey version that’s running on it, but it has very often been a somewhat stripped down version because the DSP chips in the processors have their hands full with video processing, and the capacity left for doing sound correction is somewhat limited. Based on what I have seen I would say that it is quite normal for a surround processor that the correction filters are less effective than state of the art frequency correction below 100 Hz. The lower you get down in frequency, the longer filter you need to do an effective correction, and the longer filters you use, the more DSP capacity you need….

Also, based on what I’ve seen of measured results after Audyssey I think it is safe to say that most people will get a more precise frequency correction with Audiolense. Audyssey seem to look like a half done job after correction by comparison. I bet some of this is by design, some effort to preserve the “natural” tonal character of the speaker, but in my experience, a proper full range correction towards a target that is tweaked, tested and accepted usually provides a far better result.

If you move to Audiolense XO you also have the opportunity to do time domain correction. And you can do it in two versions: One corrects the speakers towards a minimum phase target, which gives a fastest possible rise time. The other corrects the impulse towards a linear phase target, which gives the most simultaneous arrival of all frequencies. What they have in common is that they clean up in the time domain. You get better separation between direct sound and delayed/ reflected sound, and this tends to have a very positive effect on the overall sound quality.

Ad3: It is usually not a big issue. It took some three years before it even surfaced as a real problem. And now Audiolense has tools to negotiate it.

Ad4 & 5: If you have connectivity to let the computer see all your speakers as separate entities including the sub you will probably get the best results by letting Audiolense handle the bass management. It is just about as precise as a crossover between any two speaker drivers can be made to work. But even if it is more practical to let the receiver do bass management, Audiolense will improve e.g. the woofer integration since it will see woofer plus main speaker as one speaker and correct the sum of the two. If the bass management in the receiver is decent the difference between the two approaches are not likely to be very big, so I recommend that you at least start out with something that is practical given your current sound card and receiver interfaces.

Ad 6: Audiolense Surround supports bass management and all the usual surround formats. Audiolense XO has everything that Audiolense Surrond has. But it provides somewhat more flexibility for the crossover between sub and mains, support for digital active speakers, time domain correction, is generally more tweakable and will potentially give you a slightly better sound quality. You can always start with one and up- or downgrade to the other within three months without getting any “penalty” costs. Overall I think the XO version is the better buy because the price difference is quite moderate and you get a lot more tools to dial in your preferred sound quality.

Kind regards,

Bernt

--
Audiolense User Forum.
http://groups.google.com/group/audiolense?hl=en?hl=en
To post to this group, send email to audio...@googlegroups.com
To unsubscribe, send email to audiolense+...@googlegroups.com

Scott

unread,

Aug 6, 2012, 8:25:48 PM8/6/12

to audio...@googlegroups.com

Thank you for the great reply. This is basically what I was hoping to hear. My room is also pretty bad acoustically with tile floors, asymmetrical layout, closet openings and various windows, and not a whole lot of room for treatment or repositioning. I have been wondering how REW generated filters loaded into a convolver differ from the filters generated by Audiolense, but it's good to hear from someone who has tried both approaches.

I take it you are running the XO version?

Scott

unread,

Aug 6, 2012, 8:54:30 PM8/6/12

to audio...@googlegroups.com

Hi Bernt,

Thank you for the very informative reply. Your observations of Audyssey seem to be on par with what I have measured. It seems to do a reasonable job of keeping frequency response relatively flat in the upper ranges, but struggles a bit on the low end. Relatedly, how does Audiolense allow such high correction levels and achieve such flat response? I have read posts stating 12+ db correction. My very limited reading has suggested that minor boosts can take significantly more amplifier power. In the Parametric EQ school of thought, it seems they ignore nulls for this very reason.. but every Room correction screenshot I've seen seems to be devoid of nulls. Is this part of the magic of the algorithms?

My last question is more of a convenience/usability issue. Can Audiolense output to a different card than the one that is recording? My current setup has an Nvidia graphics card connected to my receiver via HDMI. The only recording interface I have is a TASCAM USB sound card which is only 2 channel. I'm assuming I'll probably have to use the TASCAM to measure two channels at a time by swapping speaker cables around on the AVR. Is there a better way I'm not thinking of?

Thank you both again for your time. I think I'm one step close to joining the world of digital room correction!

To unsubscribe, send email to audiolense+unsubscribe@googlegroups.com

Brad

unread,

Aug 6, 2012, 9:54:55 PM8/6/12

to audio...@googlegroups.com

Read this series by Mitch, a great Audiolense story.

Part 1

Mitch Global

unread,

Aug 7, 2012, 2:31:19 AM8/7/12

to audio...@googlegroups.com

Thanks Brad!

I recently wrote another piece that may be of assistance. I included binaural recordings of the before and after Audiolense. The only thing I am not happy with in those recordings is the top end is rolled-off too much in the recordings compared to listening to the speakers live. I think I have that figured now and will be posting an update soon. No fault of Audiolense, just my binaural recording. However, there is no question that you can hear how effective Audiolense is. You can hear the difference for the better:

http://www.computeraudiophile.com/blogs/mitchco/importance-timbre-sound-reproduction-systems-222/

I hope Bernt does not mind me posting a measurement here... Attached is a waterfall graph using REW that shows an overlay of before (purple) and after (blue) DRC in the critical low frequency range of 20 Hz to 400 Hz (i.e. room modes and diffusion).

For those new to a waterfall graph, the vertical scale is amplitude in dB, the horizontal scale is frequency in Hz., and the Z scale is time in milliseconds. In this case, I have measured 300 milliseconds of room decay time. Given that sound roughly travels roughly a foot per millisecond, the waterfall view is the first 300 feet of travel after the direct sound hits the mic.

The blue is with DRC applied and the purple is the overlay of my room before DRC was applied.

Audiolense generated filter has reduced the purple you see that is visible, to be in line with the rest of the blue. That is is significant reduction. Note how the blue fills in the frequency response where there were nulls and reduces those massive peaks just over 100 Hz (room cutoff frequency) and wicked harmonic at ~200 Hz. Also note I was able to extend the frequency response of my 15" woofers down another octave without overloading them or the amplifiers.

Since I have the XO version, I love TTD. If you look at the overlay waterall graph again, you will notice the Audiolense filter correcting in the time domain as well, filling in where needed, and reducing those long ringing room modes. Proof positive that Audiolense works equally well both in the frequency and time domains. Thanks Bernt!

Attached are the waterfall measurements of my listening room before and after Audiolense, so you can see how the overlay works.

That was with a 500 millisecond TTD correction window @ 10 Hz. Based on Bernt’s advice in the release notes of 4.4 Beta, was to try shorter time windows. I went from 10 milliseconds to 100 ms., in 10 ms increments. As it turns out, 40 milliseconds was the magic number that sounded the best in my particular room (I use linear phase target). Experimentation is encouraged. I plan to put up another post at CA soon about this.

Hope that helps.

Cheers, Mitch

--

Audiolense User Forum.
http://groups.google.com/group/audiolense?hl=en?hl=en
To post to this group, send email to audio...@googlegroups.com

To unsubscribe, send email to audiolense+...@googlegroups.com

Audiolense Overlay.JPG

After Audiolense.JPG

Before Audiolense.JPG

Mikkel Gylling Hangaard

unread,

Aug 7, 2012, 3:06:51 AM8/7/12

to audio...@googlegroups.com

Hello Mitch,

Your blog was a bit of an inspiration to me. I hope you don't mind a few questions to your last post, which I believe also would be interesting for Scott to hear about:

1. What perceived difference do you hear between linear and minimum phase filters?

2. How did the sound change when you went from 500ms to 10ms and then 40 ms? Any benefits or downsides noticed?

Thanks for your time and your blog entries. They make for excellent reading.

Best regards,

Mikkel

2012/8/7 Mitch Global <mitch...@gmail.com>

--

Venlig hilsen,

Mikkel Gylling Hangaard

______________________________________

Ph.d.-studerende

Cand.mag. medievidenskab
BA medievidenskab og politisk teori & filosofi

Tlf.: 3121 0418

E-mail: mikkel.gyll...@gmail.com

Mitch Global

unread,

Aug 7, 2012, 4:01:09 PM8/7/12

to audio...@googlegroups.com

Hi Mikkel,

Thanks!

re your questions. Well, that is the topic of one of my upcoming posts... In which I will also have binaural recordings so folks can hear the difference. To qualify that, on my system there is an audible difference between linear and min phase targets.

But you can try this yourself on your system in about 5 minutes. Just make a copy of your current min phase target and change it to linear phase and regen the flters. I use JRiver's most excellent Convolver and can easily switch between the two filters with only a slight silence gap in the sound (milliseconds), so makes for a good AB test.

Same goes for the TTD correction procedure designer and playing with the TTD subwindow settings. Try it. Depending on your acoustic environment and/or preferences, you will be able to dial in a setting that should make the sound totally coherent from top to bottom and anywhere in the room. I found in my acoustic environment that too long of a window in the low end, sounds a bit disconnected from the mids and highs.

Bernt's 4.4 Beta release notes are excellent and worth reviewing again. Bernt's best advice to me was experiment and that I did and it has paid off well. I encourage you and others to experiement as well.

Sorry if I did not answer your q's explicitly as I am holding off for another blog post. But again, it only takes a few minutes of your time to try AB'ing linear versus min and in my scenario, I liked the sound of linear better. Playing with the TTD subwindow in the low end may take a bit longer, but worth it.

Cheers,

Mitch

Bernt Ronningsbakk

unread,

Aug 8, 2012, 6:15:31 PM8/8/12

to audio...@googlegroups.com

Hi Scott,

I did some load simulations of how much a correction stresses the stereo system. The conclusion was that that a frequency correction with 6 dB boost produced the same overall output SPL for the same input SPL as no correction. A default TTD correction had a penalty of 1 dB. That’s not a lot if you ask me.

This was on a speaker that overall had enough capacity on all drivers to do their part of the job, but where there were normal fluctuations of plus minus 12 dB give or take.

There will be situations where the music hits hard on the frequencies that has been boosted, but there will also be situations where the music hits hard on the frequencies that have been taken down a few dB, so the load change because of correction is very dependent on the situation. But the same goes for different types of music without correction. Different types of music will load the hifi system different.

Based on experience I would advise to be careful about stressing the lowest frequencies of the low frequency driver. But a substantial lift even at very low frequencies will work if the driver is capable of moving enough air to get the job done. And if you are willing to trade e.g. 10dB SPL against an extra octave in the bass you can boost the bass an extra octave.

Apart from boosting the lowest frequencies it is my experience that the added load from correction is negligible most of the time. And the simulations I did basically confirms that. A correction with substantial dip lifting may cost you gain. The ideal situation is to have at least 10dB more gain than you need for uncorrected music at high spl levels. If your system maxes it’s useable SPL before 12 o’clock on the volume control before correction you should be fine.

Second question: During measurement, the mic feed and the speaker out should ideally be run through the same sound card. Audiolense allows for using one audio device for playback and another for recording, but this will sometimes lead to timing errors in the measurement and a compromised TTD correction. A pure frequency correction is far less sensitive to timing errors and using two audio devices during measurement is likely to give a pretty good result…

Usually it works very well to use one audio device for the measurement session and another device for listening to music, even with TTD correction. That’s quite normal among Squeezebox users, for instance.

Kind regards,

Bernt

From: audio...@googlegroups.com [mailto:audio...@googlegroups.com] On Behalf Of Scott
Sent: Tuesday, August 07, 2012 2:55 AM
To: audio...@googlegroups.com
Subject: Re: [audiolense] Re: Considering Audiolense

Hi Bernt,

Thank you for the very informative reply. Your observations of Audyssey seem to be on par with what I have measured. It seems to do a reasonable job of keeping frequency response relatively flat in the upper ranges, but struggles a bit on the low end. Relatedly, how does Audiolense allow such high correction levels and achieve such flat response? I have read posts stating 12+ db correction. My very limited reading has suggested that minor boosts can take significantly more amplifier power. In the Parametric EQ school of thought, it seems they ignore nulls for this very reason.. but every Room correction screenshot I've seen seems to be devoid of nulls. Is this part of the magic of the algorithms?

My last question is more of a convenience/usability issue. Can Audiolense output to a different card than the one that is recording? My current setup has an Nvidia graphics card connected to my receiver via HDMI. The only recording interface I have is a TASCAM USB sound card which is only 2 channel. I'm assuming I'll probably have to use the TASCAM to measure two channels at a time by swapping speaker cables around on the AVR. Is there a better way I'm not thinking of?

Thank you both again for your time. I think I'm one step close to joining the world of digital room correction!

On Monday, August 6, 2012 2:06:44 AM UTC-10, BerntR wrote: