ARC - Help my devise a test.......

[dankelly]

Hey guys, I think this is the correct forum for this thread. Although it's a bit of a techy question, there has been plenty of previous chat, albeit quite heated, in relation to the use/usefulness of IK's ARC.

I am currently studying at SAE and for my 3000 word essay I have chosen to write about ARC. My essay is titled "Digital room correction: A viable solution?", and I have chosen ARC as the platform which I am going to test.

The problem I'm having is coming up with a way of testing the software fairly. Firstly, why does the supplied microphone have to be used instead of another omni measurement mic, such as the Behringer 8000? I also need to find a way of measuring the rooms response after the ARC plugin with 3rd party measuring software such as REW. Because it is used as a plugin, I cannot think of how I can then send the "corrected" audio from my DAW into REW.
I know Ethan has ran tests on Audyssey which is a hardware device and would not have this problem. I'd be super grateful if anyone can think of a workaround for this. Cheers guys
-

[Jens Eklund]

Use a soundcard with a digital input. Use a second computer (and a soundcard with digital out), running REW and feed the output from REW through the DAW (using ARC) and record with and without ARC.

I hope you don’t limit yourself to analyzing only the frequency response. But also the things these system can’t fix; time dependent stuff like decay time and reflections (waterfalls and ETCs)

I assume that you already know about these threads:
KRK Systems - ERGO Room Correction
IK Multimedia ARC System vs Acoustic Treatment???

[Audioactive]

Originally Posted by dankelly

Because it is used as a plugin, I cannot think of how I can then send the "corrected" audio from my DAW into REW.

-

What i would do is to route the Audio interface output into its input and record the ARC result, playing a sine sweep or a noise.

then i would play the result with REW and take measurment.

but room correction is more than only adjusting frequency response.

[dankelly]

Thanks for the replies, super quick!

FYI, I intend on testing frequency response, ETC and decay time.

Jens, does this sound ok:

Macbook pro running REW, headphone out -> 3.5mm jack lead ->
Line in on macbook number 2, setup logic to use built in input ->
Out of logic via ARC through Duet to monitors ?

Or I could substitute the 3.5mm jack for a 3.5mm to 2 x 1/4" and input into the duet instead of the built in input on macbook number 2?

For a test like this can we assume that the mac built in i/o is good enough quality as i only have my duet anyway?

Audioactive, thats a good idea also. So in REW I can use a pre recorded sweep for analysis rather than the self generated one?

[mhch]

My 2 cents ...

I tried the ARC system a couple of years ago, using its automatic setup feature.

ARC showed I had a big and large peak in the 200Hz region and I ended up very surprised from the bass improvement. Upright bass parts I created using sampled sound libraries sounded much neater and clearer when using it.

The real huge benefit is that ARC triggered my learning journey of how much room treatment is important, and thanks to this forum understand modes, comb filtering, ETC, etc ... and be able to build a much better environment for my music activities.

[Sam Balaam]

Hi Dan,

I did a similar experiment not long ago, I used Fuzzmeasure pro 3, Logic Pro, a 8 in 8 out soundcard, mackie big knob and a behringer ECM 8000.

Fuzzmeasure Pro has its own sweeps and you can choose the inputs and outputs it uses. so i had its output routed to an output of the soundcard and then rerouted to one of its inputs, activated this in a channel in Logic (which you could put the ARC on) So at this point you are getting the sine sweep from fuzzmeasure into your logic channel.

The main outs of Logic were then assigned to the main outs of the soundcard which were then routed to the monitors. Fuzzmeasures sweep is now being processed by the ARC and played out of the monitors.

The 8000 was routed into input 2 of the soundcard and then activated in Fuzzmeasure, so running independently of Logic but still receiving the processed sound.

You can get a free demo of Fuzzmeasure Pro and it does really good comparisons.

Pretty sure this is what i did, hopefully this isn't too babbling but it worked for me.

Best,

Sam.

[dankelly]

Hey Sam!

Thanks alot for that great reply. I understood fully what u said and that is definitely an option aswell. My only issue is that I only have an apogee Duet so don't have the I/O for your method unfortunately.

Sam, I'm interested to hear about your experiment and the conclusion you came to regarding arc. Care to share?

[Sam Balaam]

Well basically i tried to blag a free version from IK Multimedia but they weren't having any of it and my department couldn't afford to buy it so i ended up using a series of E.Q's in logic to try and replicate what the ARC system does.

Obviously not a fair test as ARC works in a different way to just notch filtering but its all i had really. After seven hours, around 15 channel EQ's and a slight case of tinnitus later i came up with this...



Sorry for the low res, but red line is before, blue after.

It can be done but my conclusion is that a combination of both physical and digital treatment is the way forward. Output calibration, as far as i am concerned, is a good way to fine tune your listening position but not an overall solution.

Plus building panel absorbers, bass traps and clouds is so much more rewarding than obsessively tweaking digital EQ's.

Hope this helps.

Sam.

[Ethan Winer]

Originally Posted by dankelly

The problem I'm having is coming up with a way of testing the software fairly.

Jens has the right idea. But I'd use two entirely separate systems - one that hosts ARC and was set up using their microphone, and another with a separate microphone running REW. Maybe you or a friend have a laptop? Once ARC is calibrated, play REW's sweep out through the main computer system through ARC, but record into REW running on the laptop. I guess you'll need to send the laptop's output to the other system in a Through mode to play it through ARC.

Also, be sure to put your omni microphone very close to ARC's mic. The photo below shows how I did that when testing the Audyssey system.

--Ethan

________________
The Acoustic Treatment Experts

[dankelly]

I would have to have the laptop running REW outputted to the 2nd laptop, and on 2nd laptop I would have to setup logic with an audio channel fed from the line in so the sweep woul run through ARC on logics master?

Ethan, why is it necessary to have the 2 mics setup? Can I not just setup and calibrate using arc mic, see what results that claims to give, then test with REW and Behringer 8000, both with ARC enabled and disabled to compare? I only have 1 audio interface and I would need one for each computer using this method?

[dankelly]

FWIW, my room is already treated as far as it can be. Superchunks in all four corners and cathedral ceiling, first reflection points, 8" mobile bass trap for between me and back wall.....

[Ethan Winer]

Sure, I just meant that the microphone you use should be in the same place as ARC's microphone when ARC was calibrated. Which brings up a good point. Be sure to also measure at other locations around the room, both near and far. You don't have to measure 10 feet away, but see what happens 6 inches away, the space between our ears. And try 2 feet away, where someone else might be sitting.

--Ethan

[dankelly]

Ok, I'll do that. I'll be sure to post the results up here when Im done too. Thanks everyone

[DanDan]

Sam, which worked better your red before or blue after response?
In my experience a flat line response is highly undesireable. See Understanding RTA at studiotips.com or http://www.bksv.com/doc/17-197.pdf

In both cases the response is up 3dB around 100Hz and below, tilting to minus 3dB around 10K and up. This slope works well.
DD

[Sam Balaam]

Originally Posted by DanDan

Sam, which worked better your red before or blue after response?
In my experience a flat line response is highly undesireable. See Understanding RTA at studiotips.com or http://www.bksv.com/doc/17-197.pdf

In both cases the response is up 3dB around 100Hz and below, tilting to minus 3dB around 10K and up. This slope works well.
DD

Hi Dan,

Thanks for the tip, i haven't read the link yet but i will. My aim was to achieve as flat a response as possible so blue wins but is obviously flawed as only relative to one position. You get some strange drones and resonances in the lower frequencies as you slap those notch filters in also.

This was my first experience of testing acoustics and has made me rethink my approach and encouraged me to strip the room down and acoustically treat from the base up.

Forgive naivety but is a flat response not what we are going for in a mix room?

Sam

[Jens Eklund]

Originally Posted by Sam Balaam

Forgive naivety but is a flat response not what we are going for in a mix room?

http://www.gearslutz.com/board/6633549-post38.html

[DanDan]

Forgive naivety but is a flat response not what we are going for in a mix room?

Not at all. Flat response will lead to very dull mixes. Bruel and Kjaer used to sell a Test Record LP containing third octaves of pink noise.
Using it and an SLM and a third octave graphic they recommended adjusting the response exactly as described. +3 to -3. Quite a tilt.
It works.
DD

[dankelly]

Interesting stuff.

Does anyone know what the ARC is supposed to "correct" to, flat or to DanDan's suggested +3/-3 dB?

[OliverV]

Originally Posted by dankelly

Interesting stuff.

Does anyone know what the ARC is supposed to "correct" to, flat or to DanDan's suggested +3/-3 dB?

They don't publish the specifics but there are a few options. From the manual:
To cover a wide range of room types, ARC System provides a selection of four Target Curves created by Audyssey. Click on this menu (figure 5.5) to select one of these four Target Curves:
• AUDYSSEY FLAT
• AUDYSSEY HF Roll-off 1
• AUDYSSEY FLAT (Mid Comp)
• AUDYSSEY HF Roll-off 1 (Mid Comp)
Once your measurement is done, listen to material that you’re familiar with, using the ARC Plug-In on your DAW Stereo Master BUS.
To start, select the FLAT target curve and listen carefully to the speakers system.
AUDYSSEY FLAT: the system will correct the frequency response so that the final response will be very close to flat. However, the system will not extend the low frequency response of the speakers beyond their limits to avoid possible damages and excessive distortion. In fact, the AUDYSSEY Flat curve corrects the response to flat from the low frequency cutoff point that the ARC System has determined, to the upper frequency limit of the tweeters.
AUDYSSEY HF Roll-off 1: this should be selected if you feel that FLAT sounds excessively bright in your room. In fact, the AUDYSSEY High Frequency Roll-off 1 curve introduces a slight roll-off at high frequencies that accounts for the balance between direct and reflected sound for small to medium size rooms.
Both AUDYSSEY FLAT and AUDYSSEY HF Roll-off 1 target curves are also available with the Midrange Compensation, "Mid Comp", option.
AUDYSSEY MIDRANGE COMPENSATION (Mid Comp) is sometimes necessary to correct the directivity differences that often occur in that frequency range due to the crossover circuitry or horn-loaded speakers. Start listening to the corrected systems without Mid Comp, allow yourself to get used to the "new sound" of your studio, and then try the Mid Comp option. Listen to your favorite music, and judge which AUDYSSEY Target Curve sounds more natural on your speakers and room environment.

[perform/record]

Originally Posted by Sam Balaam

Well basically i tried to blag a free version from IK Multimedia but they weren't having any of it and my department couldn't afford to buy it so i ended up using a series of E.Q's in logic to try and replicate what the ARC system does.

Obviously not a fair test as ARC works in a different way to just notch filtering but its all i had really. After seven hours, around 15 channel EQ's and a slight case of tinnitus later i came up with this...

Attachment 239182

Sorry for the low res, but red line is before, blue after.

It can be done but my conclusion is that a combination of both physical and digital treatment is the way forward. Output calibration, as far as i am concerned, is a good way to fine tune your listening position but not an overall solution.

Plus building panel absorbers, bass traps and clouds is so much more rewarding than obsessively tweaking digital EQ's.

Hope this helps.

Sam.

what does blag mean?

[dankelly]

Hey guys.

Well I carried out my tests and I have results for you to make of what you will. I cant upload the rew file as its nearly 50mb with all the measurements but i will do them individually if anyone wants me to.

I am only posting the images of the frequency response and waterfalls with and without ARC at the listening position but i have results for locations 300mm in every direction from the LP

The ETC graphs show zero improvement, surprise surprise. The frequency response plots are quite alot better though, and the biggest surprise was that the decay times of some modes on the waterfall graphs HAVE been reduced! I dont know how but thats what the graphs show.

As i said previously, my room has extensive treatment already, so you couldnt expect results like this in an empty room.

I would love some opinions on my findings, and also on my room. Do you guys think ive done ok in this space? It is only 4.4m x 2.4m x 2.4m


perform/record, blag means bullshit basically. Sam Balaam tried to bullshit IK multimedia into giving him a free copy, I would have tried this myself had i not read of his failed attempt

[DanDan]

Hi Dan. The peak at 150 ish is diminished in level usefully.
I am not so sure that the mode is actually shortened.
Try dropping the 'floor', the min magnitude of the waterfall, by the same amount the freq response peak is reduced. Decay is over a span of levels. Like with like. You could also look at the third octave decays. The 400Hz hints that you may have been close to the measuring mic. The HF comb filtering suggests two speakers were driven. All too common errors.
http://www.gearslutz.com/board/studi...er-v2-1-a.html

DD

[dankelly]

Hey Dan.

I have included the waterfalls with the lower range set to 30dB, 15dB lower than before and I still think there is a definite improvement.

Both speakers are driven as i wanted to drive the modes the way that they will be driven under normal circumstances. Im mainly concerned with the modal region TBH, although I will run tests with the speakers driven individually if you think it is required to validate the test results.

I was about 1-1.5m away from the mic, do you still think thats the reason for the 400Hz null?

I have included the rew files for the mix position aswell as in front, behind and either side so you can delve deeper. Had to zip them and do it in three files due to size/compatibility though, sorry.

[DanDan]

The mode does seem a lot better, which shouldn't be happening.
REW on Mac doesn't open files. Could you export the IR as a WAV instead?
Did you look at third octave decays?
The 400Hz area can be affected by a body near the mic, however seems like you have that in hand. Try moving further away, or look for some other reflection.
DD

[RyanC]

There was a big debate about this a while ago on a speaker building forum that I was following. The upshot was that modes DO get attenuated in terms of the length of the ringing by cutting them in an EQ, and some much smarter then me were able to demonstrate the math as to why.

That said the nature of that can be relative to location in the room.

[Jeffrey Hedback]

DanKelly,

Looking very nice and I'll bet sounding smooth and tight.

here are my experiences with ARC and the JBL option:
- ARC/Audyssey tends to make poor choices by "correcting" non-minimum events. This results in flatter but more resonant responses. I've worked on two rooms where the ARC/Audyssey correction was exactly as just described. One client saw the truth and disabled while the other preferred the prettier freq response graph.
- However, the JBL DRC system seems to make very good choices...correcting minimum phase only events. And sounds good.

Once the room is treated to the point of diminishing returns and spkr/ear locations are optimized the same...DRC or properly applied PEQ is a very cool step.

No matter what anyone says, you still can't EQ a reflection.

I fully predict an avalanche of MUCH more evolved DRC systems in the next 2-3 years.

Back to DanKelly, how does it sound with ARC?

[DanDan]

RyanC, can you find that math? I am intrigued now. Dan's graph shows a serious shortening of the mode, and I can think of no theory to explain why.
Why would something resonant have different lengths at different levels.
Dan, please do post a WAV, just one would be fine.
Good to hear some detail on how the various room correction systems work internally and in practice, thank Jeff.
I saw how the Sintefex Replicator led to the Liquid Channel and LiquidMix.
Similarly Altiverb.
I look forward to sophisticated affordable room correction.

DD

[dankelly]

Jeffrey, the room sounds better with ARC TBH. When I say better, I mean more comfortable, not necessarily more accurate. I haven't managed to test it with reference material yet, only my own productions.

Ive been rushing to carry out these tests for my SAE assignment and so haven't fully got my head around it yet. What I will say though, is that since finishing my acoustic treatment i have found the room to be a little fatiguing (is this common?). This has seemed to be less so with ARC enabled. I dont know the cause of the fatigue, and i intended on investigating it further when i had time. I must admit though that i probably listen a little louder than i should sometimes which may not help.The low end also seems more rounded, clear and full with ARC enabled.

Im a super sceptic, and i had basically written ARC off before i had tested it. Ive read all the threads on this thing and I decided to use it as my essay subject as i 'already knew' what the outcome would be, and it would therefore be a simple study. In fact, parting with the £200 was the hardest part, however, you cannot deny these graphs.

Jeffrey, could you explain a little further the correction of non-minimum events please? Im not clued up on this at all.

Dan, do you have REW v5? I am on a mac and I have no trouble opening those files. I can upload a WAV no problem, just tell me how to do it

[RyanC]

I've been looking all over for it....I think it was all on DIY audio in some 50+ pager!!!

Anyone hear knows, SAC maybe? Something along the lines of the same way that boosting EQ creates ringing, cutting it attenuates it. I'm way out of my element here...

I do remember a test showing this with EQ's and REW, if you boost narrow q at 100hz you can measure the ringing, if you cut by the same amount/Q you can measure the ringing going away again. (two bands, one boosting, one cutting). And the fundamental argument being that modal ringing is mathematically the same as EQ ringing. Then a demonstration showing the waterfalls like the ones we see here.

Of course with acoustics that modal resonance could be completely different an inch to the left or to the right which would make it rather pointless for room tuning if that is the case with a specific room.

Just to make it clear I'm squarely in the camp of treat, treat, treat, then use corrective eq to fix issues that are consistant trends in the listening area.

[dankelly]

Originally Posted by RyanC

Of course with acoustics that modal resonance could be completely different an inch to the left or to the right which would make it rather pointless for room tuning if that is the case with a specific room.

I have included measurements from either side of the mix position, aswell as in front and behind. The results are not quite as good, but definitely still an improvement.

Just to make it clear I'm squarely in the camp of treat, treat, treat, then use corrective eq to fix issues that are consistant trends in the listening area.

Me too