Jump to content

Recommended Posts

Posted

Finally, Arnaud and I are still trying to sort out Tyll's data - to see if it's clean or good enough to crank out meaningful measurements. So it's still too early to give you an executive summary at this time.

I'll give you an executive summary, Al: Wooooot! We're having some fun now.

I screwed up on the day to call Arnaud, but I'm going to look at my settings after I finish a post I'm working on. I think I'm beginning to get what you guys are talking about. I'll see if I can take some screenshots of the panel in which I can make the various settings. It won't be hard to change them from now on ... it'll be pretty much impossible to go back and remeasure all the cans, but I certainly might go back and do some of the important ones again. Fortunately, even the stuff you guys are potting with current data is usable, it seems to me. Not ideal, but usable.

Anyway, I'll post back later today.

Posted

Okie dokie, here's what my screen looks like when I do the Impulse Response test:

110903_Current_screen_shot.jpg

The analog generator is outputting the Maximum Length Sequence (MLS).

It's using a 32kByte long sequence. It's using a output sample rate DAC of 65536.

The digital analyzer is using the "Quasi anechoic" mode, which is used with the MLS signal.

The ADC is set here to match the analog generator at 65536.

The picture shows 2 averages, but normally I do 16.

A raw .xls file for data from the above test is here.

Here are a variety of pix showing the options available on the various menues:

Output DAC sample rate settings:

110903_generator_options.jpg

The various type of digital analyzer modes available --- Quasianechoic is used typically for MLS:

110903_analyzer_types.jpg

Input sampling rate settings on analyzer:

110903_analyzer_input.jpg

Analizer Domain setting:

110903_analyzer_domain.jpg

Energy/time window settings:

110903_analyzer_energytime.jpg

Data display method:

110903_analyzer_method.jpg

I tried using the longer MLS

Posted (edited)

So I did another test with the longer MLS word (128K) and the next highest sampling rate:

110903_test_131072DACADCsamplerates.jpg

Speadsheet with data for this test is here.

THis shit is pretty complicated, so I figure you might need to look at the manual . Chapter 13 is about MLS and the impulse response etc.

Alrighty, I'm listening for what to do next.

Edited by Tyll Hertsens
Posted

Hi guys,

Been a very rough week. Supervisor decided to take three weeks off on vacation, so I've been stuck trying to resolve things that are a bit out of my scope of knowledge. But I've managed!

Last weekend, my brother's father-in-law was in the hospital. He had an epileptic attack at work and banged his head pretty hard on the floor. Mid-week, a close family friend had an aneurysm discovered and underwent surgery with 35% survival rate. Then on Thursday, my dad had to be taken to ER due to some very intense pains in his abdomen. Diagnosis? Huge kidney stones. He underwent surgery yesterday. It's been nothing but work and hospital "vigils" all week.

Good news, though, is that everyone is okay now.

Better news, I apologize for the delay in sending out the gear, but woke up early to do just that this morning before heading out to the hospital again.

Tyll's address actually made me chuckle for a good while: BOZEMAN, MT.

Of all things, Bozeman? As in Bose man? HAHAHAHA

Anyhow, Tyll, I sent the amp over in one box and the O2's in another.

The tube boxes are labeled in the top inside flap for the positioning on the amp (I guess they are biased for the particular socket?). There is a yellow DHL bag by the PSU section, don't throw that away. It isn't just padding, it also has the spiked feet for the PSU. The spiked feet for the amp are inside the small white box with the tubes. Umbilical cord is wrapped in some bubble wrap. I didn't send interconnects or power cord as I believe you wanted to supply your own for testing purposes. I also threw in a microfiber cloth for wiping the amp when you need to. You can keep the towel, I have plenty of those. I insured the shipment for the full $5k, just in case. Sent USPS Priority Mail, which means delivery should be Tuesday or Wednesday, I'm guessing.

The O2's were sent in their original flight case or whatever you call it, so they are well protected.

I hope you have fun doing all those tests!

By the way, I also included some sticky labels inside a folder with the BHSE amp with your addess as Sender and mine as Recipient, so you can just stick that over the old one when you ship back. I included a couple so you can use it for the O2 cans as well.

I'll pm you the tracking numbers.

Posted

Are IMD measurements in the road map?

THE BAS SPEAKER

THE PUBLICATION OF THE BOSTON AUDIO SOCIETY

November 2002 Meeting

Headphone Test Clinic

by Alvin M. Foster (with Jim Doucas and David B. Hadaway)

(...)

Intermodulation Distortion (IMD)

We settled on the traditional test: 60Hz plus 7kHz mixed 4:1 (the

bass tone is 12dB higher than, or over twice as loud as, the treble

tone).

The clinic was nontraditional in performing this test in the first

place as well as in the level used. We chose a playback level for all

the headphones of 100dBspl, and to divide the good from the bad

further, we set an arbitrary cutoff point of 10% maximum distortion as

we raised the input. When the 10%-distortion figure was reached, we

recorded the playback level and discontinued the test [at such levels,

amplifier output capabilities into very low-impedance or low-efficiency

designs could also have worsened the results significantly — AES].

Attempts to use multi-tone test signals to separate out the units

capable of low distortion and high playback levels were dropped

because the results were too qualitative, in that we could not get a

single number to represent performance.

(...)

BAS_24_04.jpg

http://www.bostonaudiosociety.org/pdf/bass/BASS-25-4.pdf

Posted (edited)

Edit: updated graphs, correct a problem with the time scale...

Tyll, I will have a look at your new data and settings. Meanwhile, I could remove the issue (lack of data points) with the current data by automatically zero padding up to 2048 samples and doing the FFTs on that "augmented" data. It does not imagine some new decay beyond 3ms but the low frequency results (below 2kHz) are smoother...

CSD_forInnerFidelity_110904R3_Tyll_AKGK701_LeftEar_3D.jpg

CSD_forInnerFidelity_110904R3_Tyll_AKGK701_RightEar_3D.jpg

CSD_forInnerFidelity_110904R3_Tyll_AKGK701_LeftEar_2D.jpg

CSD_forInnerFidelity_110904R3_Tyll_AKGK701_RightEar_2D.jpg

Edited by arnaud
Posted (edited)

Had only a very brief look but did not see much difference between your 2 files. I'll have a go at the manual.

Meanwhile, how about changing the settings in the "sweep >>Data3 >" window?

For instance: stop=6ms, Steps=2047.

Or, if you need to specify the time step rather than number of steps, that would be 0.006/2047=2.92969E-06

So I did another test with the longer MLS word (128K) and the next highest sampling rate:

110903_test_131072DACADCsamplerates.jpg

Speadsheet with data for this test is here.

THis shit is pretty complicated, so I figure you might need to look at the manual . Chapter 13 is about MLS and the impulse response etc.

Alrighty, I'm listening for what to do next.

Edited by arnaud
Posted (edited)

Not one, but two SR-009s?

Tyll's cup runneth over!

Time for some lambda and OII mk1s to materialize and keep up the momentum.

I sent him my OII's with my BHSE. I believe mine are MK1. Well, box says 007A, but the cans are brown leather/cable and champagne housing. Tyll will be able to compare them both side-by-side,

It will be VERY interesting to see how wide the gap is between the two.

Edited by screaming oranges
Posted

That's dedication, many thanks for sending your gear! And more importantly, all the best to your family.

I sent him my OII's with my BHSE. I believe mine are MK1. Well, box says 007A, but the cans are brown leather/cable and champagne housing. Tyll will be able to compare them both side-by-side,

It will be VERY interesting to see how wide the gap is between the two.

Posted

A bit off-topic, but I think it'd be nice if someone can arrange (when the time comes) a sort of donation collection so that Tyll doesn't have to shell out a bunch of cash to send all this stuff back. Just a thought (and unfortunately, no, I don't have time to arrange it).

Posted
so that Tyll doesn't have to shell out a bunch of cash to send all this stuff back.

I just want to say that my boss is totally down with what I'm doing, and down with the arraingements as is: you get it to me, I'll get it back to you in the US.

So thanks for the thought, and if anything we should collect for your cluster fuck, but Im good with working together to get these cans measured and paying return shipping.

Posted

I don't mind chipping in for shipping fees. Actually, people sending your equipment to Tyll, you're paying for shipping one way, which gets pretty expensive for something like a BHSE no??

Tyll: I reprocessed Purrin's data with the same Excel template in order to compare your and his measurement of the HF2.

The resemblance is pretty weak which could be due to:

  • a blunder in my processing (I get similar results to Purrin's CSD + my FFT at t=0 matches your original spreadsheet result though)
  • the effect of his measurement method (not traditional dummy head)
  • some issue with your original data (I don't understand why it shows up as 170kHz sample when your settings show it was actually sampled at 130kHz)
  • a combination of all the above!!

CSD_MarvData_110904R1_HD800_HF2_HP2_3D.jpg

CSD_InnerFidelityData_110904R4_Tyll_GradoHF2_LeftEar_3D.jpg

CSD_InnerFidelityData_110904R4_Tyll_GradoHF2_RightEar_3D.jpg

CSD_MarvData_110904R1_HD800_HF2_HP2_2D.jpg

CSD_InnerFidelityData_110904R4_Tyll_GradoHF2_LeftEar_2D.jpg

CSD_InnerFidelityData_110904R4_Tyll_GradoHF2_RightEar_2D.jpg

Posted

Have you had a chance to look at the data with higher resolution I made? I'm wondering if the setting are really doing something in terms of resolution.

I haven't processed the data but will check later. Meanwhile, per your last PM, I changed the batch routine to also copy the graphs to the original data (new CSD sheet after square wave graphs, graphs pasted as enhanced metafile). It saves the data file as new name with _wCSD suffix just to be safe (file grows from 2MB to 4MB due to the 4 additional graphs).

Illustration:

Tyll3_GraphCopy.jpg

Tyll3_GraphCopy_FolderView.jpg

Posted

Have you had a chance to look at the data with higher resolution I made? I'm wondering if the setting are really doing something in terms of resolution.

I processed the files now and as expected there's little difference. I did not yet check the manual but I think it has to do with the randomness of the MLS sequence but it doesn't actually change the recording settings. e.g. there's still only 3ms of data and the apparent sample rate remains 170kHz something.

CSD_InnerFidelityData_110904R5_Tyll_Standard_LeftEar_3D.jpg

CSD_InnerFidelityData_110904R5_Tyll_MLS128k_LeftEar_3D.jpg

CSD_InnerFidelityData_110904R5_Tyll_Standard_LeftEar_2D.jpg

CSD_InnerFidelityData_110904R5_Tyll_MLS128k_LeftEar_2D.jpg

CSD_InnerFidelityData_110904R5_Tyll_Standard_RightEar_3D.jpg

CSD_InnerFidelityData_110904R5_Tyll_MLS128k_RightEar_3D.jpg

CSD_InnerFidelityData_110904R5_Tyll_Standard_RightEar_2D.jpg

CSD_InnerFidelityData_110904R5_Tyll_MLS128k_RightEar_2D.jpg

Posted

Hm. So it looks like the current spreadsheets have reasonably good data, or at least data thats as good as I can get it. I could take data over a longer period of time if that would help by setting the sweep to say 5 or 10 mSec, but there's not much going on out there signalwize. Would that be helpful to look at?

Posted

Maybe the window for reading response is set to follow the stimulus duration. That would explain the short decays. No idea where to set those separately or if they're linked and you indeed need to make the sweep 5-6 ms long.

Posted

Hi guys! Just wanted to tell you that I've decided I need to get some helmet time, so I'm off on my motorbike for a week! Wooot!

I'll be back Sept 12. My wife will be here to receive boxes, so don't worry about that.

Thanks for all the help on this stuff!

Posted

I processed the files now and as expected there's little difference. I did not yet check the manual but I think it has to do with the randomness of the MLS sequence but it doesn't actually change the recording settings. e.g. there's still only 3ms of data and the apparent sample rate remains 170kHz something.

Nice result. Looks like there's some correlation. Let me know if you need any raw impulse response files of other headphones from me for comparison / tweaking Tyll's settings.

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use.