Audio Engineering Society 2023 Conference

[Beershaun]

Congratulations and best of luck with your presentation!!

 

[Phorize]

Looks like AES are yet to adopt the 'no politics or religion' approach of ASR. The immersive microphone technique session looks fascinating.

 

[192kbps]

Preference ratings and 32 magnitude frequency response curves

This repository comprises two CSV files: PreferenceRatings and MagnitudeFrequencyResponses. The former includes preference ratings obtained from 56 naive assessors (30 Danish---DK, 26 Japanese---JP) of 32 headphone curves over several music programs in several trials. The latter includes the...

zenodo.org

 

[IAtaman]

Dr Olive's seen it:

View attachment 317990

Danes ans Japanese agreeing on a preference curve is not a big deal. Get Danes and Swedes agree to it and I will call that universal.

 

[alumnicesar]

Preference ratings and 32 magnitude frequency response curves

This repository comprises two CSV files: PreferenceRatings and MagnitudeFrequencyResponses. The former includes preference ratings obtained from 56 naive assessors (30 Danish---DK, 26 Japanese---JP) of 32 headphone curves over several music programs in several trials. The latter includes the...

zenodo.org

View attachment 318109

Soundguys wasnt much worse than harman, but why are they all low?

And im guessing df isn't df + a tilt?

 

[GXAlan]

Danes and Japanese agreeing on a preference curve is not a big deal. Get Danes and Swedes agree to it and I will call that universal.

A cross-cultural comparison of salient perceptual characteristics of height channels for a virtual auditory environment - Virtual Reality

Perceptual characteristics of virtual auditory environments from three listener groups were compared. To generate convincing and pleasing virtual auditory environments, acoustic impulse responses were measured in two venues using an innovative microphone array and convolved with two anechoic...

link.springer.com

I also think it is important to talk about preferences with mono speakers versus other topologies where there is data to show cultural differences.

The fact that headphones are presumably tested in stereo is promising.

 

[CaufieldSkeptic]

Preference ratings and 32 magnitude frequency response curves

This repository comprises two CSV files: PreferenceRatings and MagnitudeFrequencyResponses. The former includes preference ratings obtained from 56 naive assessors (30 Danish---DK, 26 Japanese---JP) of 32 headphone curves over several music programs in several trials. The latter includes the...

zenodo.org

View attachment 318109

Has anyone plotted this data in a more readable manner? The preference ratings+confidence intervals are fine but the graph spaghetti is almost impossible to parse

 

[alumnicesar]

even the rating and confidence interval graph sucks. it should also have a table with the numbers lol

 

[youngho]

And, it's related, an interesting "Scientific milestones in the evolution of modern recording control room design".

You might enjoy reading https://www.audiotechnology.com/features/resonating-with-history and https://www.mixonline.com/business/cover-story-peter-dantonio-doctor-of-diffusion-part-1

 

[MayaTlab]

Has anyone plotted this data in a more readable manner?

Someone has, yes. But perhaps let's wait until the paper is actually published to know how to interpret it, as the preference score results would be quite... interesting, if someone presumes from that study a similar methodology to Harman's.
And I'm quite curious to learn how that second plot was created.

 

[alumnicesar]

Someone has, yes

where?

 

[CaufieldSkeptic]

And I'm quite curious to learn how that second plot was created.

Good point. Best I've been able to do is grab the Preference data itself and import it into an excel sheet. Otherwise, I'm lost.

 

[krabapple]

I immediately notice a small batch of presentations from the McGill U group (Wieslaw Woszczyk and collabs) ...which gives me pause, as ISTR a fair amount of 'hi rez' cheerleading coming from them some years back, and they were cheerled in turn by Bob Stuart, because who doesn't love backup from a real science department?

Then McGill ran a blind test of MQA in 2018 that seemed to reluctantly reveal it for the sham it is, so maybe it's different there now. Their presentations at this Meeting are all about virtual environments, so more likely it's good stuff.

And speaking of blind tests, Thomas Lund has written some, hmm, curious....things...about...them right here on ASR, so that's also got my antennae up, but looks like he too is concerned about other things at the meeting.

Meanwhile this is the sort of thing I would pay (but not >$400) to see if I could

AES Fall Convention 2023: Forgotten Futures / Forbidden Planet: Re...

View more about this event at AES Fall Convention 2023

aesny23.sched.com

I'm mainly sorry not to be able to see/hear Amir's presentation.

 

[burkm]

'Doof' means deaf, ' Stokdoof' means completely deaf. So don't think audio is of much interest to them .

That "name" is chosen badly, because an international audience might "interpret" it incorrectly.
Depends on the language one might consider:
"doof" in German means usually "stupid" or "dumb". "stokdoof" could be even more of this , it might emphasize "doof"... In German it means something like "strohdumm"

 

[amirm]

I immediately notice a small batch of presentations from the McGill U group (Wieslaw Woszczyk and collabs) ...which gives me pause

Dr. Sean Olive is a graduate of McGill and they have done some good work such as researching recording environments with respect to acoustic treatments. I don't think they have a specific agenda and the work simply depends on interest of the students and faculty.

 

[MayaTlab]

Good point. Best I've been able to do is grab the Preference data itself and import it into an excel sheet. Otherwise, I'm lost.

So the question I have about this plot is : what am I looking at exactly ? How do we go from the csv file to this plot ? Does it include all listeners and test tracks ? How is the data combined / averaged ? What is the upper and lower boundaries ? Etc.
If @192kbps could help us understand it, it would be very nice .

As it stands in that plot HP3 (the only ones I am certain which HPs they are) in all three forms is preferred over the two APHarm2018 targets even though this excerpt is part of the paper's conclusion according to Sean Olive :

https://x.com/seanolive/status/1711824131971531178?s=20

So it could be that whoever created this plot processed the data differently from the study's authors. Or not, IDK. Or I am misunderstanding the excerpt in this tweet.

Quite eager to read the article anyway as I suspect that the methodology used is not quite the same as Harman's.

 

[192kbps]

So the question I have about this plot is : what am I looking at exactly ? How do we go from the csv file to this plot ? Does it include all listeners and test tracks ? How is the data combined / averaged ? What is the upper and lower boundaries ? Etc.
If @192kbps could help us understand it, it would be very nice .

As it stands in that plot HP3 (the only ones I am certain which HPs they are) in all three forms is preferred over the two APHarm2018 targets even though this excerpt is part of the paper's conclusion according to Sean Olive :
View attachment 318476
https://x.com/seanolive/status/1711824131971531178?s=20

So it could be that whoever created this plot processed the data differently from the study's authors. Or not, IDK. Or I am misunderstanding the excerpt in this tweet.

Quite eager to read the article anyway as I suspect that the methodology used is not quite the same as Harman's.

I am a thief, help you @ the original author who made this table.
@ThomasXia

 

[krabapple]

Dr. Sean Olive is a graduate of McGill and they have done some good work such as researching recording environments with respect to acoustic treatments. I don't think they have a specific agenda and the work simply depends on interest of the students and faculty.

No doubt the department's been at the forefront of audio research for years. It's just regrettable that some there took a turn to hi rez promotion for awhile (this was work from Woszczyk's lab) . I think SO's doctorate took place after the hi rez woo period, but I could be wrong, and I don't know who his thesis advisor was.

 

[MayaTlab]

I am a thief, help you @ the original author who made this table.
@ThomasXia

Thank you .

 

[ThomasXia]

So the question I have about this plot is : what am I looking at exactly ? How do we go from the csv file to this plot ? Does it include all listeners and test tracks ? How is the data combined / averaged ? What is the upper and lower boundaries ? Etc.
If @192kbps could help us understand it, it would be very nice .

As it stands in that plot HP3 (the only ones I am certain which HPs they are) in all three forms is preferred over the two APHarm2018 targets even though this excerpt is part of the paper's conclusion according to Sean Olive :

So it could be that whoever created this plot processed the data differently from the study's authors. Or not, IDK. Or I am misunderstanding the excerpt in this tweet.

Quite eager to read the article anyway as I suspect that the methodology used is not quite the same as Harman's.

You can reproduce these plots in Matlab or GNU Octave softwares just by the *.m code script below and the originally downloaded *.csv data files.

As you can see, all data processing is faithful to the original data. No extra selection, filtering or error correction procedures were adopted.

Due to the lack of specific experimental details, only a general boxplot is plotted to describe the distribution of all subjective rating data for each target curve.

No further investigations into the influences of the assessors, the pieces of sound samples and the experiment trials were included in the code and the plots. I am not familiar with Sean Olive's data processing and logical inference procedures either. Happy to see further discussions and your opimions.

clear;clc
table1 = readtable("data\MagnitudeFrequencyResponses.csv");
table2 = readtable("data\PreferenceRatings.csv");

freq0 = table1{1,2:end};
targets = struct;
count = 1;
for i = 2:height(table1)
    targets(count).name = table1{i,1};
    targets(count).freq = freq0;
    targets(count).gain = table1{i,2:end};
    count = count + 1;
end
target_names = [targets(:).name];
figure(231011003);clf(231011003);
for i = 1:length(targets)
    semilogx(targets(i).freq, targets(i).gain);
    grid on;
    hold on;
end
xlim([30,25000]);
xlabel('Freq (Hz)');
ylabel('Gain (dB)');
legend(target_names, 'Location', 'eastoutside');

figure(231011004); clf(231011004);
boxplot(table2.Rating, table2.HeadphoneCurve);
ylabel('Rating');