Absolute polarity? I.e. a 180 degree phase shift at all frequencies?
-
WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!
Audibility of group delay at low frequencies
- Thread starter Kvalsvoll
- Start date
OP
- Thread Starter
- #22
It is. At 100hz, the +-10ms irregularities on the GD actually corresponds to one complete 360 deg cycle.
OP
- Thread Starter
- #23
For abs pol test you don't need anything fancy, just flip the polarity.
Ok, but how can you have DG less than 10ms in a room that isn't anechoic? Reflections will make sure GD (and everything else related to time domain) is all over the place..
OP
- Thread Starter
- #25
It isn't really 10ms, fluctuations are larger. But mean is close to zero. It will end up like this when early reflection level is sufficiently low.
KSTR
Major Contributor
That's unusual. Normally, most people agree on audibility of absolute polarity changes to specific typical test signals, for example a highpass-filtered unity step stimulus, with good headphones.
Early reflections can be "sufficiently low" only in an anechoic environment. Average living room, even if treated to be pretty decent for music listening, will have all kind of reflections (early and late) and they will not be low. Filters that I created have corrections in time domain as well and yet you saw how GD looks when measured at LP if I don't apply IR windowing.
So, is your room anechoic and how exactly did you measure GD? Was it a nearfield measurement? Have you applied any gating (windowing)?
OP
- Thread Starter
- #29
This is an ordinary REW measurement at lp. This can then be analyzed in different ways, to visualize different aspects of performance. Such as GD. The room is very far from anechoic, it is in fact quite live, but early reflection levels are low. This is due to the loudspeakers and room acoustic treatment.
For this test, it is mostly the bass range and perhaps lower midrange that matters. And this room is not very spectacular in the bass range - acoustically. The purpose of showing the room GD in the article was to show that this was audible in something that preforms like a real-world system.
OP
- Thread Starter
- #30
I tried in the article, perhaps if you read it once more, and give me some clues to where you lost it, I can try to explain a little better.
The discussion about how the GD graph is possible - or not - in Room2 is not important here, the room and that system is only used to listen to the sound samples.
The purpose of posting this here is:
- Getting the message out and have at least some people read the article.
- Someone might find a fault in how the experiment was performed, which can then be corrected.
A prerequisite is of course that it is possible to understand why, what and how I did this.
I am trying to get a sense of what you heard. The only way to do that is to have a capture of the sound produced in your room. My headphones won't match what you heard. And at any rate, literature already stated that phase effects are audible in headphones and anechoic conditions.
DDF
Addicted to Fun and Learning
- Joined
- Dec 31, 2018
- Messages
- 617
- Likes
- 1,359
Please see attached for results from other studies, including loudspeakers.
Also, "Phase Distortion and Phase EQ in Audio Signal Processing- a Tutorial Review" By Doug Preis, Tufts Univ. delivered at the 70th convention of the AES, Oct. 1981. Preis graphically represent the results of 7 studies, each using headphones and each using non-musical signals (clicks, sines, tone bursts, etc.). Each test was performed over a narrow frequency band. From 50 to 200Hz, threshold for audibility was 2.5 ms.
Laurie Fincham of KEF fame tried to tackle the issue of low frequency audibility in "The Subjective Importance of Uniform Group Delay at Low Frequencies", JAES, June '85. He found that the insertion of two cascaded all-pass filters with a Q of root(2) each (Butterworth), and maximum phase shift at 40-50 Hz did in fact "cause distinct audible differences to be observed by most of the audience in a typical lecture theatre". Controls may not have been so rigorous though.
Also, "Phase Distortion and Phase EQ in Audio Signal Processing- a Tutorial Review" By Doug Preis, Tufts Univ. delivered at the 70th convention of the AES, Oct. 1981. Preis graphically represent the results of 7 studies, each using headphones and each using non-musical signals (clicks, sines, tone bursts, etc.). Each test was performed over a narrow frequency band. From 50 to 200Hz, threshold for audibility was 2.5 ms.
Laurie Fincham of KEF fame tried to tackle the issue of low frequency audibility in "The Subjective Importance of Uniform Group Delay at Low Frequencies", JAES, June '85. He found that the insertion of two cascaded all-pass filters with a Q of root(2) each (Butterworth), and maximum phase shift at 40-50 Hz did in fact "cause distinct audible differences to be observed by most of the audience in a typical lecture theatre". Controls may not have been so rigorous though.
OP
- Thread Starter
- #34
This was difficult.
Using headphones, and could not hear any difference. Loud or not. Figured it out - sort of, after some listening - I can hear a difference, on headphones. Don't know the volume, they are not calibrated for level, but not very loud.
I made a new short 5 sec set of sample files. I listened for the 5. drum stroke, and verified I can hear a difference.
This is not very easy to hear - on low volume or on headphones. I find it easier the louder it gets, on loudspeakers.
Links to sample files are in the article, at the bottom.
Which one did you test? 20 msec or 100 msec GD?
Nice!
I think you can apply 1/12 smoothing to get a clearer picture. Btw, "abnormals" are MLs?
I can only imagine what reflections would do to a phase/GD trace of these tall fellows. My speakers have one driver firing front and another firing upward working up to 1800Hz, but these speakes have an array of drivers and each of them would generate a wave that would come at a different time to the LP and the same will happen with reflections of those waves. Yet, they seem to get good reviews.
Last edited:
I downloaded 3 samples from your page and took time to listen to them. I can't do a proper blind test as I'm using Volumio and foobar doesn't work well as UPnP controller, however I found the difference between the files not so hard to spot. I would describe it as the main bass drum kick was "thinner" and like it lost part of low energy kick while in the original the kick sounds "mightier" and deeper. I also heard difference in the higher drum kick which also became "thinner" and with less snap when GD is introduced.
I can turn the knob on my binoculars and smooth things out too.
Correct, the abnormals are the 'stats. They don't spray the room in the approved manner.
I don't think that is a good analogy.
Feel free to provide one.
