Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: EAQUAL... (Read 5648 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

EAQUAL...

Hi,
Old EAQUAL version (0.1.3 alpha) has been around for a while... Does anyone know how reliable this audio quality evaluation software is? Are the results obtained any good?

Thanks,
daniel

EAQUAL...

Reply #1 –
It's not reliable and its results are little more than rubbish. It was determined that the software algorithm incorrectly implements the underlying psymodel.

EAQUAL...

Reply #2 –
It may be decent for low bitrates, but I won't totally just rely on it.

EAQUAL...

Reply #3 –
I think it depeds on the bitrate. In my experience, it is really a golden ear for middle and high bit rate (96Kbps~384Kbps). It really can tell the audio quality..

EAQUAL...

Reply #4 –
Quote
I think it depeds on the bitrate. In my experience, it is really a golden ear for middle and high bit rate (96Kbps~384Kbps). It really can tell the audio quality..

Well, don't know if you are joking, but especially it was found out by testing and also the author made the statement that it can't tell the audio quality reliably for mid and high bitrates.
Juha Laaksonheimo

EAQUAL...

Reply #5 –
Quote
It's not reliable and its results are little more than rubbish. It was determined that the software algorithm incorrectly implements the underlying psymodel.


When it was determined? I thought that algorighm was implemented correctly - but the application of the basic ITU-BS 1387 is very limited to special areas of quality management?

Basically, ITU-BS 1387 results are not rubbish, but they are not even designed to make general conclusions about the quality. It can just show some trends.

EAQUAL...

Reply #6 –
A while back I correlated EAQUAL's results with how people ranked different codecs on dogies.wav and wayitis.wav at 128 kbit/s.  The correlation was about 0.7 with the human preferences, which is respectable, but not spectacular.

If someone is ambitious and really interested, the same kind of thing can be done for the 64 kbit/s test.

ff123

EAQUAL...

Reply #7 –
Quote
Quote

It's not reliable and its results are little more than rubbish. It was determined that the software algorithm incorrectly implements the underlying psymodel.


When it was determined? I thought that algorighm was implemented correctly - but the application of the basic ITU-BS 1387 is very limited to special areas of quality management?

Basically, ITU-BS 1387 results are not rubbish, but they are not even designed to make general conclusions about the quality. It can just show some trends.

I know it's been posted on HA before that EAQUAL uses the wrong time resolution, i.e. the time slices it analyzes are too big and misses many pre-echo artifacts.

I might consider EAQUAL as a way to compare a codec at different settings but I wouldn't use it to compare different codecs. EAQUAL's results may be generally positively correlated with results from objective listening tests but it is no substitute for them.

EAQUAL...

Reply #8 –
Quote
I know it's been posted on HA before that EAQUAL uses the wrong time resolution, i.e. the time slices it analyzes are too big and misses many pre-echo artifacts.


Resolution is not "wrong" - it is set in the (ITU BS.1387) standard, so EAQUAL is correct implementation.

Same standard has "advanced" model with much better time resolution, but unfortunately,  project was stopped.

EAQUAL shouldn't be used to compare codecs, at least not without advanced model of analysis - but that does not make it "crap" - it has its own application area.

EAQUAL...

Reply #9 –
Standard is as it is... I was wondering how good is the Eaqual 0.1.3 alpha implementation of the standard?

EAQUAL...

Reply #10 –
Quote
Standard is as it is... I was wondering how good is the Eaqual 0.1.3 alpha implementation of the standard?
[a href="index.php?act=findpost&pid=67286"][{POST_SNAPBACK}][/a]


EAQUAL implementation does not conform to conformance test in ITU-BS.1387 standard. For some MOVs it produces almost identical values, while for some (mostly important ones like NLA) it produces completely different values than reference (standard).
But EAQUAL is written based on standard, and my assumption is that reference application used for conformance testing in standard is not written as standard states.
I have the same problem in my implementation of Advanced version. I believe that my implementation does not have big errors and yet for filter bank MOVs I get very different values than reference.
There are some big mistakes in standard or in reference application. Or maybe they just don't want to reveal what is all there.