IPB

Welcome Guest ( Log In | Register )

 
Reply to this topicStart new topic
audibility of insane & standard presets
krabapple
post Jan 4 2005, 16:45
Post #1





Group: Members
Posts: 2274
Joined: 18-December 03
Member No.: 10538



Over on Quadraphonic Quad I've gotten some stone mp3 skeptics to at least consider the possibility that mp3s *can* be audibly identical to their sources. They're willing to take a test of the matter, allowing me to use 'the best' encoding. I'm going to make some test discs containing tracks ripped via EAC and encoded with the HA recommended LAME compiles at insane and standard presets.

NB I haven't made the claim that an mp3 will always be audibly identical to its source -- only that it *can* be, dependent on several criteria. They find even that modest claim very hard to believe.

My own experience is that I, and a few folks that I've tested in a quasi-ABX protocol, am unable to reliably tell the original from the APS encoded mp3 using a variety of sources. As such I've never even bothered with API, which I've always seen referred to as transparent. But my exhcange with the QQ folk got me wondering, *has* anyone here (or elsewhere) ever conclusively demonstrated an ability to identify a well-encoded 320 kbps mp3 from its source? If so, what track(s) were used? I'd like to give my testees at least the option of using stringent test tracks, in addition to track sof their own choosing.

The same question goes for --alt preset standard, btw -- I looked but didn't find any evidence that it *isn't* transparent, but I may not have looked hard enough. And in my own tests so far I may not have used the most revealing material.

Finally, is there a list of 'tough' tracks (e.g. as used for the more widespread 128 kbps tests), and any place to download them as .flac files?
Go to the top of the page
+Quote Post
SirGrey
post Jan 4 2005, 17:33
Post #2





Group: Members
Posts: 241
Joined: 8-February 04
Member No.: 11863



QUOTE
*has* anyone here (or elsewhere) ever conclusively demonstrated an ability to identify a well-encoded 320 kbps mp3 from its source?

It will be hard, but possible, as I remeber on extremely rare samples. Can not find it on this forum though sad.gif
May be castanets sample was ABXed by someone ? Can't remember...
Go to the top of the page
+Quote Post
jrbamford
post Jan 4 2005, 18:13
Post #3





Group: Members
Posts: 309
Joined: 1-December 01
Member No.: 569



QUOTE (SirGrey @ Jan 4 2005, 04:33 PM)
QUOTE
*has* anyone here (or elsewhere) ever conclusively demonstrated an ability to identify a well-encoded 320 kbps mp3 from its source?

It will be hard, but possible, as I remeber on extremely rare samples. Can not find it on this forum though sad.gif
May be castanets sample was ABXed by someone ? Can't remember...
*


I managed to ABX fatboy encoded with alt preset insane on my Sennheiser HD600s... we've all heard fatboys problems at 128k and the like.. that pumping pre echo sound gets slowly and slowly quieter the more bits you throw at it... at insane i was able to still hear it but i had to turn the volume up VERY loud to make it audible.. it hurt my ears as i recall and probably lost me a few hz off my top end permanently (and enhanced my own ringing in my ears probably) .. I could ABX it tho 15/15 times ...

I was doing it at the time to compare to MPC ... and at standard i think or definately settings above that the pre echo effect was not there and as that was all i was listening out for i couldn't ABX it.. that finally made my mind up that MPC was for me, tho i accepted i'd never listen at that volume on headphones...

My speaker system is very poor at resolving the artefact, even at 128k shockingly the pre-echo is pretty much inaudible at sensible volumes... when i get a new room for my hifi i'll buy new better speakers and will try out to see how close they are detail wise to the headphones..


--------------------
Binaural recordings of mine: http://binaural.jimtreats.com
Go to the top of the page
+Quote Post
Yaztromo
post Jan 4 2005, 19:46
Post #4





Group: Members
Posts: 236
Joined: 28-July 03
From: England, UK
Member No.: 8031



The hihat problem sample is an API killer, quite easy to abx the pre-echo.

http://lame.sourceforge.net/gpsycho/quality.html
Go to the top of the page
+Quote Post
krabapple
post Jan 5 2005, 18:07
Post #5





Group: Members
Posts: 2274
Joined: 18-December 03
Member No.: 10538



QUOTE (Yaztromo @ Jan 4 2005, 10:46 AM)
The hihat problem sample is an API killer, quite easy to abx the pre-echo.

http://lame.sourceforge.net/gpsycho/quality.html
*



That page looks to be exactly what I need. Thanks!
Go to the top of the page
+Quote Post
ssamadhi97
post Jan 5 2005, 18:48
Post #6





Group: Developer (Donating)
Posts: 1203
Joined: 10-February 02
From: Endless Water
Member No.: 1305



abxing castanets on api shouldn't be a problem either


--------------------
A riddle is a short sword attached to the next 2000 years.
Go to the top of the page
+Quote Post
Pio2001
post Jan 6 2005, 20:13
Post #7


Moderator


Group: Super Moderator
Posts: 3936
Joined: 29-September 01
Member No.: 73



FAQ :

http://www.hydrogenaudio.org/forums/index....ST&f=16&t=7783&
Go to the top of the page
+Quote Post
krabapple
post Jan 7 2005, 17:49
Post #8





Group: Members
Posts: 2274
Joined: 18-December 03
Member No.: 10538



QUOTE (Pio2001 @ Jan 6 2005, 11:13 AM)


Thanks Pio. Do I understand correctly that -Z is now 'on' by default when using APS in LAME 3.90.3?
Go to the top of the page
+Quote Post
2Bdecided
post Jan 7 2005, 18:01
Post #9


ReplayGain developer


Group: Developer
Posts: 5139
Joined: 5-November 01
From: Yorkshire, UK
Member No.: 409



QUOTE (krabapple @ Jan 7 2005, 04:49 PM)
Thanks Pio.  Do I understand correctly that -Z is now 'on' by  default when using APS in LAME 3.90.3?
*


Yes.

There are still problem samples though. IIRC guruboolez hears several for api and harpsichord music. Some of the ones in the above thread are solved in later lame versions, or different mp3 encoders, so it's not a fundamental mp3 problem.

(Those other lame versions or different encoders have other problems, which is why aps and api are best overall - though for a given sample at a given bitrate you might be able to find something that beats one of the recommended settings).

Let us know how you get on with your test/challenge.

Cheers,
David.
Go to the top of the page
+Quote Post

Reply to this topicStart new topic
1 User(s) are reading this topic (1 Guests and 0 Anonymous Users)
0 Members:

 



RSS Lo-Fi Version Time is now: 21st September 2014 - 13:29