48kbps AAC 2006 Test - Samples discussion Options

[Gabriel]

This is the thread to discuss about samples for the upcoming 48kbps AAC test:
Test discussion
Test home page

Currently selected samples (subject to change):

Debussy, Une Forêt - Classical - low volume instruments and vocal
Kraftwerk, The Robots - Electronic - impulse transients
Suzanne Vega, Tom's dinner - vocals and instruments (where to find this sample?)
french_ad - male voice
Twist And Shout, Beatles - yelling Beatle with fake stereo

This post has been edited by Gabriel: Feb 13 2006, 23:22

[IgorC]

How many samples it's planed to include?
It will be reasonable to include something like 50% hard to encode samples ( like different instruments on chanels and/or high frequences samples) and 50% normal samples. Or something like 33%-hard, 34%-normal,33% -light samples for real life encoding.

I already uploaded a sample: http://www.hydrogenaudio.org/forums/index....showtopic=41248

It sounds painfully at 48 kbit/s

PS: I didn't remeber where I saw before this sample with some stereo effects. http://www.mytempdir.com/437714 .

This post has been edited by IgorC: Feb 8 2006, 04:15

[Alex B]

How many samples  it's planed to include?
It will be reasonable to include something like 50% hard to encode samples ( like different instruments on chanels and/or  high frequences samples)  and 50% normal  samples. Or something like 33%-hard, 34%-normal,33% -light samples for real life encoding.

I already uploaded a sample: http://www.hydrogenaudio.org/forums/index....showtopic=41248

It sounds painfully at 48 kbit/s

Actually, it may not be such a good idea to include a sample of a track that has extreme stereo separation throughout the complete track. That kind of tracks are minority in today's recorded music. Tracks that have occasionally distinct, but not extreme channel separation would better represent a real life situation.

Edit: grammar

This post has been edited by Alex B: Feb 8 2006, 13:24