AAC 128 kbps listening test NAME : guruboolez LISTENING CONDITIONS : · BeyerDynamic DT-531 headphone · Onkyo AM/FM Tuner Amplifier R-A5 · Volume : average/high _______________________________________________ | | QT 128 | AN 128 | Psy STR| AN STR | B I T R A T E S |___________|________|________|________|________| | sample 01 | 3.5 | 1.5 | 4.1 | 4.0 | 128 / 123 / 139 / 125 |-----------|--------|--------|--------|--------| | sample 02 | 4.2 | 1.8 | 2.1 | 2.4 | 128 / 122 / 139 / 120 |-----------|--------|--------|--------|--------| | sample 03 | 5.0 | (1.4)* | 3.5 | 2.5 | 128 / 109 / 126 / 110 |-----------|--------|--------|--------|--------| | sample 04 | 4.3 | 1.5 | 2.9 | 2.5 | 128 / 124 / 136 / 119 |-----------|--------|--------|--------|--------| | sample 05 | 3.2 | 2.7 | 4.7 | 2.5 | 128 / 126 / 094 / 089 |-----------|--------|--------|--------|--------| | sample 06 | --- | --- | --- | --- | |-----------|--------|--------|--------|--------| | sample 07 | --- | --- | --- | --- | |-----------|--------|--------|--------|--------| | sample 08 | --- | --- | --- | --- | |-----------|--------|--------|--------|--------| | sample 09 | --- | --- | --- | --- | |-----------|--------|--------|--------|--------| | sample 10 | --- | --- | --- | --- | |-----------|--------|--------|--------|--------| | sample 11 | --- | --- | --- | --- | |-----------|--------|--------|--------|--------| | sample 12 | 4.0 | 2.2 | 3.8 | 3.4 | 128 / 125 / 142 / 127 |___________|________|________|________|________|_______________________ 24.2 11.1 21.1 17.3 | 4.03 1.85 3.51 2.88 | 128 / 122 / 130 / 115 (24.6) (15.5) * previous rating => Quicktime valuation is quite the same (24.2 vs 24.6). => Nero CBR encoding rating is even worse than yesterday. The big difference is in two first samples. I suppose that AN CBR is penalized here by the lack of high frequencies ; all new rivals are richer. Strong lowpass is shocking on cymbals ; 41_30 is very rich in cymbals, and Atrain is a jazz sample... VBR encoding made a big difference in notation. Especially with PsyTEL. · Death2.wav is interesting : - PsyTEL VBR is near perfection (on three parts : 5/5, 4/5, 5/5) - Ahead VBR is worse than Ahead CBR : obvious at the beginning - VBR encoding are both near 90 kbps ; one is perfect, the other really bad ! · 41_30 shows a enormous progress between CBR & VBR. - We can guess that Ivan took this sample for tweaking ;-) - QuickTime is worse : no more artifacts, but a stronger lowpass, not really pleasant with this sharp sample. · Atrain is the only jazz sample : - QT is still near perfection - No real progress with VBR ; cymbals are still watering, dirty. · Beautyslept is unfortunatly the only 'classical' sample (big flaw of your test !!!) - I rated QT more generously than yesterday ! - Nero CBR is unbearable ; VBR is better, but not good. Harpsichord lovers like apples ;-) · PsyTEL -streaming consumes more space than Ahead -streaming ; quality is higher, too ********************************************************************************* * (PLEASE, JUST COMPARE THE BITRATE I OBTAINED WITH THE LATEST AHEAD MP4 CODEC) * ********************************************************************************* ················ · CONCLUSIONS · ················ I would consider Quicktime CBR AAC codec as the best challenger for competition. The second may be : · PsyTEL 2.15 VBR -streaming if people are sensitive for high frequencies but not too much for artifacts · Sorenson Squeeze CBR 128 if people are not really sensitive for a 15.500 lowpass, or if they prefer a clean signal => both are awfully slow, but PsyTEL is « free », and is usable on many software (CLI encoder). I would take PsyTEL... Don't forget speed. QT is much faster than Ahead, PsyTEL and Sorenson codec. I must test with 6.3, but I timed with 6.1 a nice x5 (Nero is x2.5-x3 ; PsyTEL and Sorenson are x1.5) => Quicktime is for me the more constant in quality, the more constant in bitrate, and the faster one.