Slow Hi-sp Converting In Sonicstage

[L7R]

Can anyone explain why converting to Hi-SP is so slow compared to LP2 or other formats? Has anybody even noticed it. I think it's about 1/2 or 1/3 slower than LP2.

In my mind it should be faster than LP2 because there's less information to leave off.

Anyway, I think this is quite intresting?

Edited July 17, 2005 by L7R

[ROMBUSTERS]

this 'problem' has been around since HiMD has been introduced. I doubt if anyone knows why but here are some of my theories.

It doesn't matter the amount of information the codec chooses to leave out (see: lossy compression) because all codecs take it from a PCM source no matter the source material (i.e. MP3 > PCM > Atrac or CD PCM > Atrac)

The reasons I think it's slower are as follows:

1) The encoder is not as refined as the netMD encoders. netMD has been around for what 5 years now and has had significant time to refine their encoding techniques.

2) Because by definition Atrac3+ is a different codec than Atrac3 or MP3 it is unfair to judge them side by side. Atrac3+ uses a transform window much larger than Atrac3 (fancy way of saying it looks at more of the waveform at once) in order to better judge where the available bits should be allocated. This process of a larger view and estimation calculations are where I feel most of the computing power (and thus time) is placed. Because LP2 for example looks at much less of the wave form per cycle of the encoding process it is able to make its bit allocation quicker.

Just my thoughts

[L7R]

Thanks for clearing up this. Let's hope there'll be some improvement in next version of SS.