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: Compression regression? (Read 6593 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Compression regression?

I've got an image file "Michael Lee Firkins - (1990) Michael Lee Firkins.wv" which is 249mb. It was compressed in may 2005 - not sure which version I used. I am pretty sure I only used -hm as option, and foobar reports it as "compression high" (never had patence for those -xN options). It was ripped with EAC with cuesheet embedded.

I can't seem to recompress it to smaller than 256mb these days, whatever I do (however not bad compared to FLAC -8 which is 260mb). I've tried -h, -hh, -hhx, -hx2, -hx3 with version 4.31, 4.41, and 4.50beta, but to no avail. CheckWavpackFiles reports OK on the file.

The difference in size (and bitrate) is about 2.6% - not a big deal but it could be interesting to know what happened.

Compression regression?

Reply #1 –
Even at 260millibits, that's still only just over one-quarter of a bit. That's an incredible compression ratio!

Cheers, Slipstreem. 

Compression regression?

Reply #2 –
Quote from: Slipstreem on
Even at 260millibits, that's still only just over one-quarter of a bit. That's an incredible compression ratio!

Cheers, Slipstreem. 

Ah. mb = megabytes = size. The bitrates for the old file was 864 kbps, and about 887 for current encodes.

Compression regression?

Reply #3 –
Its normal. V4.4 was tuned different from 4.31 due to new x modes and new high mode.

Compression regression?

Reply #4 –
Quote from: shadowking on
Its normal. V4.4 was tuned different from 4.31 due to new x modes and new high mode.


There surely does seem to be an emphasis on portability with the newest releases. It seems to be a logical and reasonable progression. The new DNS for lossy aside, which is a cool new thing that I'd like to understand more, it seems the older versions might be more optimal for purely non-portable use. I'm surprised it goes all the way back to 4.31 though. I'll have to look into that version since I've been using 4.41 thinking it was the last release dedicated to PC use.

I think its worthy of a discussion to consider which versions are best or optimized for what platforms or needs.

Compression regression?

Reply #5 –
Quote from: shadowking on
Its normal. V4.4 was tuned different from 4.31 due to new x modes and new high mode.

Both 4.31 and 4.4 gave equally bigger encoded files, so this is not the reason.

However, I ripped the image again on my home computer, and tried wavpack 4.41 -hhm. Guess what? Got the same size as the original file - 249MB! Hmm. And the FLAC -8 file was suddenly down to 254MB (from 260)?

Well, I guess the wv file was originally encoded on Win32 XP. The current larger files were encoded on Win64 XP (but with the 32-bits executables). Should this make out the difference?