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: Smaller file sizes with vorbis to mp3 (Read 9815 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Smaller file sizes with vorbis to mp3

Reply #25 –
Testing 1 or 2 samples doesn't give you enough to reach conclusions.
Have you seen any graphs of guruboolez ?
He's one of a kind indeed.

It's a matter of quoting samples to get global statistic values.
Not just saying that it differs from the original,
because picking out the 80 kbps compressed sample against the original is very easy to do with decent ears.

And if I had the ears to do such nice research like guruboolez, I probably would.
Here's some statistics which don't require good ears...
A random pick of my CD collection : ripped, normalized and compressed.

second column : -V8 --vbr-new, in KB
third column : Lancer -q2 | -V8 --vbr-new, in KB
fourth column : gain in %

Bill Haley - Rock Around The Clock; 939; 928; 1,17
Blur - song2; 193; 1101; 7,71
Cappella - U and me; 2418; 2329; 3,68
Chopin - Nocturne2; 1277; 1241; 2,82
E.L.O. - Roll over Beethoven; 3238; 2705; 16,46
Elton John - Nikita; 3721; 3450; 7,28
F Lai - Theme from Love Story; 1752; 1645; 6,11
Fats Domino - Blueberry Hill; 2075; 1947; 6,17
Genesis - Jesus He Knows Me; 2689; 2399; 10,78
Kiki Dee - I Got The Music In Me; 3408; 2925; 14,17
Rammstein - Wollt ihr das Bett in Flammen sehen; 3758; 3301; 12,16
Snap - Rythm Is A Dancer; 2553; 2235; 12,46
Status Quo - What you're Proposin'; 2757; 2336; 15,27
The Carpenters - Close to you; 2093; 2002; 4,35
Toto - Stop Loving You; 2803; 2525; 9,92
Twin Peaks theme - Falling; 2966; 2750; 7,28

----------------------------------39640   35819   9,64
----------------------------------(total) (total)  (gain %)
(Sorry for the lousy formatting - I haven't got a clue how to get a table in this box)

So my calculated average gain is approx 9% with an even higher percentage for the most recent tunes (which is probably coïncidence).

Personal note for halb27 : yes, I have a small mp3 stick with not too much space on it. No, it doesn't play ogg's.
And though my ears aren't that great, wma at 64 kbps VBR is too "squishy" for me. Thanks anyway for your time, mate. It meant a lot to me.

Quote
Please educate yourself better

I can't believe you wrote that, Gambit. Again : I'm not claming anything. And I sense a lot of anger there. This remark really hurts.

Quote
so n-pass is just useless

It's very common with video compression. I'd like to ask Gabriel's opinion on this, if I may.

Smaller file sizes with vorbis to mp3

Reply #26 –
n-pass makes sense when you need a piece of material to fit exactly in a certain amount of space and still have constant quality. For video (which is often written to CD-R's) this is much more important than audio.

It'd be nice if you had more control over the VBR bitrate in LAME though, 0 to 9 is somewhat limited compared to vorbis' and musepack's floating point q.
Veni Vidi Vorbis.

Smaller file sizes with vorbis to mp3

Reply #27 –
This is getting silly. I already explained that the objective quality is degrading considerably and percieved quality will be just as bad depending on the sample. I don't know why this voodo is encouraged here  - it is more typical of other boards.

When you need bitrate control use ABR which isn't much worse than VBR.

Smaller file sizes with vorbis to mp3

Reply #28 –
a) when comparing V8 result against the 2-stage q2 - V8 result it is to be expected that quality degrades with the combined method. The best to be expected is that filesize decrease is a good trade against quality decrease.

b) when comparing the 2-stage q2 - V8 result however against the corresponding abr result (abr chosen so that filesize is identical) things are not so clear (the usual trancoding statement which is valid for a) doesn't apply here because this is a different comparison). At least it can give room for different personal preferences as these are two different methods with different properties. The combined method for instance might be better off with HF behavior, if the corresponding abr method should use a lower lowpass. This behavior then might be preferable to some people.

dirkvl, you did your comparison the b) way which to me is the most adequate comparison. While I can understand being afraid a bit of abxing in difficult situations your actual situation seems not to be difficult as you believed already to have heard differences (sorry but without an abx test it is to be considered 'belief'). You can confirm your result by abxing on your own. It's pretty easy with foobar.
lame3995o -Q1.7 --lowpass 17
 

Smaller file sizes with vorbis to mp3

Reply #29 –
This thread is killing too many brain cells. If you want lower bitrates use lower bitrate settings. Thread closed.