Skip to main content
Topic: The recommended Vorbis version (Read 8352 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

The recommended Vorbis version

... a question: Should we then change the HAK article to recommend beta5? (feel free to discuss/continue this in the Wiki Discussion forum)

Perhaps. I don't know if there is much to discuss about. I have seen very few if any valid test results that would help making the decision.

I promised to do some ABX testing when Aoyumi released b5. I actually encoded a bunch of test samples with R1 and b5. However, I couldn't quickly hear any ABXable differences and I postponed testing to some later time. Looks like that time has not come yet...

I suppose if b5 introduced any new serious problems they would have been discovered during the last few months.

The recommended Vorbis version

Reply #1
... a question: Should we then change the HAK article to recommend beta5? (feel free to discuss/continue this in the Wiki Discussion forum)

Perhaps. I don't know if there is much to discuss about. I have seen very few if any valid test results that would help making the decision.

I promised to do some ABX testing when Aoyumi released b5. I actually encoded a bunch of test samples with R1 and b5. However, I couldn't quickly hear any ABXable differences and I postponed testing to some later time. Looks like that time has not come yet...

I suppose if b5 introduced any new serious problems they would have been discovered during the last few months.


I would definitely recommend b5. I have pretty much all my music encoded in it and I have never had a problem. On a side note, I use q4 for my music but I could never abx any difference between r1 and b5. However at lower bit-rates like q0 it was easier. (I'm probably violating some TOS now) ) Still, in my opinion, b5 is a very good release.

/Kef

The recommended Vorbis version

Reply #2
Actually, IMO for b5 to be the recommended version over b4.51 (aka R1), b5 needs to fulfill the following:
- No degradation of quality on all -q
- No new artifacts introduction
- No error that may cause damage
- No error that may adversely & significantly bitrate

I haven't used b5 all that much. Most of my collection, i.e. about 95%, are already encoded using b4.51 and I don't want to reencode using b5 unnecessarily. The (very) few b5-encoded ones seem to fulfill all the above requirements.

The recommended Vorbis version

Reply #3
Why do we still need recommended versions ? Why not link the latest version like all software. If someone wants an old version they can use it instead. There was never a real defective version of anything. Above 160k, few if any could tell a difference outside killer samples on any good encoder and its been this way for many years. I suppose we could test < 128k bitrates but even then I doubt there would be major regressions.
wavpack 4.8 -b256hx6c

The recommended Vorbis version

Reply #4
I think I'll start a poll, like this:

What do you think about aoTuV Beta 5?
* It's good enough to replace b4.51/R1 as the recommended version
* It still needs more peer-review before becoming the recommended version
* Let's just drop this "recommended" moniker and point visitors to the latest version

Everybody agree?

The recommended Vorbis version

Reply #5
I recently re-encoded over 10000 files for my portable from the originals to b5 at q1.5 (Using the latest Lancer ss3 build.) I previously had all my files in b4.51/r1 and about 3-400 in b4.

I did try abxing the difference between the two but it was impossible for me at this bitrate, though in my TOS 8 breaking outside of abx opinion, I did think that b5 was generally slightly improved.

I haven't noticed any odd errors or artefacts in files and had only 3 files with errors, all down to me stopping encoding part way through the file.

Note: the main reason for the re-encode was I had a lot of new files and thought I may as well do the lot together.

The recommended Vorbis version

Reply #6
I think I'll start a poll, like this:

What do you think about aoTuV Beta 5?
* It's good enough to replace b4.51/R1 as the recommended version
* It still needs more peer-review before becoming the recommended version
* Let's just drop this "recommended" moniker and point visitors to the latest version

Everybody agree?
I'm not involved in the recommendation but it should certainly not be based on a vote!

The recommended Vorbis version

Reply #7
Heh, point taken, Jan. Which is why I don't immediately start a vote. But a vote would be a nice way to draw everyone's attention and get the collective minds of HA to mull over the answer to the question

( And no, this time, the answer's not 42 )

The recommended Vorbis version

Reply #8
Actually, IMO for b5 to be the recommended version over b4.51 (aka R1), b5 needs to fulfill the following:
- No degradation of quality on all -q
- No new artifacts introduction
- No error that may cause damage
- No error that may adversely & significantly bitrate

I haven't used b5 all that much. Most of my collection, i.e. about 95%, are already encoded using b4.51 and I don't want to reencode using b5 unnecessarily. The (very) few b5-encoded ones seem to fulfill all the above requirements.


Why won't you take some time and test the b5? I loved the b4.51 and the b5 is in my opinion better, especially in the lower bit-rate regions. Yea yea yea, I'm violating TOS but still, ...

The recommended Vorbis version

Reply #9
Actually, IMO for b5 to be the recommended version over b4.51 (aka R1), b5 needs to fulfill the following:
- No degradation of quality on all -q
- No new artifacts introduction
- No error that may cause damage
- No error that may adversely & significantly bitrate

I haven't used b5 all that much. Most of my collection, i.e. about 95%, are already encoded using b4.51 and I don't want to reencode using b5 unnecessarily. The (very) few b5-encoded ones seem to fulfill all the above requirements.
Why won't you take some time and test the b5? I loved the b4.51 and the b5 is in my opinion better, especially in the lower bit-rate regions. Yea yea yea, I'm violating TOS but still, ...
Um, does being lazy counts as a valid answer? 

The recommended Vorbis version

Reply #10
If you want a clear answer for which is better.. just encode some stuff at -q -2 and compare  (q-2 because q4 is too hard for me to abx  )

I recommend aoTuV b5 over r1 any day.  I have yet to encounter a problem with it.
Vorbis-q0-lowpass99
lame3.93.1-q5-V9-k-nspsytune

The recommended Vorbis version

Reply #11
IIRC, r1 was only b4.51 renamed to keep people happy. So, b5 would be the next logical step, it's certainly been around long enough with no complaints that I've seen.
John
----------------------------------------------------------------
My compiles and utilities are at http://www.rarewares.org/


The recommended Vorbis version

Reply #13
* Let's just drop this "recommended" moniker and point visitors to the latest version


I do think it's somewhat arrogant to believe you ("you" not you pepoluan, but "you" the HA people) are entitled to go around meting out recommendations.

It did (somewhat) make sense in the days after LAME 3.90 got released because of some politics imbroglio between Dibrom and the LAME developers that "forced" HA to make its own release of LAME. But that ultimately came back to bite us, as clearly better LAME versions were released, but (again because of politics) HA had to stick to recommending 3.90.3. In the end, as I see it, people simply started ignoring HA, and moved on. That's when 3.96.1 was finally "endorsed".

These days, the HA recommended LAME version is meaningless, it just points to whatever was last released by the LAME devs.


Anyway: that's why RareWares avoids "recommending" versions, we just post what's latest there: because it would smell of hubris, and because we would just point to whatever the developer recommended anyway...

The recommended Vorbis version

Reply #14
What is the community's stance on the "Lancer" versions?


The recommended Vorbis version

Reply #16
Soooo... should I drop the "recommended" hulabaloo altogether?

The recommended Vorbis version

Reply #17
I do think it's somewhat arrogant to believe you ("you" not you pepoluan, but "you" the HA people) are entitled to go around meting out recommendations.
LMAO.  As this board's top poster shouldn't that be "we".
I'm on a horse.


 
SimplePortal 1.0.0 RC1 © 2008-2019