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: Lame 3100m (halb27 version) vs 3.100 official release (Read 7955 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Lame 3100m (halb27 version) vs 3.100 official release

Looking to rip a few CDs for a USB stick in my car, and trying to decide between using halb27's LAME 3.100m (which I think was built off 3.100 alpha2, but I cannot verify, and was released November 2013), and the official 3.100 release from October 2017.  In my default settings on soft to moderate contemporary rock, I can't tell any difference between the two; after mp3packer, 3.100m averages about 1% larger files than 3.100 (~271kbps vs ~274).

Typical settings are -V0, -Y (aka less accurate high frequencies allowed so they don't steal too much file space), lowpass disabled, and for 3.100m, -cvbr 4.  I tried disabling -Y and adding -cvbr 0 (which raises the file size to ~316kbps) but still didn't hear a difference except maybe in the 8-12kHz range, and that could just be my imagination since I knew -Y was toggled.

I've used halb27's version for years and been happy with it, so don't see any need to switch, but thought I'd check if any differences between the base 3.100a2 and final 3.100 made switching worthwhile.  Thanks in advance.

Re: Lame 3100m (halb27 version) vs 3.100 official release

Reply #1
There's no quality change from 399 > 3100a2 > 3100 in stock lame. You can stick to the halb setting .
You can also try official lame using -V1 or -V1.5 . You get a turbocharged 'V2 / aps' WITH all the HF and
LESS bitrate than V0 settings .


 

Re: Lame 3100m (halb27 version) vs 3.100 official release

Reply #3
Interesting...I didn't know that 3995o existed.  Thanks!  Looks like halb27 released 3995m right after 3100m, then developed versions n and o, with o as the final release.

For my everyday usage (car and TV audio), I doubt there's a perceptible difference between 3100m and 3995o.  Per halb27: For high bitrate settings the differences between 3100m and 3995m are pretty negligible, especially when using stronger --cvbr levels. According to the comparison in my last post I personally prefer 3995m a tiny bit.  But he does mention later he considered n a "major step forward".  He also mentioned 3100m did better with one problem sample (trumpet_myPrince), and 3995m with another (harp40_1).  3995n and o then improved harp40_1 still more.

Overall, it seems to make sense to use 3995o over base 3.100.  Related: any insight why the main LAME developers haven't considered incorporating his upgrades?  I thought about suggesting them via SourceForge but (a) didn't feel like signing up to do so, and (b) as halb27 is the owner of said upgrades, he should be the one to do so.  Sadly, his most recent post (12/24/19) indicated he was fighting aggressive cancer, so he may not still be with us...

Re: Lame 3100m (halb27 version) vs 3.100 official release

Reply #4
@BFG , I was planning to port 3995o changes to 3100 (no ETA though). It should make it easier to submit a patch to upstream.

Re: Lame 3100m (halb27 version) vs 3.100 official release

Reply #5
Glad to hear it - thanks!  I see the developers are already working on a 3.100.1 though, to give you a head's up.

This community seems to agree halb's version creates superior MP3s, with the only expense appearing to be a slower compression rate, so hopefully the developers will take notice once you complete the work.


Re: Lame 3100m (halb27 version) vs 3.100 official release

Reply #7
I@m trying to encode using lame3995o in foobar with custom enocder settings and wondered if you would use the Q switch like -Q1 or -Q 1? Does the space between the number and Q matter to the encoder?
Dunno, but I always write it with spaces.

Re: Lame 3100m (halb27 version) vs 3.100 official release

Reply #8
Does the space between the number and Q matter to the encoder?
According to Lame's manual, there is indeed a space between the switch "-q" and the value (0-9).

Re: Lame 3100m (halb27 version) vs 3.100 official release

Reply #9
According to Lame's manual, there is indeed a space between the switch "-q" and the value (0-9).
That's actually a different switch, but I don't believe the space is critical although I haven't tested it exhaustively.

Re: Lame 3100m (halb27 version) vs 3.100 official release

Reply #10
@BFG:
Gladly I'm still with us, but things seem to get harder now.

I like to hear that my versions are still welcome for some users. I personally use it, too, but further development is out of sight. As Lame 3.100 didn't bring quality improvements it's not attractive to do a Lame 3.100 port. Also I have no ideas how to further improve quality compared to Lame3.995o. But to me quality is pretty much perfect except for special pre-echo stuff.
lame3995o -Q1.7 --lowpass 17

Re: Lame 3100m (halb27 version) vs 3.100 official release

Reply #11
That's actually a different switch, but I don't believe the space is critical although I haven't tested it exhaustively.
That particular switch appears to be not part of the official branch, I checked the manual and history from Lame v3.100.1. I found a description of the "-Q" switch in Lame v3.995o According to the description there must be no spaces between the switch "-Q" and its values.

Quote from 3995oDoc.rtf
Code: [Select]
I created a new -Q parameter to improve upon the situation (apart from the improvements related to issue 1 which the -Q setting takes care of as well).
-Q3, -Q2, -Q1, -Q0.5 yield the same average bitrate as -V3, -V2, -V1, -V0.

Re: Lame 3100m (halb27 version) vs 3.100 official release

Reply #12
I know that's what it says, but I can assure you from personal experience it's parsed perfectly well if there is a space present. ;)

Re: Lame 3100m (halb27 version) vs 3.100 official release

Reply #13
I know that's what it says, but I can assure you from personal experience it's parsed perfectly well if there is a space present. ;)

I've tried both in foobar and it seems to produce files of the same size so looks like the space doesn't matter.

Re: Lame 3100m (halb27 version) vs 3.100 official release

Reply #14
@BFG:
Gladly I'm still with us, but things seem to get harder now.

I like to hear that my versions are still welcome for some users. I personally use it, too, but further development is out of sight. As Lame 3.100 didn't bring quality improvements it's not attractive to do a Lame 3.100 port. Also I have no ideas how to further improve quality compared to Lame3.995o. But to me quality is pretty much perfect except for special pre-echo stuff.

Thanks for all your hard work on this. What setting do you use for your files personally?

Re: Lame 3100m (halb27 version) vs 3.100 official release

Reply #15
Thanks for all your hard work on this. What setting do you use for your files personally?
Look at his signature.
WavPack 5.6.0 -b384hx6cmv / qaac64 2.80 -V 100