I'm new to EAC/LAME--I just set it up a couple days ago using the version and settings recommended here. I tried to read everything I could to understand what I was doing, but I'm still confused about a few things. One thing that is puzzling me is that the bitrates I'm seeing are not what I expect based on my settings.

For instance, I re-ripped a Led Zeppelin CD for my husband that he had done in Music Match at 128 CBR, which had a lot of audible clicks. He wants a small file size and listens to his music in a very noisy environment, so that setting works for him. I wanted to see if I could get him something a bit better with comparable filesize by using the V 5 setting, though.

The command line was
-V 5 --vbr-new --id3v2-only --pad-id3v2 --ta "%a" --tt "%t" --tg "%m" --tl "%g" --ty "%y" --tn "%n" %s %d

I expected bitrates between 110-150, averaging around 130. But pretty much everything I saw flashing by in the lower left corner of the LAME window was in the 150s and 160s. The final file size was definitely too big for him. I even double checked to make sure it hadn't somehow switched itself to V 4.

Did I get this set up wrong, or does the V 5 --vbr new setting just give LAME a lot more discretion than I figured? Sorry if I'm missing something obvious/stupid--like I said, I'm new at this.

A couple related questions: when something is recorded at a variable bitrate, the player still shows a bitrate for the track--is this a peak, average, or totally meaningless?

Also, I know that the variable bitrate means the encoder can go higher for "complicated" sections of the track. But what exactly constitutes "complicated"? For example, Garcia/Grisman/Rice "The Pizza Tapes" (three guys on acoustic guitars and a mandolin, and good chunks of it are them talking) was a much larger file than I expected. I wouldn't have thought of that as "complicated."

Three guys on acoustic guitars and mandolins is pretty much the definition of "complicated", to an MP3 encoder (also: applause, castanets, trumpet)

The purpose of quality-based VBR (like "-V #") is to guarantee a certain quality level. Although the bit rate for V 5 tends to be around 128kbps, that it just coincidence. The bitrate will vary based on the material.

VBR will give you constant quality, with unpredictable file size as the tradeoff. If you want more predicted bitrate/filesize, just use ABR instead of VBR.

As for bitrate shown on a player for a VBR file, it depends on the player. A smart player would display an average bitrate, but there are many players with dumb firmware that the bitrate shown is meaningless. I wouldn't pay too much attention on it, just enjoy your music.

OK, so then what do the different presets for VBR actually do, if not define the range? I was basing my expectations on the chart in the Wiki article, but like I said everything I saw was outside the range suggested there--150's and 160's. It just seemed like something had to be set wrong. I guess if I really needed it to be lower I should have tried going down to V 6?

The range is 32-320 kbps for all -V settings. What -Vx changes is the quantization level, or something like that. It explains the different bitrate (and quality) for each VBR step. On average -V5 appear to be near 130 kbps. It's sometimes much lower (I had several ~90...100 kbps encodings) and bitrate is some other time much higher. Extreme cases also exist: from my own experience -V5 average bitrate fluctuates between 60 kbps to 200 kbps. Your music (or mastering principles) is therefore not the most bitrate consuming, but is unusual enough.


This, or ABR.

Well, i had similar problems with -V 5 due to being a real lover of Metal music. Encoding tracks featuring complex guitar sounds or mixtures of guitar and orchestral music, like they're found in the Symphonic Metal genre, usually made the average bitrate explode. For instance, Edenbridge's "Arcana" reached a horrific 154 kbit/s bitrate when encoded using the -V 5 setting of LAME 3.98 Alpha 6 - a bit too large for what i desired. Therefore i decided sticking to what you already proposed as well - a quality setting of -V 6. The resulting files still sound transparent to me, causing this solution to be quite a good compromise. I could recommend it to you as well, although it's up to your ears whether you'd agree with me in this point. Simply encode a few tracks using -V 6 and decide for yourself. That's always better than relying on subjective opinions given by some wiseacres.

I haven't heard from any wiseacres, but I wound up just recording the CD at the constant 128 he's used to, because he's more concerned with file size. Given that he's primarily listening at the airport--outside--perfect quality is not such a key factor right now. I just hoped to do a bit better since it was a borrowed CD, and it wouldn't be around if we wanted to re-do it later. I already gave it back, or I'd try the other settings.

I'll keep all this in mind for the next time the issue comes up. I'm just glad to learn that it wasn't user error.