I'm getting jittery. And so are my high-res MP3s.
April 5, 2008 2:06 PM Subscribe
Why are some of the high resolution MP3s I've encoded recently (at 320 kbps) still turning out with traces of faint, but audible jitter?
As I understand it, MP3s encoded at such a high bit-rate really shouldn't turn out with any audible jitter, and most other MP3s I've heard encoded at that bit rate are noticeably cleaner. It might be relevant that I've been using a very old version of WaveLab to encode my MP3s. I've been using fixed rather than variable bit rate encoding methods to minimize the lossiness of the compression. I'm wondering if there have been any significant improvements to the core MP3 compression algorithms over the last few years that I might not be up-to-date with?
The current machine I use for recording and mastering audio hasn't been on-line in close to five years so I've missed a lot of software updates, although I did install an upgrade to XP (I think SP2) a while back. Could it be that my audio codecs are old and inferior? Or does that even factor into it? (I'm a pretty technically-oriented guy, but some of this is a little outside my domain.)
I don't find the issue with my source wave files, so I don't think it's a problem with the wave files prior to encoding, although it's possible I haven't been using the best dithering methods when I mix down from 24 bit to 16 bit (normally I mix down and master in 24 bit, but for these mixes, I've been mixing down straight to 16-bit).
Anyway, what are some other possibilities I might explore to improve the fidelity of my high resolution MP3s?
As I understand it, MP3s encoded at such a high bit-rate really shouldn't turn out with any audible jitter, and most other MP3s I've heard encoded at that bit rate are noticeably cleaner. It might be relevant that I've been using a very old version of WaveLab to encode my MP3s. I've been using fixed rather than variable bit rate encoding methods to minimize the lossiness of the compression. I'm wondering if there have been any significant improvements to the core MP3 compression algorithms over the last few years that I might not be up-to-date with?
The current machine I use for recording and mastering audio hasn't been on-line in close to five years so I've missed a lot of software updates, although I did install an upgrade to XP (I think SP2) a while back. Could it be that my audio codecs are old and inferior? Or does that even factor into it? (I'm a pretty technically-oriented guy, but some of this is a little outside my domain.)
I don't find the issue with my source wave files, so I don't think it's a problem with the wave files prior to encoding, although it's possible I haven't been using the best dithering methods when I mix down from 24 bit to 16 bit (normally I mix down and master in 24 bit, but for these mixes, I've been mixing down straight to 16-bit).
Anyway, what are some other possibilities I might explore to improve the fidelity of my high resolution MP3s?
Oh, in case it wasn't clear.... EAC is a CD ripper, and LAME is a command-line utility that's developed and distributed separately, but which EAC can call and 'drive' for you transparently.
If you're generating your own files, then EAC obviously can't rip and get tags for you. You'll probably have to manually run LAME as a command-line utility, and will have to manually insert ID3 (identification) tags yourself. This is a bit fiddly, but there simply is no better audio compression quality available under Windows.
If you want the best MP3 files, you use LAME. Period. Suffer through whatever you must to get it working.
posted by Malor at 2:18 PM on April 5, 2008
If you're generating your own files, then EAC obviously can't rip and get tags for you. You'll probably have to manually run LAME as a command-line utility, and will have to manually insert ID3 (identification) tags yourself. This is a bit fiddly, but there simply is no better audio compression quality available under Windows.
If you want the best MP3 files, you use LAME. Period. Suffer through whatever you must to get it working.
posted by Malor at 2:18 PM on April 5, 2008
Building on what Malor said:
1. Get Exact Audio Copy
2. Get the latest LAME binaries at Rarewares.
3. Follow this guide to set up EAC and configure LAME.
LAME -V0 -vbr-new is the best MP3 quality you can get without working up a serious headache doing tons of tests and tuning.
posted by Inspector.Gadget at 2:40 PM on April 5, 2008
1. Get Exact Audio Copy
2. Get the latest LAME binaries at Rarewares.
3. Follow this guide to set up EAC and configure LAME.
LAME -V0 -vbr-new is the best MP3 quality you can get without working up a serious headache doing tons of tests and tuning.
posted by Inspector.Gadget at 2:40 PM on April 5, 2008
Best answer: Since your source is most often not a CD, just use LameDropXPd, a LAME encoding GUI with access to all command options without much more.
posted by Inspector.Gadget at 2:42 PM on April 5, 2008
posted by Inspector.Gadget at 2:42 PM on April 5, 2008
Best answer: 3rded. LAME is full of WIN. Having said that though, its possible LAME can still produce crappy results if other parts of your system are sub-par, so other suggestions would be:
--are your systems audio/sound drivers up to date ?
--is encoding the only place you hear jitter?, if so, thats probably a dead giveaway its whatever encoder you are using.
--There have been approx 80+ Windows Updates since Service Pack 2. You REALLY should consider installing them. You sound like you dont want to because that system never goes online, but the stability and performance increases you could get might help. Another thing you'll want to consider (on top of updating sound card drivers) is to update your motherboard chipset drivers. The combination of all these updates would most likely improve your system (and help protect it, if you ever do need to connect it to the internet).
--Obvious. Defrag and make sure you have enough memory and free hard drive space. Encoding MP3's shouldnt require to much, but if your machine has to struggle to cache information, that could potentially affect program performance detrimentally.
posted by jmnugent at 3:07 PM on April 5, 2008
--are your systems audio/sound drivers up to date ?
--is encoding the only place you hear jitter?, if so, thats probably a dead giveaway its whatever encoder you are using.
--There have been approx 80+ Windows Updates since Service Pack 2. You REALLY should consider installing them. You sound like you dont want to because that system never goes online, but the stability and performance increases you could get might help. Another thing you'll want to consider (on top of updating sound card drivers) is to update your motherboard chipset drivers. The combination of all these updates would most likely improve your system (and help protect it, if you ever do need to connect it to the internet).
--Obvious. Defrag and make sure you have enough memory and free hard drive space. Encoding MP3's shouldnt require to much, but if your machine has to struggle to cache information, that could potentially affect program performance detrimentally.
posted by jmnugent at 3:07 PM on April 5, 2008
Can you post a sample? And/or, can you describe the problem using a different word than jitter?
Also, are you using joint stereo, full stereo or that third option?
posted by gjc at 6:18 PM on April 5, 2008
Also, are you using joint stereo, full stereo or that third option?
posted by gjc at 6:18 PM on April 5, 2008
Another vote for Exact Audio Copy + LAME. Some more info on the Hydrogen Audio wiki.
I've never used it, but dbPowerAmp has its adherents, is at least as good as EAC at ripping. It's where the AccurateRip technology originated.
posted by unmake at 9:50 PM on April 5, 2008
I've never used it, but dbPowerAmp has its adherents, is at least as good as EAC at ripping. It's where the AccurateRip technology originated.
posted by unmake at 9:50 PM on April 5, 2008
Response by poster: for clarification, here's a more precise description of what i mean by jitter--it can be a pretty subtle thing, but it's a lot easier to recognize when you've heard the more extreme forms of it that result when you encode a really low bit-rate MP3... I'll be back with a link as soon as I can get the file up on my server.
posted by saulgoodman at 2:20 PM on April 6, 2008
posted by saulgoodman at 2:20 PM on April 6, 2008
Response by poster: Sorry I never got back with that example of an mp3 with traces of jitter. Have been busy as all hell lately. Thanks again for the tip on LameDropXPd. I used it to make this new mp3 over on MeFi Music.
posted by saulgoodman at 7:07 AM on April 23, 2008
posted by saulgoodman at 7:07 AM on April 23, 2008
This thread is closed to new comments.
EAC + LAME is _the_ solution under Windows. There is no other really worth mentioning, IMO. You'll have to fiddle some to figure out how to get it to do 320k, but it's not overwhelmingly difficult -- I think it's a pulldown option in compression configuration.
posted by Malor at 2:12 PM on April 5, 2008