Skip to main content

Recent Posts

3rd Party Plugins - (fb2k) / Re: JScript Panel
Last post by easonjim -
Dear marc2003,
Thanks very much for the reply.
I get it now. Thank you.
Sorry guys, I really should have done this before posting the thread, but…

I have just done some blind testing (for the first time in many years) and I am horrified to discover that my ears can actually hear the difference between LAME V0 and WAV! To me, the mp3s sound like they're missing high frequencies, and thus they sound slightly duller (and narrower). In contrast, the WAV sounds less like a recording and more like "real life".

It's only very subtle, but it's there. I did quite a lot of blind testing - sometimes both samples sounded the same, however in the majority of cases I could identify the lossy version. (There were no cases where I mistook the mp3 for the lossy version, I either got it right or I wasn't sure.)

So now, after this, I don't think I can use mp3 again! I'm not about to re-rip everything (some of the CDs I don't even have any more), but I definitely want to up the quality in future.

So my new question is: is there any codec somewhere between LAME quality and lossless? I don't like the idea of my music containing "useless" data, but for me at least, mp3 is not high quality enough. (Funny, I never thought I would say that - I always completely trusted the fact that LAME V0 is regarded as "transparent" - but for me personally, my ears seem tell a different story.)

Do you make music transcription?
Thanks for your answer, which was very helpful and very relevant. Yes, I do transcribe music sometimes, and when I do, I sometimes slow down tracks and use "center channel removal". But this is fairly rare - not enough for it to be important. I fully expect an mp3 to sound poor quality when you mess with it like this. It's only really "normal listening" I'm interested in (or "normal listening" with one channel).
a) I'm not sure but what your post tells me is that you do worry about quality and - if it were possible - you would prefer to go lossless.

b) Lame 3.99.5 -V0 is expected to be absolutely fine for nearly everything. You can look at the results of the last public test @128 kbps to see that Lame3.99.5 is very good even at this moderate bitrate.
If you like to go a little step further with Lame for some very special samples you can try my Lame variant lame3995o and use -Q1 (224 kbps on average for pop music) or -Q0 (317 kbps) or anything in between, for instance -Q0.5.

c) If you would like to use a more modern codec you can use opus (or AAC). opus --bitrate 128 is expected to be transparent for nearly every kind of music. You can give some headroom like using --bitrate 192 or even more to make sure that even tracks extremely difficult to encode are handled fine. According to a listening test I performed recently opus has the great merit that some tonal problems I care about are handled fine at a ~100 kbps (avg.)
 setting where other codecs need a significantly higher bitrate. (I did care about these when developing lame3995o, and it takes a ~170 kbps (avg.) setting to handle them fine).

d) If you can allow for ~400 kbps lossyWAV+FLAC is a near-lossless option. Same goes for wavPack lossy resp. hybrid. You can go even lower with bitrate when using lossyWAV+FLAC, but in the ~300 kbps region I'd prefer opus (or lame3995o or QAAC).
The best advice I can give is: Don’t panic! ;)

As well tested as LAME is it’s extremely unlikely that V0 is anything but perfectly transparent – maybe you’ll find problematic samples, but they’ll be few and far between.

But there’s nothing wrong with doing some ABX tests yourself to get back your peace of mind. If you’re not able to hear a difference between the original and the encoding in a double-blind test done in your usual listening environment(s) with the music you listen to all the time, then there isn’t any point in switching formats, is there?

If even that’s not enough to combat the paranoia, then consider solving the problem once and for all by switching to a lossless format (e.g. FLAC).
Silly comment but if
Do you make music transcription? Some audio processing plugins manipulate the phase of stereo signals to emphasize or minimize some instruments panned on a specific position to make transcription easier, the most common one is the so-called "vocal removal" technique which assumes vocal is mono and centered on the stereo field.

With mp3 or some other lossy codecs such techniques could yield poorer performance.

For only that issue I have found that AAC-LC is the best option by far from any other, but I repeat, only for that issue.

If your don't have any other of the common codec-switch issues like running out of storage on that particular device, need a more efficient codec for given (lower than transparent) bitrate or compatibility, then I don't see why you would need to re-rip everything from your catalog if everything is already transparent to you, you'll gain nothing.
I had this idea too, I'm glad I wasn't the only one.
Well, atleast 90% of all Top40 tracks you listen to the radio, if you create a duplicate in a DAW and you invert it's phase, then match the first verse with the second verse and the first chorus with the second chorus, pretty like 75% of the content gets cancelled out in the verses (the rest will be mostly some reverb and the vocals) and the chorus will be mostly cancelled out as a whole. I bet that a potential codec that could take advantage of such thing (lossless or lossy, both could get potential advantage) would have one problem. When things start to get out of phase and they are not identical anymore it makes it inefficient and hard to find a way to compress uniformly. I mean you could have a typical Top40 pop track that goes like verse1-chorus-verse2-chorus-bridge-chorus that could be compressed to crazy small files (lossless or lossy) with amazing efficiency (percivable quality or compression ratio) and then a piece of classical music that most possibly has nothing that it's phase is a douplicate of an other part and it will compress poorly. Still though, there are so many video codecs that have tuning for different types of inputs to them, creating a long term prediction subcodec/codec that is based on the idea that you may find the same thing again on the file with potentially minor differences wouldn't be as much of a bad idea as alot of people will try to claim that it is.

Your only enemies are processing power, long encoding times (because of looking waaay ahead to the file) and the fact that you need a coder that is very smart when he writes code, he must definitely be a music producer to "get it".

Also this idea is pretty similar to treating stereo content as quadraphonic where the front and back channels (when the encoder finds a match in the file) are treated as mid/side at the same time as left and right does too, so essentially you will have a mid/side for each couple of the channels.
Do you make music transcription? Some audio processing plugins manipulate the phase of stereo signals to emphasize or minimize some instruments panned on a specific position to make transcription easier, the most common one is the so-called "vocal removal" technique which assumes vocal is mono and centered on the stereo field.

With mp3 or some other lossy codecs such techniques could yield poorer performance.

Sometimes you may also want to slow the song down to pick up some fast notes, lossy codecs could also yield poorer performance.

However, in my experience, if a stereo mp3 is transparent (audibly lossless), it will still transparent when only one channel is being played.
Sure, the chorus of a particular song may have the exact same musical notation and the same lyrics, but no band plays perfectly according to the notes, 100% every time. There will always be variation, some of it is deliberate, some of it isn't. If you take that away, and basically just play a recording of the chorus every time you get to that part of the song, it'll make the songs lifeless, generic and boring....
If a (compressed) MIDI file represents the core musical information of a song, then might not "musically guided" loss represent sensible rate:distortion compromises? While repeated musical sections might be numerically different (due to noise, performer involuntary variation or performer conscious variation), it sounds like an interesting axis to work in.

I know that scores of e.g. Bach music may not be available from the composer himself, rather, some musically gifted soul listening to Bachs performance went home and transcribed the music from memory. While I find that capability amazing, I guess that there will be errors. Hopefully, the "essence" of the music survived this operation.

Perhaps because I am interested in both dsp and music, I am intrigued by the idea of lossy codecs being able to analyze a piece of music in a musically sensible manner (i.e. waveform to score) perhaps using some (algorithm du jour) machine-learning mechanism, then figuring out what matters the most to (to e.g. someone coming from a western musical tradition), and how to spend bits most wisely.

The comparision to video codecs is interesting. AFAIK, they don't even have an explicit model of our vision (unlike audio codecs), and when they track "motion" across temporal frames, they will often find "apparent" motion that does not correspond well with actual motion. I.e. they will pick up "something" that allows them to encode the residual with fewer bits, but nothing like a plausible optical flow type modelling.

Hi guys :) I could do with a little advice, please…

For the last 15 years I've happily chosen to encode my music with LAME mp3 (V0). I've always considered this perfect for me for a few reasons…

My criteria
  • I am fussy about quality - but not to extremes
  • I'm only a casual listener, with cheap earphones, I'm not an "audiophile" when it comes to listening to music
  • I enjoy music, but ultimately my collection is not terribly important to me
  • I'm happy as long as the music sounds like original most of the time - I don't mind the occasional 'imperfection' or 'glitch'
  • I'm not fussy enough to care about the whole "subconscious perception" issue
  • I do have a very large music collection, and even though hard drive space is cheap, I don't like the idea of how much space lossless would take up
  • Another key factor for me is that I'm a "minimalist" and find it satisfying to know that my music files take up as little space as possible, with all the unneeded data stripped away

So LAME mp3 has always been fine for me, on this basis.

But lately I've been wondering if it might be time to switch codec? Maybe there's something much better than mp3 these days? Perhaps ogg or something? I really known nothing about the other lossy formats.

But in particular, something else has been bugging me about LAME mp3… As a musician myself, I have a real interest in how all the parts in the track are mixed. So, while normal listeners will just focus on the lead vocal and main melody - my ears are more sensitive to "every single part of the mix" - I will often be paying attention to individual instruments in the background, or individual notes in chords - things that are perhaps really subtle in the mix - two instruments blended together, or one very quiet instrument in an orchestra hard-panned to the far right, just about audible in one ear. These are details that most people wouldn't care about. And then sometimes I will listen just to an isolated channel (left or right) just to hear what is going on in each channel.

So I guess I'm asking: Is LAME V0 good enough for me? I'm not obsessively fussy about the music being "perfect", but I do want to be able to hear all the parts and all the harmonies, in all the mix, across both the channels - including all the subtle nuances of every instrument (something a regular listener may not care about).

And can LAME V0 be trusted for playing just the left or right channel in isolation? Or does it start to break down when you do that? (To my ears, LAME V0 does sound identical to the original WAVs, but I haven't done enough testing to put any confidence in my own conclusions.)