Absolutely +1 for trying to rid your life of as many physical artifacts as possible!
However, for me as a music producer, it is unacceptable to produce lossy audio. The media of today is about remixing and mashing up - reencoding audio in a lossy way at every new version is what I call disastrous.
Also note that the perception of lossy compression greatly depends on the kind of music. I play the organ, and organ music is especially prone to the artifacts due to MP3 compression.
If the aim is to get rid of the CDs, or at least not have to re-rip them again in this lifetime, then the only rational response is that you should store them in a non-lossy format which has an open-source decoder, say FLAC. Storage is cheap. Encoding is fast. Then you re-encode to the lossy format that is best for you at the time you need it (say in 20 years).
Besides, hydrogenaudio.org have already been doing these sorts of listening tests for years, and better.
Third, if you’re the one who will be listening to these songs, how does what anyone else think matter? Only bad things can come from listening to what others say. That is, you may learn to hear artifacts which you did not previously detect. Is your life now better or worse for it? Personally this is one domain in which I’m happy to not have ‘golden ears’.
I usually encode to ~80kbps vorbis for portable.
Generally I find MP3s acceptable. But, there are some cases where they are really bad.
When I compare my MP3s of Sade to the actual CDs the CDs win out by a large margin. There’s just something about the quality of her voice on the CD that doesn’t translate to the MP3s.
I took @Gsuberland’s comments to heart, and tried encoding Aerials at different bitrates. Upon careful listening, I could pick out the 128kbps version. CBR 192, ABR 160, ABR 192, and CBR 320 sound exactly the same as the FLAC to me.
Casually, I can only tell whether it is 56kbps and below, or “good enough”. 64-128 CBR sound good enough to me.
That said, I always use ABR 192 to encode.
I wrote a simple PHP script to do the double-blind encoding, just generates X random numbers in an array, writes that array to a file, and then encodes using the value at each array index as the filename. Random enough for me.
Ripping to a lossy format looks like a misguided optimization, I think.
The hard labor-part is the ripping: you only want to (have your nephew) rip your CDs once in a lifetime and never again, so you want to get the most out of that manual process. If you let EAC rip to FLAC you loose nothing and can always batch-encode all songs to MP3 with whatever lossy settings that suits you and your player. I’ve ripped my entire CD-collection to Flac and use dbpoweramp to batch-transcode it from FLAC to MP3 for storing on my iDevices and it works really great - the entire library can be converted completely unattended overnight. And if a better format comes along in 10 years I’ll just convert to that. Sweet.
Let’s look a a few numbers, based on my own FLAC-converted CD-collection:
My 517 CDs takes up 169 GB. That is 0.33 GB/CD.
A new 2TB harddrive costs $120. That is $0,06/GB.
This means that storing a CD as FLAC costs $0.02.
That’s 2 cents for each album. You’re paying your nephew $1 per album just to rip it. Does it really make sense not to rip it in a lossless format that adds 2% to that price? (Disregarding the admittedly extra hassle of storing both an archived format and also eg an iPod-friendly mp3 format)
First off, saying something like “I […] think I have pretty good ears” is a ridiculous statement. The only way to know is to get your hearing checked. Hearing is not binary (as in hearing vs deaf), and you can easily loose a large spectrum of your hearing before you actually realize it.
Also, I agree with most of the other posts here. Not ripping to FLAC or WAV is just plain idiotic. The cost of the space is nothing compared to the time it would take to re-rip everything. In 5, 10 or 20 years, a few 100G will be the sample space on Google drive or equal to what you carry around on your credit card. Considering the cost and time to rip (not to mention the limited shelf life of physical media), its not worth the risk for literally zero gain.
Anybody who claims that lossless formats are never better is just as ignorant as anyone who thinks that 192KHz is better. Also, if you must have lossy audio, go with a format (ogg) that doesn’t require licensing.
“Not ripping to FLAC or WAV is just plain idiotic. The cost of the space is nothing compared to the time it would take to re-rip everything. In 5, 10 or 20 years,”
In 20 years time I’ll care even less about audio quality. Meanwhile, the likelihood is high that lossless copies will be around for future generations because someone else will have cared enough to make them.
20 years ago I was mostly using computers that didn’t have hard drives, but – as far as music goes – things have been at the ‘good enough’ stage for ages.
There doesn’t seem to be much point with accurate ripping if you are going to throw data accuracy away altogether by using MP3. It is like measuring a piece of metal to 10 decimal places and then cutting it with an axe rather than a laser.
The flaw in the argument against higher resolutions is that it presumes that all equipment is theoretically perfect. Although CDs are theoretically perfect to 20KHz, the entire upper half of the spectrum comes out smeared across time in practice. This is why a cymbal sounds like a cymbal on a high resolution recording and merely an undefined hissing sound on a CD. In any case, there is so much more to music than mere frequency response.
Also, the claim of “inaudible difference” is merely “inaudible on my current equipment”. A bit like assuming 8-bit colour is fine for archival photographs in 1995 because that’s all your monitor can do.
Here’s an interesting aside: if your sound card supports 24-bit, try switching to this mode for playing 16-bit CDs (Windows 7 Control Panel). I am amazed at the difference, even on small computer speakers, that using a more accurate decoding path can make. It is this accuracy that makes the difference — higher resolution audio is easier to decode cleanly than lower resolution, because the unwanted artefacts are further separated from the signal to begin with and do not need such harsh and harmful measures to be rid of them. (BTW iTunes does not seem to support 24-bit playback, you need to use WMP to hear a difference).
Far from the claim that some diehards do not understand sampling theory, those who say “it makes no difference” often do not understand the practical limitations and trade-offs of physical implementation. This error has been going on for years: people (even in the industry) saying that all digital sources are perfect and sound exactly the same.
Actually, the reason to use EAC is because problematic disks produce pops and cracks on ripping. Very audible no matter what encoder you’re using.
I tried to participate, but the sample material is intolerable. Seriously, “…worst-slash-best rock songs of all time…”? No “slash-best”! Definitely worst song of all time and definitely not rock. I can’t take it, the heinousness of the song completely taints any objective evaluation of the sound quality.
Also, unless your CD collection numbers in the tens of thousands, the space usage is insignificant by today’s standards. I have around 500 (perhaps not an “audiophile” size collection, but more than most people I know) and I store them in their natural .wav format. Only consumes ~210 GB. Only slightly more than 10% of the drive upon which they reside.
-
“Uncompromised studio quality” has more to do with conversion flexibility, forwards compatibility than it does some “holy grail” of SQ.
-
Whether or not you can hear the difference depends heavily on a) your equipment (most people’s is subpar) b) the kind of music you’re listening to c) how that music was recorded and mixed and d) whether you’ve trained your ears to be sensitive listeners or not.
and I personally suspect 3) Just like there are “supertasters” out there, there are also “superlisteners” out there.
and 4) “good enough” is not a constant baseline in analog reproduction, and never has been, be it sound or vision. What was “good enough” in 1950 is not “good enough” now, and what’s “good enough” now isn’t going to be “good enough” in 2050.
Who cares if you can hear the difference between “good enough” and “absolutely uncompromised”? If there’s no significant downside to the latter then the relative quality of the former becomes moot, as does the former itself.
I have FLACs in my music library. I don’t give a crap about what bitrate they are or how heavily they’re compressed because I don’t have to. I can encode them to any format I want without worrying, be it MP3 or Apple’s format du jour or Microsoft’s own special format.
I don’t have to worry if in 10 years all those 160kbps files I can’t re-encode sound lackluster compared to the new baseline, like all those 128kbps files from the 90’s.
As tech changes, so does people’s expectations, and I suspect also their ability to perceive relative quality differences. It’s called change, and adaptation. It happens.
The question is not “why FLAC”, the question,as tech and perceptions change over time, is increasingly “why not”?
I did the test, didn’t notice much difference, but have noticed big differences between 128kbps and FLAC before, which your sample songs don’t show. So I think the sample song is flawed. It uses synths heavily, which are already samples, this is a big problem.
A live acoustic recording with multiple instruments would’ve been a better test (and preferably one that’s known to be a good reference recording). One way HQ audio is better is good instrument separation and a sense of 3-d space, you feel more like you’re in the room (whereas with LQ you get a more mushed together sound that sounds like it’s coming from one place). With the recording you’ve selected, this difference can’t be highlighted, even on the best recording, it sounds like it’s coming from the same place.
Like most people here, I store my audio in FLAC, and for similar reasons. When I have 7+TB of space, then a few 100GB for FLAC is a no-brainer. I encode down to OGG normally to put on my phone, and stream from the server when I’m at home (no point duplicating the data) since I have a gigabit network that handles it fine.
I agree that nobody can hear the difference between 160 and 192Kbps, especially VBR, though some people can hear the difference between 160 FBR and lossless FLAC but they are few and far between.
I completely understand the desire to rid oneself of excess physical artifacts, but Jeff, you CAN NOT get rid of the CDs that you ripped. When you rip a CD, you’re shifting the format of the music, but since you still own it, that’s a perfectly reasonable thing to do.
If you then sell or donate the CDs, then you DON’T OWN THE MUSIC ANYMORE.
If you rip then sell, you’re a pirate just as if you’d pulled the album off a torrent.
Think of it like this: Would it be OK to copy the CDs with your burner and then sell them on? Of course not! But if you rip and then sell them on, how is it any different? It’s not!
Or think of this: You buy a CD new and then rip the CD and sell it. The next guy buys it used, listens to it. You’ve just created extra copies of the music. Not cool.
If you really need to minimize the physical artifacts, then just dump the discs on a couple of spindles and throw the cases in the trash.
I agree with your sentiment, Jeff: It is time to do away with media and concern ourselves only with the information.
I agree with posters that flac is the only sensible way to blanket archive the audio information of CDs: Flac (and wave) represent the base data of the CD, and for an insignificant storage cost, we can entirely avoid the debate of the ability of lossy formats to accurately represent the audible information of a CD, and, particularly, to maintain quality through transcoding.
But I’m intrigued that noone has mentioned the other qualities of a CD album - the liner notes. I buy CDs instead of mp3s because, apart from wanting a lossless copy, I feel I am missing out on an important aspect of the album if I just get the mp3s and a cover image jpeg. Sometimes this is misguided (a 2 page booklet), but sometimes it is the artist themselves who has produced the artwork, and it is a significant part of a concept album. Does anyone care about this part of the experience? Has everyone been iTunesd?
Also, I can hear no difference between the sample tracks. (Foobar 24bit to Sennheiser HD280Pro via Audigy front panel. I am 26 though. Hearing going.)
Haven’t bothered to give this a shot, mostly cause I don’t have confidence in my audio analysis. But what about the subconscious effects? It seems entirely plausible to me that listening to compressed tracks, where your mind sometimes fills in the details, might be more mentally taxing, or slightly stressful. I’d love it if anyone knows of any research in the area.
The reason I keep my audio in FLAC is because I want to keep it forever. For me, it’s a question of compatibility.
If you encode in a lossy format today, what will happen in the future when a new encoding format is released, or when devices stop supporting MP3? Your only choice if you wish to update will be to use lossy encoding on your lossy encoding, resulting in further degraded quality. What happens when your awesome new proprietary (e.g. Apple) device only supports AAC5 and not MP7?
Encoding in lossless means that you can encode in any other lossy format at any time in the future.
I have submitted rates; no cheating, I heard that the two top rated samples have different compression level and the other three have different high frequency cut but with almost the same compression setting. used my $35 creative earplug headphones, with my wife snoring beside, heavy trucks 50 ft. away, and boeing airplanes 1000 ft. above about landing
Shenanigans! This can’t possibly be a valid test, proper audio tests must be TEH DOUBEL BLINED!
Your influence is clear in your choice of cheese, Everyone knows which cheese is the best and which is the worst in that list. You even tried to cover it up by REVERSING TEH ORDER! Best Cheese = worst quality!
New test, this time use boy band member names, no-one likes boy bands. Or Vegetables! Animals! No, wait, Animals have Order, breeds of rat!