The Beatles released their "remastered" catalogue on a USB drive in 24-bit 44.1kHz FLAC.
Lots of fans claim the 24-bit versions sound amazingly better than the 16-bit CDs, e.g. see the Amazon reviews...
http://www.amazon.com/Beatles-USB/dp/B002VH7P4O (http://www.amazon.com/Beatles-USB/dp/B002VH7P4O)
...but we all know how reliable such subjective impressions are!
However, it made me wonder if, just maybe, they didn't apply peak limiting to the 24-bit versions, hence giving them greater dynamic range than the CDs I think this is unlikely - your typical Amazon purchaser (who can't even play a FLAC file without jumping through hoops!) would probably think a quieter version without peak limiting sounded worse. But still - does anyone have one of these - or even a 30 second snippet of one of the tracks - that they could share?
It would be interesting to compare with the 2009 remastered CDs.
(I'm not sure it would make me buy it - why not put both the stereo and mono versions onto the USB memory stick?)
Cheers,
David.
I have such, which track would you like a 30 second clip from?
One of the discs had no ID Tags, and all albums had [24 bit remastered] in the name (which I am sure the majority of people would get rid of)
Brilliant.
How about the first 30 seconds of the very first song - I saw her standing there?
Many thanks.
Cheers,
David.
I just recently read that these are 24-bit files. When I first read about the USB version, I assumed it was identical to the CD version. But I'm disappointed that the USB version is much more expansive than the CD stereo box.
I'm also interested in a comparison and hope that I won't hear any difference. But maybe some of the hidden messages will be clearer in the 24-bit versions
The Beatles released their "remastered" catalogue on a USB drive in 24-bit 44.1kHz FLAC.
[snip]
It would be interesting to compare with the 2009 remastered CDs.
A chap called Phil Leigh did such a comparison using Audio Diff Maker (see post #5 in original thread on the Slim Devices forum (http://forums.slimdevices.com/showthread.php?t=72852)).
He found that the differences are at very low levels (requiring 50dB of boost to become audible), and only at low and high frequencies.
My take on this is that the differences he found are simply due to the noise-shaped dither added to the 16 bit versions when preparing the CD releases.
Thanks cliveb. (There's some rubbish talked in that thread too, isn't there?)
So, nothing useful to be gained. I'd still like to take a look though - to see how small the difference is.
Cheers,
David.
P.S. I find it amusing and depressing that there's all this talk of how much better it does sound at 24-bits, and could sound at a higher sample rate - as if 44.1kHz 16-bits was the limiting factor on these recordings! For goodness sake - have none of these people listened to The Beatles Love CD? Do none of them understand how many "bits" of difference peak limiting makes? Just shows that Apple can take the same masters, release them at 88.2kHz, and charge some people all over again. Which probably means we won't see versions drawing on the original 4-track tapes in my life time (and I'm only in my thirties!) because they can make extra money with far less effort by selling 24bit and 88.2kHz versions.
Thanks cliveb. (There's some rubbish talked in that thread too, isn't there?)
It's the "Audiophile" section of the forum - what do you expect? :-)
Actually the Audiophile part of the Slim Devices forum is unusual: an approximate 50:50 mix of down to earth types and loonies. Amazingly it rarely descends into full-on flame wars. Sometimes Sean Adams (the Squeezebox designer) pops in with some very insightful objective comments. I rather like the place.
Are the stereo mixes really that bad, or are you guys only hopeless romantics?
I want this.
I am interested in 24 bit, but only if it has mono too. I was pretty disappointed with the stereo mixes and would definitely like to se a remix, otherwise I'm sticking to mono for just about everything, with some exceptions. If there's a 24/96 or something similar, especially with remixed stereo, I'll buy it all over again.
I have the USB box set and have borrowed the Stereo Box set for comparison purposes. So far, ACTIVELY SEEKING DIFFERENCES, I can find none. Never mind ABX testing, I can't convince myself they're different when I'm peeking.
Given I didn't really expect much here, but I WAS kind of hoping that we'd be getting the non peak-limited versions on the USB. It looks from other posts that these are most certainly cut from the same cloth as what came out back in September.
That said, I DO prefer the Stereo mixes over the Mono box, which I also purchased. While I understand they didn't put ANY additional dynamic range compression on the mono remasters, I don't think that's the case when they cut the original masters. The early LPs seem to have more "pop" in stereo than in mono.
Well that's disappointing that there's no audible difference, but it makes me feel better about purchasing the CDs. It's not that hard to rip, and I think it's ridiculous for them to charge the same amount (or more) for files as they do the physical media. The mono box is so beautifully packaged, it ALMOST makes me feel like I got my money's worth. The stereo CDs less so, but it's still nice to actually get the discs for your money.
I was trying to convince myself to get the 24 bit files, and I'm glad I didn't. The 44.1 kHz is actually what stopped me; I can hardly believe they didn't release it at 88.2 or 96...it would have been SO easy. I'm sure that'll be another release.
I was trying to convince myself to get the 24 bit files, and I'm glad I didn't. The 44.1 kHz is actually what stopped me; I can hardly believe they didn't release it at 88.2 or 96...it would have been SO easy. I'm sure that'll be another release.
I'm sure your cat misses the ultrasonics.
I really hope there are no further releases of the catalog. The very idea of releasing the 24 bit files three months after the stereo box set (and purposefully not announcing it earlier) was clearly intended to get people like me to double-dip. The liner notes of the Mono Box Set ironically make a point of reminding us that the Beatles frowned upon putting singles on albums, being that it would force fans to buy the same songs twice. So much for that principle.
My post was deleted... why???
Not a very nice welcome I must say.
Your post was moved to here (http://www.hydrogenaudio.org/forums/index.php?showtopic=77756&hl=), apparently because you made a statement about audible quality differences without backing it up. (See Terms of Service #8)
My post was deleted... why???
Not a very nice welcome I must say.
Your post was moved to here (http://www.hydrogenaudio.org/forums/index.php?showtopic=77756&hl=), apparently because you made a statement about audible quality differences without backing it up. (See Terms of Service #8)
Uhmm I see...
I said that the USB and the CDs sound the same to my ears... What's wrong with that? or what am I supposed to do to prove that??
Is this some kind of scientific forum where everything must be put in measurable terms??
pretty much, yes. This is about the only place where I can come get info on audio and at the same time keep my sanity.
I really hope there are no further releases of the catalog. The very idea of releasing the 24 bit files three months after the stereo box set (and purposefully not announcing it earlier) was clearly intended to get people like me to double-dip.
I think so too. I ended up buying the stereo remastered CD box set but I am not about to plop down $200 for the digital files on an Apple. I went with the stereo remastered set as it came with more content and was about half the price of the mono box set. That is all I needed; every Beatles album on stereo CD so that I can rip them for playback on my computer and portable device. I still would have went with the stereo CDs if the stereo box set, mono box set, and USB Apple were all released at once. I think $200 is way too much to spend on a USB storage device for digital files. I want something to hold in my hand other than a plastic apple. I want to read through the liners, have a nice big box sitting on my shelf, etc. Having the FLAC files would be very convenient but I don't think it is worth the extra ~$50 for a few hours of my time.
It would be interesting to conduct a few ABX tests to determine if I can actually hear a difference between the 24-bit files and 16-bit files. I doubt I can. I have a 24-bit 44.1KHz FLAC file that I converted to ALAC (24-bit) and ALAC (16-bit). I cannot differentiate between the two ALAC files for the life of me. I cannot even effectively ABX the 24-bit ALAC track and a 16-bit lossy version (at -V 2/-q0.5/192kbps VBR constrained Lame/Nero/iTunes).
I'd also be interested if low-volume parts are ABX-able if you boost the volume, e.g., the squeaking chair at the end of "A Day in the Life".
In other words, if the original tapes provide more than 16 bit of dynamics. Maybe we can discover some new hidden messages in the flac version .
I said that the USB and the CDs sound the same to my ears...
You said, "
sometimes I feel a little more 'air' or spaciousness in songs with quiet moments (in the USB), but then I compare them with the CDs and they are
basically the same." So no, you you aren't saying they sound the same to your ears at all. You're saying you can tell the difference and chose to use subjective fluff to describe the difference earning your post its own cozy little space in our recycle bin.
Quite impressive for a first post!
Maybe someone should have pointed out the problem in the thread. I was too lazy to - sorry. I assumed it wouldn't take more than five seconds for someone to point to TOS 8.
The bigger problem for me was the statement "I'm sure there are situations where 24-bit files sound much better than their 16-bit versions", and pdg answered this correctly with "The only cases in which the difference is significant is when they were mastered differently. If the 24 bit version is properly converted to 16 bits then the difference is virtually inaudible."
On the 1987 CD, that squeaking chair has so much tape hiss that 14 bits would still be more than enough. I have to admit that I haven't listened to the Pepper 2009 remaster that carefully yet!
Cheers,
David.
pdg answered this correctly with "The only cases in which the difference is significant is when they were mastered differently. If the 24 bit version is properly converted to 16 bits then the difference is virtually inaudible."
So where's the TOS#8 police on this one? What does "virtually inaudible" mean? I'm pretty confident that you'd hear a difference between 16 and 24 bit playback if I recorded or mastered at -60 dBFS.
Looking back over ten years or so of following HA, and r3mix before that, I only recall one instance where one person could reliably differentiate 16 bits from 24 bits, on specific material.
Of course, anyone can make up any bizarre situation to prove a point.
I was trying to convince myself to get the 24 bit files, and I'm glad I didn't. The 44.1 kHz is actually what stopped me; I can hardly believe they didn't release it at 88.2 or 96...it would have been SO easy. I'm sure that'll be another release.
I'm sure your cat misses the ultrasonics.
Hey, I just think he wants to get his money's worth. We're payin' by the bit here, after all
Okay, I just made the mistake of looking over at the Amazon reviews for the USB box set. Apparently according to everyone there you can totally and easily tell the difference between 24 and 16 bit. Either I have absolutely terrible hearing or there's a lot of justification going on. It's a shame as I think this will convince a lot of people to plunk down $220 they could be spending on lots of other worthier things. Ah well.
Agreed about the comments on Amazon.
I've compared the two properly now (sent to me via PM - thank you!). I've attached spectral plots to show one moment from I Saw Her Standing There. I chose a comparatively quiet bit (i.e. between drum beats, rather than on the beat!) so that the contribution of 16-bit noise shaped dither above 20kHz would be visible. You have to look carefully though!
In essence, the two are the same, except for the 16-bit dither, and a 0.2dB level difference (24-bit is louder - neither hits digital full scale). I "corrected" this difference by attenuating the 24-bit version by 0.2dB before subtracting one file from the other.
Strange thing is, they're not actually identical. Even when the gain of the 24-bit file it changed to match that of the 16-bit file, it doesn't match throughout. There's another ~0.01dB gain discrepancy creeps in at certain moments - not enough to be audible in itself, but it stops the two files from nulling out completely during those moments. For one moment only (after the quieter "1, 2, 3, 4" - at the exact moment the track really starts) the gain difference is greater - about -50dB down.
Otherwise the difference is -64dB down, and consists of the noise shaped dither only.
It looks to me like they ran the peak limiting separately for the 24-bit and 16-bit versions - using near-identical settings, but getting a non-bit-identical output.
So sadly there's no greater dynamic range in these 24-bit versions - in fact there's nothing of use that isn't already there in the 16-bit version.
If you try to ABX (good luck! ), be aware that a 0.2dB level difference might just be audible under some circumstances - certainly far more audible than any other difference between 16-bit and 24-bit.
Cheers,
David.
> Apparently according to everyone there you can totally and easily tell the difference between 24 and 16 bit.
Those are people who just spent a lot of money on the USB apple and have to justify their purchases. So of course they're going to hear HUGE differences between them.
Nothing will affect your hearing like spending $240.
> Apparently according to everyone there you can totally and easily tell the difference between 24 and 16 bit.
Those are people who just spent a lot of money on the USB apple and have to justify their purchases. So of course they're going to hear HUGE differences between them.
Nothing will affect your hearing like spending $240.
Post-puchase rationalization?
http://en.wikipedia.org/wiki/Post-purchase_rationalization (http://en.wikipedia.org/wiki/Post-purchase_rationalization)
If anybody could hook me up with the samples 2BDecided received I'd appreciate it (and hopefully you'll soon appreciate it too!)
I uploaded a new sample. All new releases (mono, stereo and 24-bit) and one old release (1987 CD) are included:
http://www.hydrogenaudio.org/forums/index....showtopic=77836 (http://www.hydrogenaudio.org/forums/index.php?showtopic=77836)
Listened to the samples, and honestly couldn't tell a difference between the 2009 stereo cd remaster and that 24-bit flac. Not that I have great hearing anyway, but maybe there just really isn't an audible difference (for humans anyway, I'm sure bats love the new versions!).
Just an FYI, the "bat" thing has do do with the sample rate (which is the same as the CD version), not the bit depth.
Ahh yes, blunder on my part. Good catch.
There's a somewhat enjoyable back-and-forth on this topic in one of the Amazon reviews - the only one who gave the USB box 2 stars.
My uncles purchased the box set. I compared the 24-bit FLAC files to my 16-bit ALAC files by completing a couple blind ABX tests with some tracks. I could not determine a difference between the two. I even compared Nero AAC lossy versions (encoded at -q0.5 with version 1.5.3.0) from the 44.1/16-bit masters to the 24-bit FLAC files. I could not pass the ABX tests to save my life (either my my headphones or my semi-calibrated, about as good as it will ever get, home theater system). The reviews on Amazon.com make for a humorous read though. No point in arguing with people like that. They are the same type of people who believe that $300 HDMI cables produce superior quality, a $1000 ethernet cable gives a superior advantage when streaming audio/video content, freezing CDs increase their sound quality, and that the Sun orbits the Earth.
There's a somewhat enjoyable back-and-forth on this topic in one of the Amazon reviews - the only one who gave the USB box 2 stars.
Thanks for pointing that out.
On the Amazon Beatles message boards, people seem to think that more bits = better, and higher sample rate = better.
e.g. the Love DVD being "only" 96kHz - but they had 192kHz in the studio - what are we missing out on?!
It's like putting more pixels on point-and-shoot digital cameras for people to make 6x4 prints - though in that case more pixels can sometimes reduce the quality!
I blame dumb marketing and dumb people. More doesn't necessarily mean better!
Cheers,
David.
It's like putting more pixels on point-and-shoot digital cameras for people to make 6x4 prints - though in that case more pixels can sometimes reduce the quality!
At the risk of derailing the thread, I kind of take issue with that. In many tech/photo sites, forums and blogs, the "megapixel race" fad has turned now to the "anti-megapixel-race" one. Now that DSLRs are extremely popular and accessible, everyone's an expert (not attacking you here, just talking in general) educating the unwashed masses with their lowly point-and-shoots about the evils of more minute pixels. I am no expert either, but I take my info--like here at HA--from them (which 98% of the time are optical/electrical engineers who are hobby photographers, rather than just pro photographers).
Other than in poor-image-processing cameras, more pixels per se don't cause more image-level (photon-)noise. When looked at 100% (1:1 mapping) on the screen, yes, there's more noise, but there's also more pixels. The objects in the picture will be bigger too. So that extra noise is offset by the extra pixels themselves, and at the image level, they're balanced. This can also make the noise less coarse.
And besides all that, there are actual advantages to having more pixels,
even when you don't gain more resolution (e.g. if you're resolution-limited by the lens). You can do leveling, rotating, perspective correction and other such interpolating procedures, without
losing as much resolution.
There might be the point where pixels are so small that electrical noise (not photon or shot noise) from the more complex sensor will again offset this balance, but that is electrical noise and it can be minimized with technology advances, it's not, physically, inherent to small pixels. Furthermore, there's no evidence that current cameras (even point-and-shoot ones) have reached that point. The horrible IQ of
some super-high-megapixel cameras is not because of the pixel count, but because of the horrible noise-reducing algorithms. When RAW images from high-MP cameras have been compared, there's been no evidence of more noise in the whole image that I know of.
Crappy pictures from small cameras and cellphones are not because of the small pixels, but because of the tiny sensors. A 1 MP cellphone picture will be as crappy or more than a 5 MP one.
See here (http://forums.dpreview.com/forums/read.asp?forum=1010&message=32739559) for some images. Also, check John Sheehy's and ljfinger's posts in that thread. Sheehy is one of the experts on analyzing RAW data, he has posted examples many times there, but I don't have the direct links.
@ andy_o,
Yes, I understand - hence my use of the word "sometimes"!*
Otherwise the analogy was a good one:
For photos, for a 6x4 print, there is a point where more pixels will make no visible difference to the human eye.
For audio, for a given replay volume, there is a point where more bits will make no audible difference to the human ear.
For photos, where the lens in front of the sensor can't focus anywhere near down to the individual pixel level, it may become the limiting factor, rather than the human eye or pixel count.
For audio, where the analogue tape feeding the converter has typical amounts of noise, it may become the limiting factor, rather than the human ear or bitdepth.
And besides all that, there are actual advantages to having more pixels, even when you don't gain more resolution (e.g. if you're resolution-limited by the lens). You can do leveling, rotating, perspective correction and other such interpolating procedures, without losing as much resolution.
Assuming you're already a good factor beyond the resolution of the lens (simple Nyquist isn't good enough for imaging, because you can't use sync reconstruction filters - 2x or 4x Nyquist sampling is needed), this isn't true. If the rotation isn't sharp at the pixel level, then upscale, rotate, downscale - no extra loss - no need for more pixels in the original sensor. If you're not already 4x oversampled beyond Nyquist for the lens, then more pixels could be a little useful. Problem is, there are probably far more useful things that don't go into a single "ours is bigger" number on the box.
I agree that, for the digital equivalent of medium format, we need far more pixels. But for point-and-shoot? Come on - no one will ever see the difference if we go beyond what we have now. Just like audio.
I followed your link. I'm familiar with dpreview from picking cameras in the past, and have never seen such a "religious" debate!
Cheers,
David.
P.S. * = more pixels can only be benign in the way you suggest if the gap between pixels scales in the same manner as the pixels themselves, and if camera designers don't overdo the noise reduction due to concern over the appearance at 1:1. Hence, as I said, "sometimes"!
While your points about more pixels increasing the noise level per pixel, but decreasing it per image are correct - we are quickly approaching the point where digital sensors are diffraction-limited at sane apertures, even on large-sensor DSLRs.
You know, I was gonna include the "sometimes" into my thing, as a sort of caveat. True story.
I understand what you're saying, but also part of my point is that there's no evidence that current point-and-shoots have reached the limit. I'm not saying that 10 or 12 MP isn't enough, just saying that the outcry every time a new MP jump occurs (or the great excitement when it goes in reverse, like the Canon cameras) is largely unwarranted. For one, 4x the lens' resolution is still very far off. Noise as you saw in the pics, is not raised in the whole image, and there's still more detail at least a low enough ISO.
About the "religious" debate, you'd be surprised how much similar are these kinds of "debates" on photo forums, as they are on these audio forums. Basically, you get engineers and science-types vs. self-important "pro" photographers that pretend to know more just because their pictures are prettier. You make an argument, sound as it may be, and someone invariably comes up with the "so, where is your gallery?" misdirection.
Agreed about the comments on Amazon.
I've compared the two properly now (sent to me via PM - thank you!). I've attached spectral plots to show one moment from I Saw Her Standing There. I chose a comparatively quiet bit (i.e. between drum beats, rather than on the beat!) so that the contribution of 16-bit noise shaped dither above 20kHz would be visible. You have to look carefully though!
In essence, the two are the same, except for the 16-bit dither, and a 0.2dB level difference (24-bit is louder - neither hits digital full scale). I "corrected" this difference by attenuating the 24-bit version by 0.2dB before subtracting one file from the other.
Strange thing is, they're not actually identical. Even when the gain of the 24-bit file it changed to match that of the 16-bit file, it doesn't match throughout. There's another ~0.01dB gain discrepancy creeps in at certain moments - not enough to be audible in itself, but it stops the two files from nulling out completely during those moments. For one moment only (after the quieter "1, 2, 3, 4" - at the exact moment the track really starts) the gain difference is greater - about -50dB down.
Otherwise the difference is -64dB down, and consists of the noise shaped dither only.
It looks to me like they ran the peak limiting separately for the 24-bit and 16-bit versions - using near-identical settings, but getting a non-bit-identical output.
So sadly there's no greater dynamic range in these 24-bit versions - in fact there's nothing of use that isn't already there in the 16-bit version.
If you try to ABX (good luck! ), be aware that a 0.2dB level difference might just be audible under some circumstances - certainly far more audible than any other difference between 16-bit and 24-bit.
Cheers,
David.
Thanks for the post. I haven't seen any AD/DA converters ever hit the maximum for the theoretical limits of 16 bit or 24 bit. They are getting closer, but they aren't there from a linear standpoint. They used the Prism Converters, but a lot of the studios, including Abbey Road, and others are now using DAD AD/DA converters. I don't think they are going to remaster them all over again to see if there is a difference in the converters and maybe put out 24/192 or DSD128.
One question I do have is have you listened to them and what equipment did you use in terms of DAC, cables, speakers, just for curiosity sake?
I have checked out other recordings from HD Tracks and have noticed a difference in their different conversions as I might be able to turn up the higher res files a little louder. I usually only listen to my content at around 80 to 90dB, but sometimes, i do play them a little louder where the peaks will hit over 95dB. The problem with my system is my DAC converts 16 bit to 24/96 automatically and it runs it through a Proprietary Apodising filter and I've compared 16 bit recordings to 24/96 and the 24/96 were always better than the 16 bit up sampled running through the filter. But the up sampled versions are better than not being up sampled. At least on the equipment I've have used.
Thanks again.
For those interested, I am using a Mac OS X computer using Pure Music and I can set the preferences so it plays back in mono. So you might have to use a 3rd party s/w product if you want to enjoy it in mono.
Have fun!
Hi RichDavis.
I didn't do any meaningful audible comparison of these files. I did spend some time comparing some original UK vinyl with the CD re-issues because I expected to hear, and did hear, a difference. You may call it closed minded, but I don't have any expectation of being able to hear differences ~90dB down. I started this thread because I was hoping to find that the 24-bit Apple USB was more faithful to the original releases in terms of dynamics, but I was disappointed. In this context, it's not really a big issue for me - I listen to and enjoy the new CDs. I like some of the EQ and "fixing" decisions. I still enjoy the original vinyl.
Plenty of other threads around similar topics...
http://www.hydrogenaudio.org/forums/index....showtopic=97914 (http://www.hydrogenaudio.org/forums/index.php?showtopic=97914)
http://www.hydrogenaudio.org/forums/index....showtopic=97241 (http://www.hydrogenaudio.org/forums/index.php?showtopic=97241)
http://www.hydrogenaudio.org/forums/index....showtopic=84869 (http://www.hydrogenaudio.org/forums/index.php?showtopic=84869)
HA is an objectivist audio site. If you want to describe differences you hear without backing your statements up with evidence to prove that you actually hear a difference, you will find such posts tolerated on many other forums but not on HA. It was created as a strictly evidence based site.
Cheers,
David.
P.S. I don't think I understand your point about mono. I wouldn't convert the stereo mixes into mono by summing them; I'd listen to the mono mixes. They have the same content on both channels, so I don't see what change you'd need to make on your PC. For an authentic experience I guess you could mute one channel - is that what you meant? Alternatively you could disconnect one speaker, assuming your amplifier won't mind this. Some power amps, especially valves ones, don't like being run open circuit like that.
HA is an objectivist audio site.
Speaking as someone with a bit of a beef regarding Ayn Rand and her disciples, I'd prefer the term "objectively minded" or "scientific", but I get what you mean.
This MIT Cables guy looks like one more kook selling snake oil. Haven't seen anything even remotely persuasive.
I had a look at the mis-tagged files, and it looks as if Yellow Submarine had a nice error in its tags. "Pepperland Laid Waist", hah. And the filename was already correct, I think.
I had a look at the mis-tagged files, and it looks as if Yellow Submarine had a nice error in its tags. "Pepperland Laid Waist", hah. And the filename was already correct, I think.
True... and Rubber Soul had no tags at all, at least on my USB.