Quantitative comparison of 16 vs 24 bit Beatles remasters

more people would be able to play back 44.1 24 bit, than the 192...

 

Jorge,

I think this is one of those "lost in translation" issues - my signature is a joke, and when I say "good system" I don't mean to imply it's the best, it's just my best system (as opposed to my computer speakers or ipod). The .3 dB volume difference is very small. I'm sure there are tests out there about where humans can hear volume differences, and I'll bet it's not much bigger than 0.3 dB! I listened for the timbre, soundstage, and overall presentation of both sources (16 and 24) and heard no noticeable difference.

I reiterate that this thread is meant to be primarily about the *quantitative* differences - these are indisputable, assuming my methods are valid (they are ). What I am willing to concede is that the quantitative comparison is *not* necessarily the full story because of the fact that difference DACs may treat different bit lengths and sampling rates differently (I could be way overstating this). My DAC, the rather affordable but in my opnion awesome Cambridge DAC Magic upsamples incoming streams to 192/24 so it does not surprise me (because I believe the upsampling is being done intelligently) that 44.1/16 and 44.1/24 sound identical.

 

Maybe they couldn't get a USB big enough for 192/96k files and mp3s. To have included two or three would probably have driven up the price. Not only that, *most* average people may not be able to handle such files on their computers or other playback devices that accept USB. And, a lot of damage can be done if people do not have clean SRC software.

 

Damn those majority types!

Seriously, thanks. I just figured this item was not geared towards the everyman.

 

Just use a bigger apple.

Seriously, you may be right. As I said above, I figured item was not geared to the masses.

 

You're kidding, right? I can find 64 gig USB flash drives at Newegg. On my computer the Beatles USB shows up as 9.18 gig. I upsampled Taxman from 44.1/24 to 192/24 and it increased in size by about 2.5 times. If that's a rough estimate for the rest of the discs they'd all fit on 32 gig flash drive with several gigs to spare.

32 gig USB flash drive for $65: http://www.newegg.com/Product/Product.aspx?Item=N82E16820171402

 

You got any idea how much space you need for 200+ at 199KHz/24-bit? It wouldn't work. Taxman is less than three minutes. Remember all those longer cuts.

 

There are also many tracks that are smaller. The only way to know will be to do the math not speculate!

 

lost in translations i'm feeling very stupid after yours clarifications Leigh. (really lol)
ok, this is the very important part in your post for me, don't need to quote more. i agree with you and understand.

cheers, Merry Christmas and Happy New Year!

 

I did some math.

 

How about some numbers? ie: show your work! if not it doesn't count

 

Want to make a bet?

So what? If they were 192/24 they'd be roughly 2.5 times larger than a 44.1/24 file. 2.5 * 9.18 gigs = 22.95 gigs. Want me to convert them all to 192/24 just to prove you're wrong? I'll do it.

 

Good Little Space-Heater, Though

From the OPs profile:

The veritable Dynaco ST-70 with JJ KT-77 tubes​

I don't believe the Old Dyna 70 has anything like 24 bits of resolution. Nice sounding amp but the combined noise floor and innate lack of resolution of that piece of gear leaves the results of your listening test moot.

 

Grant, I just converted Please Please Me and With The Beatles from 44.1/24 to 192/24 using Goldwave.

Please Please Me:
44.1/24 = 408 mb
192/24 = 979 mb
Increase of 2.39 times


With The Beatles
44.1/24 = 465 mb
192/24 = 1009 mb
Increase of 2.34 times.

Amazing, very close to 2.5 times larger like I said it would be. All of the albums would easily fit on a 32 gig flash drive with several gigs to spare. The rest of the albums are still converting, I'll post again when it's finished.

 

You didn't read the rest of the description of the amp, apparently. I have updated the power supply (SDS Labs) and replaced the driver board with the Triode Electronics board. I have spent 15 years with various incarnations of this amplifier and it is plenty resolving and has a very low noise floor - it is dead silent in my listening chair, low level hum and pink/brown noise with my ear pressed to the ProAcs. The only thing left in the amplifier that is original - not even the wiring - are the output and power transformers and the chassis. It is probably not even correct to call it a ST-70.

I did a listening test with a tone sweep through the amp and began to clamp my hands over my ears at 16 KHz - meaning either that's where the amp or preamp rolls off or that's where my ears roll off. It is a plenty nice sounding amplifier, and your dismissal of my test as moot is.... moot.

Plus, I don't trust my ears, or yours. I do trust my quantitative analysis, however, which is what this thread is primarily about.

 

Plenty nice isn't the point, 24 bits is.

I've seen plenty of tests of gear claiming 24 bit resolution where they delivered more like 19—in addition to everything else, 24 bits also describes the noise floor. I've only read one subjective review where someone had digital playback gear for the apple stem that approached state of the art and that was a rave. I won't deny that the potential sonic virtues of 24 bit files will be beyond all but a very few listeners. But that's not the same thing as "therefore, there is no point in having 24 bit files of this material."

 

I can hear some difference, mostly a little clearer sounding and also subtle details come out more in the USB 24 bit files. Now I wouldn't say the sound quality is vastly superior to the cd versions, but there is a slight higher quality sound with the USB tracks.

 

The types of differences claimed by the owner of the "state of the art gear" are far too great to not be heard by someone with modest gear. Night and day and all that. Or, maybe his playback system is really bad for 16/44.1, and the difference reflects this poorly performing stage.

I have looked at the difference between the files. They are numerically consistent with what you'd expect with downsampled data and decent src software and dithering. The original tapes of the Beatles and the equipment used in the mid-to-late 60s are hardly "state of the art" and you can't polish a turd and all that, and I am highly skeptical of grandiose claims of night-and-day differences in files which are different on average by -60 dB where much of the energy is above 16 kHz where there isn't much useful energy anyway.

I started this thread to show the quantitative differences for those who are interested. If you want to talk about the (real or imagined) virtues of 24 bit audio reproduction in general, that is for another thread.

 

Thanks for the info. Would fit nicely. Plus you showed your work! Maybe Grant was using the "new math"?

 

A turd? What exactly is considered inferior with the state of EMI/Beatles recording in the mid-to-late 60s?

If anything has come along to better the fidelity of a Studer J37 1" 4-track fed through a tube-driven EMI console in the past 50 years, I'd like to know.

 

Sorry, but I just don't find that sound to be anything special compared to much of the gear which followed. The fidelity of much of the Beatles catalog sounds pretty so-so to my ears, nothing terribly special, compared to much stuff recorded on modest modern digital gear.

Concerning frequency response, I think you'll find that Studer rolled off around 15-16 KHz.

 

From what I've read that's all they have. The tapes were transfered to digital at 24/192 but that's just the flat files. When they mastered the albums they used mostly analog gear so when they re-recorded the eq'd audio they did it at 24/44.1. That means the highest resolution for the mastered material is 24/44.1.

 

Unfortunately, that is correct.
(That extra conversion stage bugs me no end)

 

Looks like my math was off a little, it came in at only 18.7 gigs for all 226 tracks at 192/24 resolution.

 

Better than sample rate conversion old boy,...192 down to 44.1 is nasty maths. Eqing from a 192 source in the analogue domain and converting that result at 44.1 24 bit is the most sensible course of action....no machine doing the src