HDX rip better than iTunes rip?

Gosh. I wish you were right.

Doug

FWIW I too have a degree in Computing and over 20 years in the business as development, consulting and tech support and I'm firmly in the camp of "To Err is human, but it takes a computer to really f**k things up."

There are loads of places where very minor differences which may result in the same data but slightly different files may occur, including:

1. Reading pits off a disc is an analogue process.

2. Where does the pit start and stop? The much trumpeted new blue laser CDs from Sony which are meant to have higher sound quality (as yet uncorroborated) are only better (if so) because the pits themselves are better written by the blue laser! In fact it's simply that the pits are burnt into the material more quickly and do not leave 'wrinkled' edges.

3. Once read, the interpretation and encoding of those waveforms into bits (the digital part) is down to the controlling software which can cause minor changes in the eventual digital code, so minor that any file with decent checksums will be interpreted correctly, but a data stream such as a music file may not be.

Microsoft have more important things to do than work out if there's jitter in music streams. They're happy the thing works at all. Anyway, most leading technical spheres are led by small bands of specialists, not mega corporations who are more interested in reliability than outright performance. Take almost any sphere of endeavour you please, and you'll find some little specialist company doing amazing things which leave the big boys cold.

Microsoft bought Pacific Microsonics which was way ahead of things (HDCD) and now ... is nowhere to be seen, technology and all.

On the subject of buffering, one would thinhk that once you've buffered the stream in a DAC and reclocked it, it should make absolutely no difference what you used as a source, be it a computer with a USB out, a Rega Apollo or a Panasonic bluray player. However, using any of the above into a Chord DAC64 (or the newer QBD76) on its 4 second buffer yields different results in each and every case, and one can never tell which way the pendulum will swing, but USB is almost always the loser, probably thanks to its high levels of jitter. The only worse connection is HDMI (an order worse).

In the HIFICRITIC article mentioned earlier, the journalist mentioned that they'd managed to get the HDX rip off and put it onto a USB pen drive, along with other rips of the same track using variously iTunes, DB Poweramp, FreeRip and EAC. The USB stick was plugged into the HDX and all the tracks were played directly off the stick. I find it interesting that Malcolm still found the difference of the HDX rip to be so high considering that the USB interface is in the way here. However, I am perfectly willing to believe him when he says he didn't know which rip was being played at any stage in the process.

I also find it interesting that he had such differing results with different drives and configurations of machines.

The HDX is an optimised bit of kit where attention to detail has been paid to get the best rips and best replay from a computer source as possible. There are few who have spent the time learning how this can be done while aplying Naim's unique expertise in power supply management. Why do you find it so difficult to believe they can wring extra performance from the unit than you can (or Microsoft for that matter)? That said, if you spent enough time on an XP machine and designed a bespoke power supply, chose the most appropriate disc drive and optimised XP (HDX is embedded XP), the drivers and wrote your own ripping engine, and put it in a properly designed vibration reducing case, you might eventually get similar results in the ripping section. The DAC section would take a lot more work obviously.

quote:

Originally posted by Doug Graham:
Gosh. I wish you were right.

What a tease that comment is! Says everything, whilst saying nothing, much like most MPs.

I was just speculating. I'd be very surprised if the *current* average gate price is 2K, I'd probably speculate it would be closer to 1K when you factor in all the 5i, 5x, 122x, 150x, Nait5, Headline, Stagelines etc.

Plus I don't know if Fraim / Hiline / Powerlines are serial numbered.

But you knew all that, and I know you knew that. And I still bit!

Cheers

John

John,

If Naim's numbers were as high as you'd said, Doug really would have a Bentley in his parking spot.

quote:

Originally posted by Frank Abela:

Microsoft have more important things to do than work out if there's jitter in music streams. They're happy the thing works at all. Anyway, most leading technical spheres are led by small bands of specialists, not mega corporations who are more interested in reliability than outright performance. Take almost any sphere of endeavour you please, and you'll find some little specialist company doing amazing things which leave the big boys cold.

I thought we had agreed that jitter was a playback issue, not one connected to the issue of whether music files could sound different (when played back through identical equipment) even though they appeared to be (and could be proven to be) identical byte for byte.

The issue (I think) we are discussing is whether Naim has actually come up with a way to rip a CD so that, even though the resulting file may be identical ( using all generally known tests and criteria) to a file ripped using another software program, it could still sound different (better) than that other (apparently identical) file. If that is the case, they must have somehow discovered some property of digital files that is capable of affecting playback and that has eluded the entire computer industry and every researcher in the entire world, including organizations that spend billions on computer experts and research. That is what I find very hard to believe.

Frankly, if they really have managed to develop such knowledge, they really ought to patent it as it is likely to be extremely valuable indeed with possible applications in a wide range of fields outside of audio, and not taking all steps to protect their ownership rights might be construed as negligent.

quote:

a way to rip a CD so that, even though the resulting file may be identical ( using all generally known tests and criteria) to a file ripped using another software program, it could still sound different (better) than that other (apparently identical) file

I think the "identical" part may be the key here–simply matching the end result against a database of known rip parameters (versus an individual bit-by-bit comparison) is akin to the differences between a land survey and counting each blade of grass by hand.

On the other hand–what if the Naim rip is "wrong", but sounds better than the accurate rip? How could one explain the perceived result versus the expected data?

All worth thinking about....

quote:

Originally posted by David Dever:

I think the "identical" part may be the key here–simply matching the end result against a database of known rip parameters (versus an individual bit-by-bit comparison) is akin to the differences between a land survey and counting each blade of grass by hand.

Fortunately computers are very good with integer numbers, if they weren't we wouldn't be having this debate, there would be no IT industry and no internet. People have checked files using checksums and bit by bit comparisons, if the files are shown to be identical then they are.

quote:

Originally posted by Frank Abela:
If Naim's numbers were as high as you'd said, Doug really would have a Bentley in his parking spot.

I know, I was just having some fun back.

And it really pointed out how crazy Microsofts figures are, and how far from average we are that we put some effort, and money, into our passion of music and its reproduction.

TTFN

John

quote:

Fortunately computers are very good with integer numbers, if they weren't we wouldn't be having this debate, there would be no IT industry and no internet. People have checked files using checksums and bit by bit comparisons, if the files are shown to be identical then they are.

I remember well the days when playback quality on a hard-disk recording system depended on the drive one played the data back from–and though our technology has gotten significantly better over the years, the basic constraints remain (at perhaps smaller levels).

The problem with early hard disc recording software was that you had to contend with slower processors, much less system RAM and slower hard drives. The main problem was making sure you didn't run out of data so people would spec the fastest hard drives they could afford. Even then the only effect the hard drive could have would be clicks, pops, gaps or synchronization problems between audio tracks all caused by buffer under runs. There isn't anything the hard drive can do to alter the data in a subtle way and as far as anything external to the drive is concerned data is recovered or an error is reported.

Anyhow this is moving more into the area of playback and away from the point that you can make a bit perfect rip of a CD on a computer using any one of a number of software packages, yet people claim they can hear a difference between files which contain the same information.

A good test for this would be:

1) Rip a track in a number of ways including HDX.
2) Make four copies of each file.
3) Mix the order of the files up and make sure they are labeled in a way so test subjects can't tell where the original rip came from.
4) Copy the files to the HDX.
5) Leave people to listen and attempt to group the recordings, no other person in the room while test underway.
6) Also have a control group where all the files are from the same rip.

I'm willing to bet that no statistically meaningful groupings come from this test.

quote:

Originally posted by matt303:
I'm willing to bet that no statistically meaningful groupings come from this test.

and so do I. But blind testing has always been shied away from by the hi-fi industry because they don't get the results they want unless someone is "leading" the test.

Matt, that's pretty much exactly what Malcolm Steward did. The only difference was that they put all the rips on a USB stick and plugged it into the HDX for replay, instead of placing all the rips on the HDX's drive.

Blind testing is usually just as guilty of its own problems. More often than not, you're asked to listen to a snippet of sound as opposed to engaging with a proper body of music). Given a minimum level of quality, we're not interested in better quality sound. We're interested in how music affects us, and for that you need to listen to substantial parts of tracks. Listening to snippets just tells you how something sounds, not how it makes you feel.

I've noticed in the article on their website that they've started talking about CDs sounding different, which if they were mastered with differing amounts of jitter is possible. But two identical files played on the same hardware can only sound the same there is no additional magic information in the file.

This is my last post on this because this is the equivalent of a bunch of people looking at a car and wondering where the hay goes.

quote:

Originally posted by matt303:
I've noticed in the article on their website that they've started talking about CDs sounding different, which if they were mastered with differing amounts of jitter is possible. But two identical files played on the same hardware can only sound the same there is no additional magic information in the file.

This is my last post on this because this is the equivalent of a bunch of people looking at a car and wondering where the hay goes.

That's why you need a truck, yee haaaaa!!!!!

This thread has really been an education for me in terms of the amount of BS and snake oil that is widely peddled in this hobby (by some of the most reputable sources in the industry, BTW). For that, I am very grateful indeed.

jazzfan,

Hasn't it just!!

quote:

Originally posted by matt303:
I've noticed in the article on their website that they've started talking about CDs sounding different, which if they were mastered with differing amounts of jitter is possible. But two identical files played on the same hardware can only sound the same there is no additional magic information in the file.

I agree, but the elephant in the room is that the files ARE NOT identical. Thus all bets are off. An Apple Lossless file is not an HDX file, or am I wrong about this? Even if the numerical data representing the music contained in the files is exactly the same, they at least have a wrapper that indicates what kind of file each format is. And since bits are bits, there must be a codec or some definition that tells the computer what to do with the data in the respective files (i.e. here is how to play Apple Lossless, here is how to play HDX, here is how to corrupt a Word document). It is in that definition which acts as a filter between the data and the eventual electrical signals produced by the computer, not the file itself, in which a whole host of variables can affect how the exact same data can produce a different sound.

I am not arguing that one sound is better than the other having never even seen an HDX, or even that there is a difference between EAC, HDX etc. I am only saying that it is plausible to me <obligatory d*ck schwing>as a software developer with a mechanical engineering background and over a decade of experience who first played Pong when he was five and once took a tour of an Italian castle with a beautiful woman named Monica</schwing!> that identical data can produce different results.

@Naijeru

The claim is that the HDX produces better rips period. Nothing about Apple Lossless, WAV, AIFF whatever. HDX produces better rips to the same format. WAV for example.

The statement many of us are making is that two bit for bit identical files including the header when replayed through exactly the same chain have to and will always sound identical.

Therefore, if an HDX rip to WAV sounds better than an iTunes rip to WAV the files have to be different.

(The header itself is transparent in that you can observe what is in it and determine whether any differences are material or not. In a WAV it has the frequency/word length, whether the words are big or little endian and thats about it)

Bottom Line.

If you cannot hear the diff between FLAC, WAVs, ALAC. Or HDX, EAC or LINN etc...

you are not in the market for an HDX....

just wait for the nDAC and rip your own way, and store on your own drives.

It should be pointed out that Naim on the HDX webpage do not claim that rips made on the HDX sound better* than rips from other devices only that it makes rips that are bit perfect.

*with the exception of the track start/end

@matt303

The claim has been made, by NAIM or NAIM guys, that the HDX rips sound better than iTunes and this claim is also made in HiFi Critic.

quote:

Originally posted by Harry H. Wombat:
The statement many of us are making is that two bit for bit identical files including the header when replayed through exactly the same chain have to and will always sound identical.

Agreed.

I'd imagine that you'd have problems with the ASA if you put in print that your bit perfect rips sounded better than another companies bit perfect rips.

If this is really the line taken by Naim I feel it's yet another step away from Naim products being based on quality engineering and towards HiFi marketing and myths. Which is a shame because the Naim products I've owned have been great and they don't need to resort to these types of claims.

@matt303

Ah - but this is not the claim. The claim is that the rips sound better but this is never backed up by a demonstration of the differences in the ripped files or any claims that other rippers are not bit perfect.

See Paul Stephenson's reply here:
Sounds better subjectively

Harry, you've repeatedly comment on the bit perfect/check sums and therefore, the argument seems to be that all of us who have heard differences are essentially "wishing" these to be the case.

I think DD makes some very pertinent points. However, I have listened to different files with different software ripping engines and check sums having been doen to show that they are "bit perfect" yet they sound different and it didn't take more than 20 secs of listening to hear that difference.

I'm not a technical guy, but I can only go by what I hear and my hearing isn't better than the average person's yet the differences were clear and heard by many people. I don't sell stereo equipment so I have no reason to find one better than another. I also listened extensively before I bought.

So, please explain this and I mean this sincerely and not as a challenging remark since subjectively is all I can really go by.