Vote

that certainly looks like an Ortofon 2M Bronze in the second pic, not the $99 Red...

looking foward to listening to these

So the loudest cartridge is the best one? (Obviousy kidding, but it is going to be hard to normalize the volumes).

So, we have 10 files. The tenth file is the Anna. Is file number 10 the Anna? It says above: "The tenth is Ortofon’s $8499 Anna...That file is the “control” against which the others should be compared, except it will not be identified as such." 

I guess I could listen and see if file 10 is that much better than the others, that way I'll answer my own question. Seems like we should be picking the second best file if we're ranking the budget cartridges?

D

otaku2 has a good point. If these files aren't adjusted to be the same percieved volume within 1 dB, it will be difficult to judge accurately. I haven't downloaded the files yet so this point may be moot if the gain is carefully adjusted.

That's it! I've managed to repetedly talk myself out of a DAC for at least 5 years now. But this is just too damn cool to ignore. If I were to snag a Cambridge DacMagic 100, or perhaps the Audioquest Dragonfly, would either do these files any justice, or would I just be wasting my time/money?

All these files are at different levels so need running through a wave editor to normalise them to 0dB. Two files already peaked at 0dB but the others were all under.

This is a great idea - and withholding the identity lets you listen without preconceptions. 

I'm pleased to see you're using a Graham Slee Era Gold V as I use this, along with an even better Graham Slee Reflex M, so the comparisons make it easier for me to relate to. You could do an 11th file - the Anna with the Era Gold partnered with a GSP Elevator EXP. That way the Era Gold's presentation would be common to all :o)

First of all, thank you for a fantastic idea and for the opportunity to hear the products myself, rather than just reading about them. I have yet to "take the test", but will and are looking forward to it! I really hope this practise will become a new norm in audio reviews (though obviously not possible for many types of products).

This has led to some thoughts I wanted to share, and hopefully get some reasonable answers to, or just an interesting discussion (there are many clever contributors to this excellent site).

If vinyl and analog sound has a special quality that digital formats do not, will this test be able to give us a reasonable impression of what each cartridge can do? Or are we missing something here (namely the essential analog quality)?

I am not trying to question whether analog sound is different from digital, or even that it is better (this would question both my sanity and the justification for my record collection, so let's quickly move on!). I simply raise the issue if there is a logical flaw in this experiment. To phrase this in another way: by setting up this test, is Mike actually saying that hi res files can satisfyingly reproduce the qualities of good (and even the best possible, re: MC Anna and Continuum) analog sound?

Perhaps the test is good enough to be able to hear just some of those most essential differences between the cartiridges and this is the motivation for the test.

This reminds me of a similar "logical problem". Mike describes bringing recordings from his analog setup to audioshows. On a CD. As far as I remember, the sound on those discs are much better than commercial CD's of the same music. I can't remember if Mike's are hi res files. Would this not indicate that a major component of the inferior sound on CD's is to be found where the CD's are mastered/produced, rather than in the format itself?

Again, I am not questioning the premise for this site or Mike's excellent work, which I very much enjoy and respect. I simply thought these questions were interesting and couldn't provide a good answer myself.

A comment from my own experience of doing this same exercise with my own cartridges...

The tracks need to be normalised on their Average RMS levels and not on their peak levels.

Peak levels above the Av RMS will provide a quantifiable measure of cartridge dynamic ability.

Normalising on peak values will of course accentuate dynamic differences perceptually.

looking foward to going through these files

Thank you for the chance to "listen" to these cartridges back to back

David

I pulled the tracks into Adobe Audition and did some quick statistics on them...

All figures in dB - Peak is RMS average

Track 1 peak -2.08 Av -21.98

Track 2 peak 0db Av -19.15

Track 3 peak -2.08 Av -21.98

Track 4 peak -6.15 Av - 26.35

Track 5 peak -7.73 Av - 27.79

Track 6 peak -5 Av -25.19

Track 7 peak -2.88 Av -21.93

Track 8 peak -3.32 Av -23.02

Track 9 peak 0 Av -17.76

Track 10 peak -2.53 Av -22.61

Comments based on the stats:

1) Track 9 and track 2 are clipping - this will introduce some hash - how much depends on how often it clips, and how well the ADC copes with it. (also how well the individual listeners DAC copes with it too!)

2) The tracks are NOT normalised - highest level is track 9 at -17.76, lowest is Track 5 at -27.79, a 10db difference is huge

In terms of dynamics they range from a low of 19.7db (difference from average to peak) to a high of 20.2db - 0.5db difference, not huge but enough to be noticeable (Track 2 seems clipped by around 0.8db, Track 9 seems clipped by around 2.5db)

The tracks should be normalised to the same Average/RMS level - probably around -24db - which would provide for a touch of headroom to keep people's DAC's happy.

I may be accused of being obsessive compulsive for making these comments (and measurements) - but when I did this experiment myself, the tracks were dramatically more similar once normalised.... which demonstrates the huge influence of amplitude on psycho acoustics.

Without normalisation the results cannot be valid....

(I would be happy to normalise and republish the tracks if that is helpful)

bye for now

David

This is really great. I do agree with those who state that the tracks need to be volume matched for best results. I note that the leading vote getter is currently also the sample with the loudest average RMS. 

I volume matched in my own audio editor, reducing the volume of all tracks except the one with the lowest average RMS so that they are approximately volume matched. This makes it easier to hear differences in tonality, resolution, etc. because volume difference is always going to be most obvious thing perceptually.

The way to get the levels right is to capture a 1KHz reference tone from a test disc and adjust the analogue gain into the ADC so as to at least get that equal for each test. This is more valid than normalising either the peak or average levels. This has to be done by the recordist, obviously. There's not much the end user can do about the levels really objectively without this.

In any case it's great that MF has done this and presented it blind.

I noticed the different types of cartridge screws used. Due to weight they change they change tonearm mass do they not? Does this alone hinder an apples to apples comparison? Why or why not?

1.   Third picture down: Audio-Technica AT7V.   Did you test it

with only three wires attatched?  Thanks.