I tried to download the files yesterday without success. Will try again later today.


/Peter

How can we tell what difference there is in the converters given we dont know what the oriiginal sounds like without it going through the Fantom's D/A?

I understand you are trying to test the A/D side of the converters but it would make more sense to record to Ptools via the Fantom's digital outs,then go out of each converter and back into it. It woudnt isolate what the A/D sounds like (you haven't really done that anyway). But it would tell us what each converter's sound is compared to the Fantom tracks that did not go out of a converter.

Same as if you imported a CD file and made the loop.

For anyone who buys either converter, he would be mixing via its D/A anyway so it would tell a bit about the "sound" of each brand.

I checked and it works
http://www.thecoldchamber.com/test/test.zip

This is not a test for "mixing" but only for tracking.
When you record a song from a set of keyboards to your PT rig it will be converted at the source anyways so it doesn't change anything. the original is not converted, digital to digital then you can compare it with fantom digi to analog the analog back to digi via converters yes it's converted twice but it would be as well when anybody doing music would use a sound module plugged into converters. it's not a test for converters in general but for this specific use.

like someone else said...
wouldn't a Y cable right before the 2 converters be the best way?

same source, same mic, same pre, etc....

I'd say 1 is the Aurora, 2 is the Rosetta, 3 is the M-Audio interface.

But you won't really hear what a good converter is capable of using the output of a synths DAC of course...

I'm planning on purchasing an Aurora 8 or Rosetta soon and would like to know what differences you hear between the converters in the samples (rather than just guessing which brand the samples represent).

I can't download the samples on my Mac or PC, for some reason beyond me, so very curious to know what you're hearing between the samples.

There's an enormous bass bump on the second file. The third one has less detail and worse transients than the first.

But you still have nothing to compare it to.heh
So again
My point is that the Fantom Digital track isn't a ref of anything.

When you compare to it what the converters are doing, how do you know which differences are because of the Fantom A/D?

We are listening to the Fantom through a few sets of converters to see which sounds best to us but with no ref. (I see what you're doing, I just don't get the point of the digital ref)

results here it goes:

numbers are good for the downloadble Wav and for the order in the protools snapshot:

Original
1= M-AUDIO
2= Rosetta 200
3= Aurora 16

Here is my point of view.

The M audio really changes the frequencies, if you look at the graph they don't really match with anything else.
The Rosetta Match pretty good BUT literally CUT OFF a lot of the "lower" info...that's a bummer... even the M audio doesn't do that.
The aurora is the closest to the spdif to PT (original) file
About the sound now
The m audio sound harsh and digital
The Rosetta sounds good but Has a lot of bass, it actually puts the vocals behind in the mix and i would if mixing with the rosetta put the vocals louder..which is in fact not needed. The Aurora is the cleanest most colorless sounding one out of the 3.
When trying to enjoy music in a listening kind of way, the rosetta is more pleasing to the ears BUT when Mixing I find the Aurora Way more transparent with also more details (as shown on the graph).
The M audio is Not that far actually (considering the price) but Will change the frequencies and make everything sound Harsh.
MY opinion.

The digital ref was just a"ref" with No converters....you don't see a point to that??

No conversion ...against conversion from fantom and bacl to Digital with m audio and same with rosetta and aurora

The Original is the "real one" since it is not altered by anything else.

Now sure the fantom DA does some conversion in this test but it's the same for each take! so no matter how much it does at the end the one looking or sounding closest to the Conversion-less file is a winer in my book. (of course it will never be exactly the same as the original)
After going thru the fantom conversion and the converters the final results will still be different from one another but one is still closer and THAT is the Aurora hands down
PS: remember I own BOTH so I'm not one of these guys trying to justify their choice and purchases . I tell it How I ear it and see it.
Even tough this test might not be the best one this is all I can put on a web site. Hearing the real thing isalso dofferent and the Aurora is cleaner. maybe the Aurora compares more to a AD16 DA 16 from apogee then the Rosetta series

What mdjice says make sense.

Considering the Aurora being transparent in the few loopback and "pass thru" tets I've done I'm not surprised at these results in this test.

I will do some serious blind tests with a panel soon on some converters in different price segments. The listening will not be A/B but before and after the AD/DA trip thru the converters.

The Aurora measures extremly well and that seems to translate into very good audible performance.

mdjice,

do you have the RMAA software at hand?

I would be very interested in seing measured performance of the contenders... and especially the Rosetta.

I downloaded the files but couldn't listen to them for some reason. Seems like some plugin was missing in my WMP and did't find it on the net..


/Peter

Wow, I prefered M-Audio to the Aurora... Listening through an Aurora16 myself...

But looking at the files again, I noticed the M-Audio file is 0.4 db louder. You've got to match levels accurately in a listening test, louder is always better...

I've done some loop tests with my Aurora. There is a slight loss of high end with every pass, but the audio stays intact otherwise and won't get harsh.

When you say prefered, does that mean "sounding the best" or "closest to reference file"?


/Peter

I only compared the 3 recorded files, the reference is useless since it hasn't passed the synth's DAC.

BTW, with this test it's all too obvious that Apogee's repeated claim on this board about their converters sounding closest to the source and not adding colour is bull****.

I wouldn't say the ref is useless.. after all it's the "ref". The other can be as good (clean) or altered to one degree or another. The file sounding closest to the ref would get my thumbs up.

Listening to a couple of files/components (like in an A/B test) and choosing the one that sounds subjectively best is not wise IMO. Unfortunately that is often how things are done and it's often all about compensating an error with another error. When the DUT allows for an absolute testmethod for investigation of it's effect on the audiosignal that is the one to go with IMHO.


/Peter

Peter I don't have the soft but Will look into it.
About the M audio it was the loudest, I did my best to match the gain it might be a tiny bit different but again I did my best you have the files you can adjust as well.
Also about not being able to open the files, remember these are 24 bit so Not every Consumer soft will open, you have to use cubase, reaper, PT logic wave lab etc...
And yes the Myth about the apogee is (to me at least) busted. Are they great converters? Sure, are they transparent? NO! do they color sound and miss lower frequencies when converting...YES
that said the M-audio again don;t sound all that bad...just a bit harsh but is it worth investing thousands...Good Question ..However remember that this is only 1 track! if you record 48 tracks thru each converter then the Little harsh sound from the M audio might turn into a BIG HARSH sound when all stacked..which is why using the aurora when multi tracking is still worth it.

With stuff recorded from the source (mic, tape machine, analog synth etc.) rather than a dgital synths DAC (they don't use high end converters in these devices) the difference is huge. So is the difference when processing audio internally (because the converter problems artifacts become more audible) and even more so when using outboard gear. I've tried several budget converters, older high end converters as well as modded converters and the Aurora is noticeably better than all of them.

Ah, about 24bit.. should have understand that myself.
Will try in another software.

/Peter

Well, I wouldnt go so far as to say that.

I just looped his digital file through the outs on my rosetta 800 and back in.
I know it doesn't have the Fantom D/A in the picture, but it's "supposed" to be the same converter as the Rosetta 200. I don't hear any of the characteristics that are shown in this test. The Bassyness and cloudyness that I hear in his Rosetta file I'd expect to hear to some degree comparing my rosetta file to the digital file. I don't hear it.

I'm about to record my Fantom playing the demo and level match to see what I get.

Again I know the flaws of my quick and dirty loop back through the converters, but I expected to hear some of the characteristics of his rosetta file and I don't.

Now if I don't get those characteristics when I record the fantom demo from the fantom's outs should I conclude that the Rosetta 200 and 800 are different?

The tests I've done so far with the Aurora8 is the following;

Looping a couple of 16/44.1 tracks thru the converter from the hard disk.
These I listened to with an Audiofire4 and HD600 headphones. Don't think I could hear a difference. The material was some Eva Cassidy, Pink Floyd (Money) and a very good recording of Vivaldi four seasons.

What I was listening to and comparing was the ripped PCM directly from the HD via the AF4/HD600 combo and the same but after having passed the DA/AD of Aurora8.

I also did a test popping in the Aurora8 between my SACD player and amp listening live to what it did to the signal (hi resolution speakers and decent room).

I used material that was either direct to DSD or analog tape to DSD. I did the tests myself and had no way of calibrating levels so I set them as good that I could. I don't think I could here the Aurora8 in the chain. Possibly a small effect with 44.1kHz that was shrinking the soundstage a tiny tiny bit bit that may very well be due to different levels or the cheese cake I had for breakfast. I imagined that this small effect was absent at 192kHz.

All in all very promising results but a more rigorous tests is planned.



/Peter

No, I really don't see a point to that. To listen to the "ref" converters will be involved at some point. We can't listen to a digital file without converting it to analog.

You are correct that the Fantom D/A does some conversion in this test, but for the test really to be valid, we would have to be able to hear the original signal through the Fantom's converters since that's what the other three converters are "hearing". Without that we really can't know which converter is most accurate.

Of course the test is valid. We have the ref file which for this test only is affected by the DA in your monitor setup.

Then we have a, b and c that has the Fantom and "a" or "b" or "c" in the chain.

Let's assume for a while that the Fantom and your monitoring DA is transparent then the only thing that can differ between the files is the AD in the contenders.

At the least, an AD obviously always needs to be tested with a DA.


/Peter

That's the problem...you can't assume that the Fantom and your monitoring DA are transparent, because they are not. I suppose that if the test were done through a particular converter, and the people who were downloading the files had that same converter, the comparison might have some relevance...but if we're not able to hear what the converters heard (the analog signal coming out of the Fantom) then it's a flawed comparison.

I don't assume that they are transparent.. but one error one place in the chain does not mask everything that happens up or down stream. Shure, there is a little chance that an error in the Fantom cancells an error in one of the AD's but that is not very likely.

In this test also it is very likely that the one that sounds closest to the ref has the most transparent AD in the chain.

But sure, as I have mentioned earlier, a loopback or a direct live before/after test may be better. But in the end how do you know that the AD and DA in the particular converter doesn't cancell some errors that may show up in other combnations?


/Peter

Please do, I recorded the fantom demo using trs to xlr Mogami cables stright in the rosetta volume al the way up in PT, we'll see what you get.
I also know now that you have a Rosetta so admitting the results from my test will be a difficult thing from any apogee owner

It doesn't matter if thy are transparent or not since all signal (including the "original" will be affected the same way by your DA monitoring!! you are listenning to the samples from the same setup for each one right??
and as far as the original it has no conversion true so when the conversion has gone thru the aurora m audio and rosetta they all had the same source to begin with !! (The fantom converted to analog signal) at this point what is changed by the fantom conversion can only be accentuated by the other converter. The more the difference from the original, the more the converter (apogee lynx m audio) did to the "already" modified sound (by the fantom DA)

No, but what people argue about with these tests is you have to hear the source! We can't hear the source, so we don't know how accurate it is. If you hear the original track, then hear what it suffers from after going through each units A/D/A loop, then you'll know what it sounds like compared to the original.

Thanks for posting man! I think I preferred the Aurora. Unfortunately I saw the results before listening, thankfully I'm not partial for anyone of them.