A/D Converter Shootout (with samples)

[Spartacus]

Originally Posted by gtr-cl

When I listened to these tracks with a AKG K530 headphone on a DELL Precision M6400 notebook I could not hear any difference at first but after repeated listening I convinced myself that #1 was differernt, felt less connected to and I was just planning to order the HEDD today! Any way in a blind test before I read the posts, I chose 2 and 5 and like 2 the most.

If you read all the topic, you will see that there might be an issue with the file number 1, it looks like compressed or something...
other Hedd measurements showed better result :
A/D Converter Shootout (with samples)

[gtr-cl]

Thanks, I indeed missed at least those few but apparently important posts to consider!

[blusound]

Hello to everyone
very nice posts!! I read all 13 pages but there's some audio files that are missing for send space time expiration. Can someone upload again 2192 loopback, hedd loopback and lavry blue loopback?
Very thanks in advance!!

[finleysound]

3 - my favorite, liked what it did to the kick and bass
4 - least favorite, sounded a bit muddy and congested

[edjay]

I chose 3 because I initially thought it sounded like it had slightly more higher detail: When I listened a few times I couldn't decide at all?

[Muser]

problem is.. I'm listening through my conversion.. so I don't really know what I'm listening to.

but, as it goes, they all sound pretty good to me.
but I wouldn't choose, for the above reason.

thanks for the files though.. some useful data should be in those files.

[daves]

Thanks. This was helpful. I was surprised how much difference there was between units. It surprises me sometimes that I can even hear differences on crappy PC speakers. I ranked the top three as Universal, Prism and Mytek, but it was more because of what the result sounded like than which was closest to the original, which kinda made it a crap shoot between them. It seemed like they all changed the sound in some way, it was just which you liked best.

[Spartacus]

Originally Posted by daves

I was surprised how much difference there was between units. It surprises me sometimes that I can even hear differences on crappy PC speakers.

congrats !

[...Jazz!,]

Originally Posted by Muser

problem is.. I'm listening through my ears.. so I don't really know what I'm listening to.

but, as it goes, they all sound pretty good to me.
but I wouldn't choose, for the above reason.

thanks for the files though.. some useful data should be in those files.

[WBM]

Lavry Gold best sound far.
Heed 192 good eq and warm A/D D/A sound difference to the lavry, I use it for different kind of music.
I trusted Forssell with my eyes closed and his plate on NSEQ-2 Amazing!!, I bet he build a hell of converter.

[karumba]

thx for the tests! since i'm also currently in judging converters for my studio (i'm also testing the mytek 8x192), i've run your loopback test through my evaluation scripts. beside the important blind listening tests, i think its also cool to have the "mathematical view". pls find the results below:

just a few notes what my script is doing:
1. subtracted each converter file from original in time domain
2. calculate the FFT of each file (32k window) & substract the converter results from the original.
3. plotting the curves of absolute differences in time & frequency domain

the figures to the left are showing the errors in time domain & to the right in the frequency domain.

from my experience with this type of evaluation (which is of course some kind of a special view) i have to say, that all converter are performing on a very high level. if we look on time & frequency domain at once, the lavry gold seems to perform best followed by the uaudio. the all the other converters are roughly on the same level, maybe except the forsell whose performance seems a little bit worse compared to the others.

[imaginaryday]

they all sound good to me. so, my vote goes to all of them.

cheers.

[didier.brest]

Measurements from Audio DiffMaker:

Time offset, level difference, correlated depth (the higher is the better, relates to the linear distortion), RMS level in dBFS of the difference track generated by Audio DiffMaker (time aligned and gain compensated with respect to the original) and measured in Wavelab.

2 first silent seconds removed because otherwise it corrupts the gain correction.


Mytek + Crane-Song
-6,901usec, -0,067dB (L), -0,021dB (R)..Corr Depth: 39,3 dB (L), 40,1 dB (R) Diff: -52,5 (L), -52,2 (R)


Mytek + Lavry
-68,95nsec, 0,079dB (L), -0,003dB (R)..Corr Depth: 35,5 dB (L), 37,7 dB (R) Diff: -54,8 (L), -56,0 (R)

Mytek + Mytek
parameters: 2,52usec, 0,023dB (L), -0,007dB (R)..Corr Depth: 37,3 dB (L), 38,0 dB (R) Diff: -56,4 (L), -56,5 (R)

Mytek + Forsell
parameters: 834,2msec, -0,002dB (L), -0,001dB (R)..Corr Depth: 32,6 dB (L), 33,2 dB (R) Diff: -45,8 (L), -45,6 (R)

Mytek + Prism
-5,032usec, -0,028dB (L), 0,022dB (R)..Corr Depth: 36,2 dB (L), 37,1 dB (R) Diff: -52.63 (L), -52.52 (R)

Mytek + Universal Audio
-239,3nsec, 0,000dB (L), 0,068dB (R)..Corr Depth: 38,7 dB (L), 39,4 dB (R) Diff: -57,8 (L), -57,7 (R)

loops from Ivo

DAD + DAD
-0,515dB (L), -0,545dB (R)..Corr Depth: 29,5 dB (L), 30,2 dB (R) Diff: -50,4 dB (L), -50,7 dB (R)

Forsell + Forsell
-1,096dB (L), -1,073dB (R)..Corr Depth: 31,9 dB (L), 32,5 dB (R) Diff: -46,0 dB (L), -45,9 dB (R)



One shall have

RMS level of the original (~-20 dBFS) - correlated depth = Linear distortion level < difference level

because the difference level consists in the linear distortion and an uncorrelated component, which is noise and non linear distortion.

Ranking based on linear distortion
Crane-Song > Universal Audio > Mytek > Lavry > Prism > Forsell > DAD

Ranking based on linear distortion + noise
Universal Audio > Mytek > Lavry > Prism > Crane-Song > DAD > Forsell

[Seditionary]

Apogee AD-16X -- converter #7 -- ad-16x converter.aif

Ran the original file in Logic 8 through an AD-16x advanced option routing... then to a Rosetta 200, into an Aurora GTQ2 at line level 0. (don't know what DA or pre was used in the shoot-out test)... and then back in through the Apogee AD-16X.

This is the most unbiased Apogee Ad-16X test I could think of to do. I think it stacks up (unless you compare it to the Lavry Gold (i also really like 5, the orpheus for this track))... man, the lavry gold sounds good though.

Would love to hear the same through a Symphony. Apogee folks? Hook me up.

[Adebar]

Originally Posted by Muser

problem is.. I'm listening through my conversion.. so I don't really know what I'm listening to.

but, as it goes, they all sound pretty good to me.
but I wouldn't choose, for the above reason.

I wouldn´t choose either for the same reason. It gives you some idea in which direction each converter sounds related to each other. But I wouldn´t decide based on this comparison.

[sat159p1]

Interesting that UA is almost identical to Lavry Gold from these charts

[Seditionary]

Originally Posted by sat159p1

Interesting that UA is almost identical to Lavry Gold from these charts

it doesn't sound at all the same to my ears though. it is strange, the prism orpheus almost sounds more like the lavry gold to me.

[Jason Rabbit]

Originally Posted by Seditionary

Apogee AD-16X -- converter #7 -- ad-16x converter.aif

Ran the original file in Logic 8 through an AD-16x advanced option routing... then to a Rosetta 200, into an Aurora GTQ2 at line level 0. (don't know what DA or pre was used in the shoot-out test)... and then back in through the Apogee AD-16X.

This is the most unbiased Apogee Ad-16X test I could think of to do. I think it stacks up (unless you compare it to the Lavry Gold (i also really like 5, the orpheus for this track))... man, the lavry gold sounds good though.

Would love to hear the same through a Symphony. Apogee folks? Hook me up.

First, The Stereo image is too different than other converters.
And it sounds like it's comped.
Sounds so Punchy to me.

[DAZZER]

Originally Posted by bjorn

hmm I don't know
My interface is metric halo ULN-2 2d
my 2192 is hooked up via AES (2192 is master clock)
I recorded using metric halo's MIO console record panel
it always records as mono split
so I combined the mono split recording file by using apple's terminal command
but when I made stereo interleaved it become aiff
so basically mono split (wav) -> stereo interleaved (aiff)
maybe the way I combined the mono split did something?

I uploaded the original mono split file
could you check it one more time?
if it is still less precise, what would be the problem?
either 2192 or uln-2 is wrong.. or my recording chain..?


Download 2192 mono.zip from Sendspace.com - send big files the easy way

Error(

[DAZZER]

Originally Posted by mischa janisch

I finally had some time to do a Loop using Lavry Blue DA-AD. The converters are level matched within 0,1db @ 1kHz. I personally prefere it to the Mytek Loop (#3) that most seemed to favor in Robins examples. To me it shows that it is very important to judge the entire Loop, not just the AD...


Download Lavry Blue Loop.zip from Sendspace.com - send big files the easy way

Error

[FLYINGJAY]

Wow I can run the original threw my converters! And....line it up!

Sent from my PC36100 using Gearslutz.com App

[arf]

Originally Posted by karumba

thx for the tests! since i'm also currently in judging converters for my studio (i'm also testing the mytek 8x192), i've run your loopback test through my evaluation scripts. beside the important blind listening tests, i think its also cool to have the "mathematical view". pls find the results below:

just a few notes what my script is doing:
1. subtracted each converter file from original in time domain
2. calculate the FFT of each file (32k window) & substract the converter results from the original.
3. plotting the curves of absolute differences in time & frequency domain

When subtracting files, don't they have to have a common clock and time base for the null to be valid? Each DAC/ADC loop has it's own clock and time base that will differ from that of the original file. ADC clocks are also going to shift slightly causing a moving target over the duration of the resulting capture. Even with a master clock running the DAW and the ADC, won't the DAC's latency throw off the sample alignment of the resulting capture?

[DAZZER]

Originally Posted by ISedlacek

Unfortunately I realised that I don t have Lavry Black DA here at the moment ... Will do Lavry DA DA it later

We are still waiting...

[CMO Mastering]

Great test!!! Makes me even happier that we own good quality equipment!

[CMO Mastering]

Forgot to mention my favourites being 3 & 5

[soundgeezer]

Hello everybody, this might interest you:

I also own 2 mytek 8x192 converters and paid a lot of money to get high quality digital and analog cable, same goes for power cable etc.

I ran the original file through my mytek and would like you to compare to Robin's sample of the mytek (number 3).

Do you here a difference? The analysis guy can also maybe jump in to check if there is a measurable difference.

Here is the file: Mytek SW.wav

[didier.brest]

Updated list of Audio DiffMaker results.

Mytek + Crane-Song
-6,901usec, -0,067dB (L), -0,021dB (R)..Corr Depth: 39,3 dB (L), 40,1 dB (R) Diff: -52,5 (L), -52,2 (R)


Mytek + Lavry
-68,95nsec, 0,079dB (L), -0,003dB (R)..Corr Depth: 35,5 dB (L), 37,7 dB (R) Diff: -54,8 (L), -56,0 (R)

Mytek + Mytek
parameters: 2,52usec, 0,023dB (L), -0,007dB (R)..Corr Depth: 37,3 dB (L), 38,0 dB (R) Diff: -56,4 (L), -56,5 (R)

Mytek + Forsell
parameters: 834,2msec, -0,002dB (L), -0,001dB (R)..Corr Depth: 32,6 dB (L), 33,2 dB (R) Diff: -45,8 (L), -45,6 (R)

Mytek + Prism
-5,032usec, -0,028dB (L), 0,022dB (R)..Corr Depth: 36,2 dB (L), 37,1 dB (R) Diff: -52.63 (L), -52.52 (R)

Mytek + Universal Audio
-239,3nsec, 0,000dB (L), 0,068dB (R)..Corr Depth: 38,7 dB (L), 39,4 dB (R) Diff: -57,8 (L), -57,7 (R)

loops from Ivo

DAD + DAD
-0,515dB (L), -0,545dB (R)..Corr Depth: 29,5 dB (L), 30,2 dB (R) Diff: -50,4 dB (L), -50,7 dB (R)

Forsell + Forsell
-1,096dB (L), -1,073dB (R)..Corr Depth: 31,9 dB (L), 32,5 dB (R) Diff: -46,0 dB (L), -45,9 dB (R)

Loop from soundgeezer Mytek 8 x 192
-0,014dB (L), -0,065dB (R)..Corr Depth: 37,2 dB (L), 37,8 dB (R) Diff: -56,8 dB (L), -56,9 dB (R)

[mmarra]

This is a great post ...as I'm in the market for a A/D converter because I'm moving over to the analog realm and this shoot really helped as I needed a bit of help choosing a A/D D/A.

[Spartacus]

Originally Posted by soundgeezer

I also own 2 mytek 8x192 converters and paid a lot of money to get high quality digital and analog cable, same goes for power cable etc.

Originally Posted by didier.brest

Updated list of Audio DiffMaker results.

Mytek + Mytek
parameters: 2,52usec, 0,023dB (L), -0,007dB (R)..Corr Depth: 37,3 dB (L), 38,0 dB (R) Diff: -56,4 (L), -56,5 (R)


Loop from soundgeezer Mytek 8 x 192
-0,014dB (L), -0,065dB (R)..Corr Depth: 37,2 dB (L), 37,8 dB (R) Diff: -56,8 dB (L), -56,9 dB (R)

This should prove that spending lot of money for cables is useless and confirm what some serious engineers think about cable manufacturer...