Big A/D converter test session

[funka]

Here is the link with updated files(some converters added:SONOSAX SXR4 INT, TC Finalizer INT, AETA Mixy (mic in with 0dB gain), AETA Mixy(line in), LUCID 88192, FOSTEX FR2, BEHRINGER DEQ 2496, METRIC HALO ULN8):
Tests de convertisseurs - fichiers audio

NOTE: DO NOT READ PAGE 2 AND FOLLOWING PAGES, OR DO NOT OPEN RESULTS LINK IF YOU WANT TO DO THE REAL BLIND TEST BEFORE KNOWING THE RESULTS .

The pretenders were:

- SSL ALPHA LINK

- NAGRA VI

- LAVRY AD10(with different settings)

- DIGIDESIGN 192

- FORSSELL MADC2

- PRISM ORPHEUS

- LYNX AURORA16

- METRIC HALO ULN2

- LAVRY BLUE(with different settings)

- RME FF 800

- SONOSAX SX-ST

- STUDER D19

- UNIVERSAL AUDIO2192



The test was done like that:

We recorded a short piece of instrumental/vocal music(very strange and experimental), just to have some diversity in the recorded test signal, to try testing stereo imaging, percussive sounds, treble, bass...

The mic was a Schoeps sphere KFM6(stereo mic), into studer preamps(from a 962 console), into a Studer A810, well calibrated.

Then the analog signal recorded on tape served as the source for all converters, levels were calibrated each time to match 0VU = -18dBfs.
Digital signals were directly recorded into PT HD3.



We tested out the converters at 44.1(internal clock only) & 96 kHz(internal clock or external with a Tektronix clock).
Note that the STUDER D19 is only 44,1/48kHz, and that the Forssell can't be externaly clocked.



Now guess which is which?

Results can be seen here: Results

The tape was kept preciously, to be able to test further converters in the future...while available.

[space2012]

hi,

do you have a digital camera?
can you tell the reference of the tektronix ?

53% downloading.
Orbit Downloader: the ultra file & social media (YouTube etc..) download manager

all recorded at 96khz? exept 1?
what software you used to make the 96 to 44.1Khz files ?

Thanks.

[space2012]

Originally Posted by funka

- SSL ALPHA LINK
- NAGRA VI
- LAVRY AD10(with different settings)
- DIGIDESIGN 192
- FORSSELL MADC2
- PRISM ORPHEUS
- LYNX AURORA16
- METRIC HALO ULN2
- LAVRY BLUE(with different settings)
- RME FF 800
- SONOSAX SX-ST
- STUDER D19
- UNIVERSAL AUDIO2192

theres a converter in that bunch that sounds amazing.
im shure is not Metric Halo, Prism, Lavry, RME, Digi 192io, SSL.

[space2012]

Hi,

did you disabled the SynchoLock(TM) in the Aurora?
did you disabled DDS in the RME?
what brand & model of cables did you used?

Thanks

[funka]

do you have a digital camera?

Yes, but it was done at a Friend's Studio in Paris...
Anyway, here are some photos:

can you tell the reference of the tektronix ?

I'll ask, it wasn't mine.

all recorded at 96khz? exept 1?

To be clear, we made 3 different converter tests:
- 44.1kHz self clocked
- 96kHz self clocked (no STUDER D19 here)
- 96kHz slaved by the tektronix clock (no STUDER D19 here, nor Forssell MADC-2 as it can't be externaly clocked)

what software you used to make the 96 to 44.1Khz files ?

None! We did not use any sample rate conversion. When I talk about 44.1kHz or 96kHz, it is the original conversion clock(be it internal or external). I'll add that there is no effect at all, only microphone -> preamp -> converter -> PT.

did you disabled the SynchoLock(TM) in the Aurora?

No, it was activated.

did you disabled DDS in the RME?

I do not think so...I'll check that.

what brand & model of cables did you used?

Schoeps Sphere-> Studer console linked with 8 meters Neumann/1 meter Gotham/9 meters Canare Starquad L4E3 loom.

Studer Console -> A810 tape linked with 16 meters Canare Starquad L4E3 loom

A810 tape -> Hendicott console(used for level calibration when converters could not be calibrated directly) linked with 16 meters Canare Starquad L4E3 loom

Hendicott console -> Tested converter linked with 20 meters Canare Starquard L4E5

Synchro:
Tektronix Clock -> PT HD3 "Sync" interface, video clock sync type with 6 meters 75 Ohms Canare LV61S + 1 meter video patch ADC

PT HD3 "Sync" interface -> Tested converter: "Word Clock" Sync type with 20 meters 75 Ohms Canare LV61S cable.

Recorded Digital signal(depending on selected/available output):
AES out from converter -> PT HD3 "IO192" interface linked with 20 meters AES Canare 110 Ohms DA202 loom + 1 meter into the interface.

S/PDIF out from converter -> PT HD3 "IO192" interface linked with 15 meters 75 Ohms Canare LV61S cable.

So the total of analog wires is about 70 meters long, 21 meters for digital & 27 meters for synchro signals(not all used at the same time).

[funka]

Concerning the Tektronix clock, it is the sum of three racks in fact:

1) ECO 170A that ensure the stability/control of the clocks.
2) TSG371 is the first clock source
3) TSG271 is the second clock source

[mobius.media]

Wow! Awesome work. Looking forward to results.

[Ethan Winer]

Originally Posted by funka

the analog signal recorded on tape served as the source for all converters

Excellent! Thanks for doing this. I can imagine it was a huge amount of work.

--Ethan

[space2012]

Originally Posted by funka

No, it was activated.

Sad.
if the Tektronix clock is better than Lynx SynchoLock(TM)
"probably"
the SynchoLock(TM) limits the Tektroniks to "SynchoLock(TM) levels"
...
same happens with RME DDS SteadyClock(TM), if DDS its turned on, limits the external clock to "RME Levels".
Sad.
great work anyway.

[space2012]

Originally Posted by funka

there is no effect at all, only microphone -> preamp -> converter -> PT.

Schoeps Sphere-> Studer console linked with
8 meters Neumann
1 meter Gotham
9 meters Canare Starquad L4E3 loom.

Studer Console -> A810 tape linked with
16 meters Canare Starquad L4E3 loom

A810 tape -> Hendicott console(used for level calibration when converters could not be calibrated directly) linked with
16 meters Canare Starquad L4E3 loom

Hendicott console -> Tested converter linked with
20 meters Canare Starquard L4E5

Synchro:
Tektronix Clock -> PT HD3 "Sync" interface, video clock sync type with
6 meters 75 Ohms Canare LV61S +
1 meter video patch ADC

PT HD3 "Sync" interface -> Tested converter: "Word Clock" Sync type with
20 meters 75 Ohms Canare LV61S cable.

Recorded Digital signal(depending on selected/available output):
AES out from converter -> PT HD3 "IO192" interface linked with
20 meters AES Canare 110 Ohms DA202 loom +
1 meter into the interface.

S/PDIF out from converter -> PT HD3 "IO192" interface linked with
15 meters 75 Ohms Canare LV61S cable.

So the total of analog wires is about 70 meters long, 21 meters for digital & 27 meters for synchro signals(not all used at the same time).

interesting, that could explain why they sound similar, exept 2.

was Tape-->AD converter, no Eq?

[tsvisser]

3C.wav in the 44.1 group has its left / right channels swapped.

I'm hoping for a good showing from the Nagra VI, as that is the only converter on this list that I own.

[Shy]

Listened to the 96kHz recordings. In short: they all sound very good to me and it's extremely difficult to spot any subtle differences.
P.S. very nice recording.

[djmukilteo]

I liked 7, 6 and 2 in that order on the 44.1....pretty subtle differences between all though....I used headphones to monitor....the rain stick and chimes in the first section seemed a little smoother sounding on 7 than any of the others, 6 was close to that and 2 was also smooth....
There was a second drum strike in that first section that was pretty beefy and saturated to me almost distorted or clipped and had a nasty ring at the top of the sound....every track had it....but maybe that was the microphone or tape....
voice was fine on all hardly much difference to my ear, the rain stick in the first section was a good test to focus on and crispy on almost all and the chime on some of the other tracks rang a little too much for me....
I'd like to know which wav's were which now!

[sonicdefault]

Thanks for this- Very interesting


-SD

[djmukilteo]

Ok...so...listened to all of the 96k files...
only one I really liked was 10b on the Tek clock
Everything else sounded pretty close to the same...very hard for me to tell....could be my ears...again used headphones....so closed monitoring with no room coloration.

All tracks clipped in both sections.....

One in the first section, same drum hit noted before and several clips occurred in the second section, again mostly the drum hits but the buildup in the last portion of the second section was where most of the clipping surfaced.....actually I don't think there were any clips in the 44K group....so I'm not sure if that was a level problem or what caused that....I'll have to check that again!

It was a bit annoying either way.....it certainly didn't provide any sense of dynamic range or noise floor reference.....which BTW...noise floor was really good on everything in all sample rates, even at 96K it was pretty quiet...not quite as as sensitive in the 44K group of course....I didn't really hear a difference in the clocking's....but then with the source material I'm not sure you could tell.....

It would have been nice to have actually had a musical piece for the sample...something with actual dynamic range with both quiet and heavy parts....not that I'm criticizing your sounds, it's just easier to have something familiar with normal tones to compare....that would have been the real test in listening to the same musical passages that you can compare space and spread and overtones and harmonics...
piano piece, synth, sax or horns.....rain stick could stay....oh and a full sweep with the windchime always works....

[funka]

was Tape-->AD converter, no Eq?

Yep, as I said, it was
- mic => preamp => tape, then
- tape => console for calibration => converters => PT

No effect, no eq, nothing else.

Thank you for your comments, indeed, it was a huge amount of work, it took us(as we were 7 in the studio for the tests) a full day to record all those converters! Tape was ajusted before by the owner of the studio, as "he knows his shit" very well.

3C.wav in the 44.1 group has its left / right channels swapped.

Oops, maybe, I'll check that.

Waiting before revealing the "who's who"...to allow other people to listen and give their favorites.

[The Listener]

Only 44,1 files - 9c, 5c, 7c, 4c, 11c, 8c, 14c sound good to me.

I liked 9c & 5c the most.

[studio freak]

many thanks for doing this!!!

so, what we are hearing is the AD convertion of 2 tracks stereo, right?

I was wondering if the results would be much different if the test was made with, lets say 12 or 16 tracks going from tape to AD.
would the diferences remain so subtle as the track count sums up?

[soundgeezer]

Merci pour le test !
I've been testing converters for the last 6 months and I know that there are some differences, even though they are pretty small when having a certain level of conversion (let's say starting at the aurora level).
I am still not decided which one to buy, and blind tests are the most valuable part, cause you hear good things, knowing it's an expensive unit.
So this is a great test. In my experience you will only be able to hear differences in a well treated room, with an excellent setup (monitors, conversion etc) ... so I'll check that later in the studio. For now I can only say that there is no audible difference on my crappy laptop speaker ...

So after the buy I hope my clients won't listen to my stuff on there computers

[funka]

so, what we are hearing is the AD convertion of 2 tracks stereo, right?

Yes, the Studer A810 tape is a stereo one.
We converted this tape analog stereo signal with each converter straight into PT.

I was wondering if the results would be much different if the test was made with, lets say 12 or 16 tracks going from tape to AD.
would the diferences remain so subtle as the track count sums up?

Maybe not...to test that, we would need to do the same with a multi-channel tape and multi AD converters(but it is harder to gather several channels of some high end converters).

I agree it is a tough listening test(many files), and to hear subtancial differences, you must listen to the files with a good monitoring chain otherwise, the subtilities would be less hearable...


Keep up posting your thoughts!

[Bill Moriarty]

Let's get those answers!!! I really want to know which is the SSL

[Smithers]

Am I the only one who care's that there is a human being with access to 13 of the top converters in one sitting?!?!?!

There are pictures of converters on crates with CD cases as spacers!!!!!!

Pictures!!!... Converters!!!... Many!!!!.. Same room... In a row... With PICTURES!!!!!

I'm supposed to concentrate on listening to audio files after seeing that???tutt

um, this is GEARSLUTZ, not EAR SLUTZ...... priorities people, priorities.

[Tom Hakala]

Originally Posted by Bill Moriarty

Let's get those answers!!! I really want to know which is the SSL

Let's wait... I wanna hear more of people's opinions.

Great test funka!!

[jazzymike]

greaaat test
I sort of like 6c best, both spatially and in terms of frequency separation and fullness.
Some certainly clip sooner/harsher than others.


Looking fwd. to the unraveling

[funka]

May I mention that most of the converters were only borrowed from friends or distributors.

My contribution was the aurora16 and Forssell MADC-2.

The studio owner where the tests were done has 24 ch of Studer D19 and the digi 192.

Sonosax was coming from a friend who mainly works in cinema world.

Let's wait... I wanna hear more of people's opinions.

Yes, I am sure we will get many more opinions...It would be sad to reveal all that early!
Moreover, it takes time to download the files that are huge, and it also takes time to listen to them and make an opinion!

[studio freak]

Originally Posted by funka

Yes, the Studer A810 tape is a stereo one.

of course!!! my fault, I was reading in a hurry,
I´m feeling really stupid now!!!


pleasssee, dont take too long, give us the answer´s

[space2012]

Originally Posted by Bill Moriarty

Let's get those answers!!! I really want to know which is the SSL

in this test the cables are too long, and digital cables are not Silver or Silver plated true OFC, makes them sound similar.

in other tests the SSL sounds inferior to Aurora and Lucid:
- A-D Conversion Test & Redux
- SSL Alpha Link AX vs. Apogee Rosetta 800 vs. Lucid ADA8824

[lowswing]

very interesting

[trubac]

If I may say something : it is quite difficult to ear any differences in a "not ultra carefull" listening, to me they all sound the same at first glance, a little bit like high end clean preamp shootout on one single take ...
Like the preamps the differences are really significative when you add more tracks(or passes) of a same preamp(A/D/A),it would be , IMHO, more easy to compare if the audio has been encoded several times with the same unit :
Audio ---> A/D ---> daw ---> D/A ---> Audio (1st gen) ---> ....---> Audio (10è gen) ...
This the main issue with poor converters : you can't go in and out your daw several times without destroying your audio.
I know it means D/A also and that is not the prupose here, but all the machines there can do it, no ?

Gilles

[didier.brest]

In the 44.1 kHz set (c), I select 1, 2, 3, 8, 10, 13. In the end, I might prefer 3.