You are here

Measurement based converters evaluation database

All about the tools and techniques involved in capturing sound, in the studio or on location.

Re: Measurement based converters evaluation database

Postby Del-Uks » Thu Aug 15, 2019 3:08 pm

Hugh Robjohns wrote:It's not a perfect indicator of perceived quality by any means, but it does tally pretty closely with my own perceptions...

So, here are my A-weighted AES17 dynamic range measurements (in decibels) for tested A-D converters, ordered from best to worst. Although these are all nominally 24-bit converters, the equivalent conversion word-lengths run from 20.6 down to 17.1 bits...

    RME ADI-2 Pro 124.0
    Lavry AD11 123.0
    Universal Audio 2192 122.0
    Lynx Hilo 121.3
    Current Focusrite ISA card 121.3
    Merging HAPI 121.0
    Grace Design M108 120.5
    Apogee Symphony 120.0
    Orignal Focusrite ISA card 119.5
    Prism Lyra 2 118.0

[...]

For A-Ds, I would categorise converters below 106dB as using antiquated technology or having a compromised or poor design. A-D converters measuring between 105-111dB are decent project and semi-pro level, while anything devices achieving 112-118 dB can be categorised as an excellent professional-quality A-D converter.

Anything in the 120dB region is a genuine flagship high-end product, but A-Ds exceeding 121dB define the current state-of the-art.

Here are the A-weighted AES17 dynamic range figures for tested D-A converters (in decibels). Again, the equivalent word lengths run from 21.4 down to 17.4 bits.

    Apogee Symphony 129.0
    Merging HAPI 126.3
    Benchmark DAC2 HGC 125.0
    Universal Audio 2192 125.0
    RME ADI-2 Pro 121.0
    Antelope Eclipse 384 121.0
    Lynx Hilo 120.5
    RME ADI-2 DAC 120.1
    Grace m905 119.7
    Focusrite RedNet 119.0
    Crookwood M1 119.0
    Focusrite Forte 118.0
    UAD Apollo 118.0
    Benchmark DAC1 117.6
    Drawmer HQ 117.0
    Prism Lyra 1 116.0

For D-As, the AES17 numbers dynamic range are generally a little higher, but anything managing more than 125dB is exceptional. 116-122dB is genuine top-notch professional quality, while 110-116 is good solid project studio fayre. There's no excuse these days for anything achieving less than 110dB...

Dear Hugh,

My beloved UAD 2192 has served me well for 15 years, but this summer’s extreme heat in Europe has killed it (melted caps and burned matrix board).

According to my technician, there’s nothing we can do to repair it. He also suggests me to look for something else since the UAD 2192 is not made anymore and most parts are very difficult to find (PSU is a well known exemple of potential disaster to come).

At first I was aiming at the Prism Lyra and the Lynx Hilo, but then I found your review of the RME ADI-2 Pro (pre-FS)... and needless to say that I’m very tempted by this (much) cheaper option.

As you know the UAD 2192 was not the most neutral converter by design... and I might say that’s probably why I liked so much and I didn’t look for any replacement until my unit dies.

I’ve read some ADI-2 Pro’s filter options are more "musical" than others... at some point I was also considering adding a Portico 5042 (at 15 IPS with the saturation knob fully counter-clockwise) in my signal flow in order to get a fuller character overall (more "analog").

What would you suggest to help me to get a sound closer to my late and sorely missed 2192.

Any other ADDA converter I should look at ?

Thank you for your attention.

Cheers,

Aleks
Del-Uks
New here
Posts: 4
Joined: Wed Nov 16, 2016 6:28 pm

 


Re: Measurement based converters evaluation database

Postby Hugh Robjohns » Thu Aug 15, 2019 5:22 pm

Del-Uks wrote:At first I was aiming at the Prism Lyra and the Lynx Hilo, but then I found your review of the RME ADI-2 Pro (pre-FS)... and needless to say that I’m very tempted by this (much) cheaper option.

The Lynx Hilo is much more than just a converter, so you need to ask yourself if you'd be paying for facilities you don't need or want with that one.

The Prism is a good converter, of course, but no one matches the track record of driver support that RME has, so if the ADI-2 or ADI-2 Pro does what you need, then I'd endorse that decision without hesitation.

As you know the UAD 2192 was not the most neutral converter by design... and I might say that’s probably why I liked so much and I didn’t look for any replacement until my unit dies.

If you like a bit of character in your converters, then I'd suggest checking out the Burl converters.

H
User avatar
Hugh Robjohns
Moderator
Posts: 24324
Joined: Fri Jul 25, 2003 12:00 am
Location: Worcestershire, UK
Technical Editor, Sound On Sound

Re: Measurement based converters evaluation database

Postby MOF » Thu Aug 15, 2019 6:11 pm

the Prism Orpheus doesn't come out especially well in the AES17 tests, yet it sounds stunningly good

Could this be due to some companies’ obsession with a ‘ruler flat’ frequency response in the design of their analogue stages causing time domain/phase issues? So on paper they are better but the listening tests show otherwise.
The analogy is a bit like trying to do room/speaker compensation with a graphic eq.
MOF
Regular
Posts: 356
Joined: Thu Mar 06, 2003 1:00 am
Location: United Kingdom

Re: Measurement based converters evaluation database

Postby Del-Uks » Thu Aug 15, 2019 6:13 pm

Thank you very much Hugh for your quick and wised answer.

Naturally, I've been looking at Burl (since Rich Williams also designed the UA 2192), but I really need to get a more compact (and less expensive) solution.

A single ADDA converter unit is what I'm looking for.

I think I will get the ADI-Pro 2 and I'll see for myself if I'm gonna need to find a way to add the right amount of character (warmth) to it.

According to your experience, which ADI's filter option would you suggest ?

I plan to use it in conjunction with a RME UCX interface... I guess the ADI-Pro 2 FS should be the master clock, right ?

Thanks again for your precious help.

Cheers,

Aleks
Del-Uks
New here
Posts: 4
Joined: Wed Nov 16, 2016 6:28 pm

 


Re: Measurement based converters evaluation database

Postby Trevor Johnson » Thu Aug 15, 2019 7:16 pm

I have owned a Lynx Hilo for about six years and as Hugh said, it does much more than its role as an A/D:D/A converter. Compared to other good converters, I find it has a very neutral sound, so is probably not what you are looking for. Although for me, I prefer to introduce 'character somewhere else in the recording/playback chain.

Incidentally I note that you have had problems with extreme heat. The Hilo normally runs fairly warm and that is in a very well ventilated position (not rack mounted), so possibly something else to bear in mind.
Trevor Johnson
Regular
Posts: 206
Joined: Sat May 15, 2010 12:00 am

Re: Measurement based converters evaluation database

Postby hobbyist » Thu Aug 15, 2019 7:24 pm

Mr.Michaelz wrote:Hello everyone,

I view converters as a simple studio commodity, something that should be technically accurate.
While it's up to every individual to assess if a given device will fit their workflow, I cannot see why it would be impossible to compare converters on a technical basis ; maybe with something like loopback tests performed in a standardized and stable environment.

Does something like this already exist?

You can certainly identify the technical factoids for all the converters along with cost, reliability , risks, and other factors.

But mathematically there is no logical way to order them so A>B>C....

You could weight the parameters to match how your personal risk factor or other criteria to come up with a list of 'best'. But they are only best for that particular set of weighting factors.

With cost being high on my list I wonder who needs >106 dB DR as one esteemed poster suggests.

With a -36dB headroom there is still 70dB DR that will below my ability to hear. And recording more than 40dB below my peak would be wasted anyway. In the end I will compress it to 20dB max and more like closer to 15dB DR or somewhat less.

So reliability , distortion, cost , and other factors would tend to outweigh the extra bit depth and DR and SNR that could be done with the higher end gear.

YMMV
hobbyist
Regular
Posts: 294
Joined: Mon Jul 01, 2019 12:52 am

Re: Measurement based converters evaluation database

Postby Del-Uks » Thu Aug 15, 2019 7:42 pm

Trevor Johnson wrote:Although for me, I prefer to introduce 'character somewhere else in the recording/playback chain.

Sure thing, but the UAD 2192 has some very pleasant "analog mojo" I'd like to retrieve somehow with my new converter.

I will experiment with the Portico 5042 and/or the Chandler TG2 preamp the A‑D side of the ADI-2 Pro FS when it arrives...
Del-Uks
New here
Posts: 4
Joined: Wed Nov 16, 2016 6:28 pm

 


Re: Measurement based converters evaluation database

Postby Sam Spoons » Thu Aug 15, 2019 11:27 pm

I'm not sure I see the fuss about converters, surely they should simply convert analogue to digital with little or no affect on the audio (i.e. if you pass some audio through a AD/DA the output should sound exactly the same as the input)? Converters that introduce 'character' are actually lower 'fi' than those that don't?
User avatar
Sam Spoons
Jedi Poster
Posts: 10012
Joined: Thu Jan 23, 2003 1:00 am
Location: Manchester UK
Finally taking this recording lark seriously (and recording my Gypsy Jazz CD)........

Re: Measurement based converters evaluation database

Postby Martin Walker » Fri Aug 16, 2019 10:52 am

Sam Spoons wrote:I'm not sure I see the fuss about converters, surely they should simply convert analogue to digital with little or no affect on the audio (i.e. if you pass some audio through a AD/DA the output should sound exactly the same as the input)? Converters that introduce 'character' are actually lower 'fi' than those that don't?

I agree with your basic premise Sam, that converters should ideally be transparent (for D/As that goes without saying, because you need to exactly what you're mixing) , but I'm not sure I'd go as far as saying that those A/Ds that advertise 'added euphonic mojo' are actually 'lower-fi'.

Some people will pay a lot of money for extra hardware mojo - personally I'd much prefer to add that via a preamp rather than A/D converter so it's an optional extra when required, but everyone is different (thank the gods! ;) )


Martin
User avatar
Martin Walker
Moderator
Posts: 14235
Joined: Wed Jan 13, 2010 9:44 am
Location: Cornwall, UK

Re: Measurement based converters evaluation database

Postby Sam Spoons » Fri Aug 16, 2019 6:15 pm

Martin Walker wrote:
Sam Spoons wrote:I'm not sure I see the fuss about converters, surely they should simply convert analogue to digital with little or no affect on the audio (i.e. if you pass some audio through a AD/DA the output should sound exactly the same as the input)? Converters that introduce 'character' are actually lower 'fi' than those that don't?

I agree with your basic premise Sam, that converters should ideally be transparent (for D/As that goes without saying, because you need to exactly what you're mixing) , but I'm not sure I'd go as far as saying that those A/Ds that advertise 'added euphonic mojo' are actually 'lower-fi'.

Some people will pay a lot of money for extra hardware mojo - personally I'd much prefer to add that via a preamp rather than A/D converter so it's an optional extra when required, but everyone is different (thank the gods! ;) )


Martin

Well, I was being slightly mischievous I guess (probably should have added a smiley), but, 'fidelity' is defined as "the degree of exactness with which something is copied or reproduced." so altering it by adding 'mojo' must be reducing the fidelity ;)
User avatar
Sam Spoons
Jedi Poster
Posts: 10012
Joined: Thu Jan 23, 2003 1:00 am
Location: Manchester UK
Finally taking this recording lark seriously (and recording my Gypsy Jazz CD)........

Re: Measurement based converters evaluation database

Postby Watchmaker » Sat Aug 17, 2019 11:53 am

In a pre amp, or many other places, its easy to adjust gain and alter harmonic content - in a convertor, my limited understanding leads me to think whatever color is there is a fixed "value." How much variability in character is achievable in conversion, and where would it theoretically be best placed - A-D or D-A?

Given that "sounds like tape" means largely compression artifacts, with some other "mojo," I'm not all that interested. Now if there were parameters on convertors I could alter that would give me some level of artistic control, that I could see being useful, but a color that's simply printed on everything I do is not appealing.
User avatar
Watchmaker
Frequent Poster
Posts: 613
Joined: Wed Apr 13, 2016 12:00 am
Location: Upstate NY, USA
Take my advice, I'm not using it.

Re: Measurement based converters evaluation database

Postby Del-Uks » Sat Aug 17, 2019 1:36 pm

Watchmaker wrote:Now if there were parameters on convertors I could alter that would give me some level of artistic control, that I could see being useful, but a color that's simply printed on everything I do is not appealing.

You might want to (re-)read Hugh's review of the UAD 2192.
Del-Uks
New here
Posts: 4
Joined: Wed Nov 16, 2016 6:28 pm

 



Who is online

Users browsing this forum: No registered users