A/D D/A Converters?

[echoshock]

Hi all,

What is the difference between an A/D D/A Converter and an audio interface? (besides $)

They are listed as different categories in this audio catalog I'm looking through.

Is the difference really just quality of the converters?

Thanks

[MikeRivers]

 

Hi all, What is the difference between an A/D D/A Converter and an audio interface?

 

 

An A/D or D/A converter (or both in the same box) has analog on one side and one of the standard digital audio interfaces - S/PDIF, AES/EBU, ADAT Optical ("Lightpipe"), TDIF, MADI, AES50, etc - on the other side. If you have one of those, you'll need something with a digital interface to connect to it.

 

"Interface" in this context usually means that it's a box with an A/D converter, a D/A converter, and an interface to a computer, like USB, Firewire, or a direct bus interface like PCI or PCMCIA.

[echoshock]

I notice most of the a/d converters just don't have preamps. Is that one of the main differences with an 'interface'?

The reason I'm asking these questions is because I'm looking to improve the quality of my recordings. (as Angelo knows). Is it better to go with a so called a/d converter because its a box dedicated to that job as opposed to an interface which usually includes preamps and lesser quality a/d converters?

Thanks

[Ethan Winer]

Just to be clear, a sound card is an A/D/A converter. And most modern sound cards are perfectly adequate for pro quality recording. Here's a report on my comparison of a $25 SoundBlaster card with a $6,000 Apogee converter:

 

http://www.3daudioinc.com/3db/showthread.php?t=12836

 

I'm not saying the two are equal, but IMO they're so close that for all intents they might as well be equal.

 

--Ethan

[UstadKhanAli]

As someone pointed out in the thread that Ethan posted (Lynn Fuston's forum), and as I have noticed by intensive listening, differences in A/D become more apparent when you "stack" them. This makes sense since obviously, any obvious points of a converter (or pretty much anything else) become magnified when they are done numerous times.

 

When I recorded a single vocal track or a single acoustic guitar track through the Digi001 and through an Apogee Rosetta 800, the difference was apparent. Apparent, but not startlingly different.

 

When I transferred (from tape) the same performance of vocals, drums, bass, guitars, and keyboards through the Digi001 and Apogee Rosetta 800 converters and then listened, the differences were far more obvious...I mean, really startlingly obvious.

 

Any strong points, weak points, anomalies, etc. become more obvious when they are "stacked".

 

Now, I realize that in comparing a Digi001 with an Apogee Rosetta 800, I am comparing something that completely blows with something that sounds rather good. It's an extreme comparison. And I also realize that the Digi001 is "turn-of-the-century" technology, rather old in comparison. But it hopefully illustrates the importance of evaluating converters (or anything else, such as mic preamps) by listening to them when they are used over the course of a session, and not just when you are recording a single track. It can really become far more obvious what is going on.

 

There are some good quality mic preamps out there that don't cost an arm and a leg...some of the PreSonus stuff is highly regarded as good bang for the buck. You do have to listen, though. If you hear a slight difference in comparing one track, it just may be that over the course of twenty or thirty conversions, it might sound like a little more than a slight difference...

[UstadKhanAli]

Here's an example where a slight difference seems like it might just be that...a slight difference.

 

I have two FMR RNP mic preamps. And I have a Neve Portico mic preamp.

 

I've done whole sessions where I've used the FMR RNPs. Sounds great. And whole sessions where I've used primarily the Neve Porticos in clean mode.

 

And you know what?

 

There's not a giant huge difference.

 

You can hear a small difference between the RNP and the Neve in clean mode (the Silk Mode is not subtle, and there's a huge difference between the two there) when comparing a single track of vocals or acoustic guitar.

 

And over the long haul, I thought, "Maybe there's a larger difference when they are 'stacked'".

 

But no, there's not a huge difference.

 

And I mention this because it's difficult to evaluate from one track just how large a difference there is between two items.

 

In the case of the Digi001/Apogee comparison, well, it wasn't a tiny difference...comparing one track to another was still apparent. But nevertheless, it was a *MONSTROUS* difference when stacked. Totally not subtle, and not something for just Golden Ears. I mean, everyone noticed.

 

With the RNP/Neve (clean) comparison, the initial difference was very difficult to hear. I had to really play things back many times to hear the difference. Very subtle. And subsequent stacking bore this out. Subtle.

[MikeRivers]

 

I notice most of the a/d converters just don't have preamps. Is that one of the main differences with an 'interface'?

 

Yes, most "interfaces" offer a complete one-box solution, with mic preamps, sometimes equalization and dynamics as well, a headphone amplifier, and some sort of useful routing for monitoring. There's an article of mine (probably in two parts, with this addressed in the second part) coming out in Recording soon that's about this very subject.

 

Is it better to go with a so called a/d converter because its a box dedicated to that job as opposed to an interface which usually includes preamps and lesser quality a/d converters?

 

The "system" approach is always better if you can afford it. By using separate boxes for each function, you can choose your favorites rather than settling for the manufactuer's choices that he made to meet his chosen price point. And you can upgrade incrementally - when you want a better mic preamp, you can change just the preamp without changing the A/D converter.

 

Still, everyone has a budget and there are some very good "interfaces" out there that won't stand in your way of making good recordings.

[Ethan Winer]

differences in A/D become more apparent when you "stack" them.

 

Good point Ken. I guess now I have to do another SoundBlaster test recording ten passes in a row.

 

--Ethan

[Dean Roddey]

The gotcha with using a pure converter is that you'll probably connect it to the card in your computer using a one way protocol like S/PDIF or something. This may introduce jitter because any one way, non-packetized protocol like that requires that the receiving PC card sync itself to the incoming samples, whose timing it cannot control.

 

Some of these devices also have the option for you to put a firewire interface card into them. Then they become the 'sound card' as well and you avoid that extra jitter. Firewire is a two way, packetized protocol. It will add a small amount of latency to the system becaues the converter box packages up a small bundle of samples and sends them a packet at a time. But that means you get error checking and there's no need for the receiving PC to have to sync itself to an external signal clock.

 

Devices like the Apogee, Prism, Aurora, all provide Firewire options so that they can avoid the problem of multiple clocks. And, BTW, there's a number of discussions on GS about why external high end clocks are not a good idea as well, again because of the issue of having multiple clocks in the system and the jitter that introduces.

[UstadKhanAli]

Good point Ken. I guess now I have to do another SoundBlaster test recording ten passes in a row. --Ethan

 

 

Oh, no, I didn't create more work for you now, did I? Sorry 'bout that, buddy!!!

 

I think it'll make the listening tests all the more interesting...and wouldn't that be fascinating if you could stack ten or twenty tracks with your SoundBlaster, and it'd still sound really great when compared to Apogee? Then it'd be MY turn to cry!!!

[Ethan Winer]

Shall we put money on it Ken?

[UstadKhanAli]

I'm pretty sure the Apogees will still emerge triumphant...probably a bigger difference than with just a one-track comparison.

 

Still, though, wouldn't it be something if the SoundBlaster still sounded quite good in comparison, even after listening to 20 tracks "stacked"?

[Ethan Winer]

I just wasted half an hour trying to do this, but had trouble routing things in Sonar. I'll try again when I get a chance.

 

--Ethan

[Ethan Winer]

 

wouldn't it be something if the SoundBlaster still sounded quite good in comparison, even after listening to 20 tracks "stacked"?

 

 

Okay, I managed to get this to work by playing from Sonar and recording into Sound Forge. I loaded a tune fragment to a track in Sonar and played it through my Delta 66, which was patched into the SoundBlaster's Line Input. I recorded at 16 bits and did a total of 20 passes. After each recording I put the newest file into Sonar, then played that while making the next recording.

 

I would have rather played through the SB too, but couldn't get that to work. It's supposed to work! But the sound was hollow after only one pass, even though I had Input Monitoring turned off. Go figure. I wasted an hour in the SB's Control Panel trying to get a clean recording and finally gave up. But 20 passes through the SB's input A/D (only) is still telling.

 

I'll give away the ending. It still sounds like the same music after 20 passes, but you can definitely hear a change in quality. Here are four files, each 2 MB:

 

Original source file

After 1 pass

After 10 passes

After 20 passes

 

--Ethan

[blue2blue]

Just to be clear, a sound card is an A/D/A converter. And most modern sound cards are perfectly adequate for pro quality recording. Here's a report on my comparison of a $25 SoundBlaster card with a $6,000 Apogee converter: http://www.3daudioinc.com/3db/showthread.php?t=12836 I'm not saying the two are equal, but IMO they're so close that for all intents they might as well be equal. --Ethan

 

:snax::snax:

 

I'm sending out for pizza as I catch up to the rest of the thread...

[blue2blue]

Shall we put money on it Ken?

Methodology question: are you going to allow correcting for DC offset?

 

'Cause my ol' Blaster Lives (used them for testing and everyday listening as well as for their dedicated Soundfont DSP) used to tend to have a fair chunk of it.

 

 

_____________

 

 

UPDATE: All caught up now. I DL'd the original, the one-pass, and the after-20 pass snippets.

 

There's definitely a clearly discernible diff between the original and 20 pass versions -- but it's not nearly as bad as I would have imagined, all things considered.

 

LOVE to see some tape afficianadi try this on their Studers. Or their TASCAMs...

 

 

[i did some cheezeball 'double blind' testing -- first comparing the original, one-pass, and 20-pass (in a noisy environment/fan on in the other room) and the ability to pick out the correct one was pretty random. But when comparing only the original and the 20-pass using randomized playback (had to go to the silly Yahoo Music Engine player to get the kind of dumb azz 'real' random play behavior of playing the same song 2 or 3 times in a row -- WinAmp is too smart and will simply alternate between two tracks, even when set to random. No prob for YME, though, it's as dumb as a rock... no 'smart random' shuffle for the geniuses at Yahoo.) In that far less than ideal listening environment, I was able to discern the diff at better-than random -- but far from everytime. Clearly, in the course of my highly informal test, there was a fatigue factor -- I had a run of about 4 incorrect answers in a row.]

[Ethan Winer]

LOVE to see some tape afficianadi try this on their Studers. Or their TASCAMs...

 

 

[Ethan Winer]

LOVE to see some tape afficianadi try this on their Studers. Or their TASCAMs...

 

I was thinking more about that today. I've stated many times in forums that a $25 SoundBlaster card beats an $80,000 Studer in every way one could possibly assess audio fidelity. And every time I say that half a dozen people tell me I'm crazy. I don't have a Studer to do this test, and I think we all know the sound will be down the toilet after far fewer than 20 passes. But if anyone here has a Studer in good condition, I'd love to get a Wave file of the 20th generation of my original.wav file above! Anyone?

 

--Ethan