D/A A/D Converter. Essential?

[atticus]

Just in case anyone is curious we have some comparison tests posted on our website in which we tested several a/d converters in a calibrated test. Here is the link:

MYTEK DIGITAL USA

I'm only posting this link because Mytek's name has been brought up several times and the files located on our site are perfect for the type of test being discussed. Thanks!

[desotoslo]

Originally Posted by atticus

Just in case anyone is curious we have some comparison tests posted on our website in which we tested several a/d converters in a calibrated test. Here is the link:

MYTEK DIGITAL USA

I'm only posting this link because Mytek's name has been brought up several times and the files located on our site are perfect for the type of test being discussed. Thanks!

The test we're talking about here is Mytek vs. Digi 002 converters/M Audio Delta, etc.

That's why this is such a heated debate. Ethan's premise is that the converters would make 0.4% difference. (or is that number based on different room sources, Ethan?)

I don't think we're not interested in Mytek vs. Prism/UA, etc. That's like apples with apples.

Though maybe your Mytek vs. Protools HD test would be good.. I haven't listened to that one..? When you say Protools HD what AD converter is used? (excuse my naivety here, I'm a happy logic user)
...

A really good test for this thread would be to have a heavily multitracked recording done in the same room through the same mics at the same time, split, into two wildly different (price wise) conversion systems, like Digi 002 and Mytek.

Then, comparing those results.

David

[masp84]

I was ready to shell out 2Gs for ADC but after hearing mytek shootout, I don't know if somethings wrong with my ears but I can't hear any difference at all, maybe, Mytek is slightly (0,4%) clearer. My money goes to room treatment god knows I really need it, Then I'll think about ADC-DAC.

I listened the clips through Saffire LE Analog Ouput > Event TR8 Monitors > untreated unbalanced room, maybe that kept me from hearing the difference, would it?

[desotoslo]

Originally Posted by masp84

I was ready to shell out 2Gs for ADC but after hearing mytek shootout, I don't know if somethings wrong with my ears but I can't hear any difference at all, maybe, Mytek is slightly (0,4%) clearer. My money goes to room treatment god knows I really need it, Then I'll think about ADC-DAC.

I listened the clips through Saffire LE Analog Ouput > Event TR8 Monitors > untreated unbalanced room, maybe that kept me from hearing the difference, would it?

Look, on that website, the shootout files are between Mytek converters and other high quality professional converters... of course the differences will be minimal.

I am able to clearly hear differences between the tracks, but I am also using pro quality headphones.

And yes, your untreated room will make a difference of course... but listen closely. Even in an untreated room (not a complete tragedy... come on, guys... ) you should be able to hear subtle differences between the files.


What I and others would like to know is if Mytek AD makes only a 0.4% improvement over say, the AD of a Digi 002 interface.

That's what's on the table. Am I right?

[atticus]

I'm sorry if my post was mis understood a bit. The samples on the site include the information of the SACD titles and player that was used as the source for the files. If anyone happens to be out there that has those same SACD's could use them to compare their own converters with the files on our page. Thanks!

[ISedlacek]

Originally Posted by Ethan Winer

Seriously, outboard A/D/A is probably the last thing you need to worry about, and it comes way after having good microphones and good monitoring in a good room.

--Ethan

I have to respectfully disagree and I feel this is one of the surviving myths in professional audio. In my practical experience, AD DA has practically the same importance for the resulting sound quality as mic, preamp and room. What is the use of a great mic and preamp when a mediocre converter is not able to translate all the subtle nunances in the digital domain and the original sound loses details and depth, with some extra disturbances ?

For sure, if we talk about low/mid end converters, it really does not matter whether you use Digi 00x, MOTU etc. etc. It is really the same. BUT - if you switch from Digi to say Mytek or Lavry Blue or Black, you suddenly get into a new sound dimension (using the same mics, preamps etc, as before). A small revolution happens.

And if (keeping the same stuff) you change from Lavry Blue to Gold or Prism or DAD AX-24, the revolution will be even bigger and quite amazing... I myself have experience a sudden jump from Blue to AX-24 recently and it brought me in a completely new sound dimension and league within my studio.

Convertors are indeed one of the essential parts of the audio chain and not less important than the rest ... The better convertor you have, the better sound results you will get (of course, in conjuction with the other parts of the chain)

[Ethan Winer]

Originally Posted by atticus

the files located on our site are perfect for the type of test being discussed.

Excellent David! Thanks for posting that link. Yes, these are exactly the types of comparison people need to hear in order to have informed opinions.

--Ethan

[TurboJets]

Originally Posted by Ethan Winer

Excellent David! Thanks for posting that link. Yes, these are exactly the types of comparison people need to hear in order to have informed opinions.

--Ethan

So, what a person hears in their own experience in their own studio (and in other studio's) is not an "informed" opinion?

There's a vast difference Ethan between the phrases "in my opinion" and "in my experience" and I'm confused at this point as to which one you're basing your rhetoric in; experience - or opinion?

[Ethan Winer]

Originally Posted by ISedlacek

In my practical experience, AD DA has practically the same importance for the resulting sound quality as mic, preamp and room.

Well, let's examine this in detail. Here's a graph that shows the bass response in a typical untreated room:



Besides the severe peaks and nulls, there's also modal ringing at the peak frequencies which is not shown here. The typical 15 to 30 dB spans from each peak to the adjacent null can be reduced to less than 10 dB with bass traps. But even then the response is seriously skewed.

Now, if a manufacturer released an A/D/A converter with a response like this they'd be the laughing stock of the industry. Indeed, even a $2 sound card chip built onto a cheap laptop's motherboard will be literally 1,000 times better than what this room does to the sound. Yet you say this is not more of a problem than conversion?

--Ethan

[Ethan Winer]

Originally Posted by TurboJets

So, what a person hears in their own experience in their own studio (and in other studio's) is not an "informed" opinion? There's a vast difference Ethan between the phrases "in my opinion" and "in my experience" and I'm confused at this point as to which one you're basing your rhetoric in; experience - or opinion?

I don't understand your point or your motive. Opinions are formed from experience, no? Did you listen to the files David posted? What exactly are you disagreeing with?

--Ethan

[TurboJets]

Originally Posted by Ethan Winer

What exactly are you disagreeing with?

--Ethan

You.

[jslevin]

Originally Posted by Ethan Winer

Now, if a manufacturer released an A/D/A converter with a response like this they'd be the laughing stock of the industry. Indeed, even a $2 sound card chip built onto a cheap laptop's motherboard will be literally 1,000 times better than what this room does to the sound. Yet you say this is not more of a problem than conversion?

Yes, but you could say the same about many well-treated rooms — frequency response variation that would be considered quite reasonable or even excellent for an acoustical design would be considered horrendous for a converter.

The fact remains, however, that these are separate issues. As critical as monitoring environment is — and with everyone in the world except you, Ethan, I'm on the other side of this argument — we all acclimate to our monitoring environments, but you never acclimate to a mediocre ADC. A good monitoring environment can help you overcome recording issues, because at least you can hear them and adjust. But it can't help you overcome poor ADC, excessive jitter, weak phase coherence, etc. It's just sitting there, on your tracks, forever.

It's kind of like comparing an AT4047 with a real FET47 design, like a Lawson. Just as with converters, the frequency response differences would be absolutely trivial when compared to the graph you posted above — hell, forget the AT4047, that's a nice little mic, you could probably find an LDC for $100 that had a trivial response difference from the Lawson, relative to the response differences in your graph. Trivial! But to the ears? To the music production process? Certainly not trivial.

I'm not sure why this has to be an either/or for you. Many folks don't spend 1/10th the time on acoustical issues that they should, I'm totally with you on that, and in fact you're one of the main people who helped me understand it. But that doesn't mean that conversion issues aren't important, tangible, measurable and worth addressing — even if they measure out to being far smaller in absolute terms than acoustical issues.

The difference between relatively good converters — an RME Fireface and an Apogee Rosetta, just to throw out two — is still very significant, and any of my engineers could have told you the differences between an AD-16x and a Mytek 8X96 after we switched. Beyond that, however, the difference between mediocre conversion and outstanding conversion is absolutely enormous in its own right — regardless of whether acoustical issues are enormous at a different and larger scale — and absolutely worth addressing.

And if you can't hear that ... well, I'd still love to hire you to treat my room, but you sure won't be allowed to engineer anything there. When you're trying to make great recordings of great musicians, measurements and specs alone simply won't do.

JSL

[stevens119]

I just wanted to add my two cents. I have an Mbox 2. I also have an Apogee Mini-DAC. There is more than a considerable difference I hear when the DAC is in the chain. It makes my mackies sound like a pillow has been taken off of them. It's that big of a diference. I would say, if it's affordable to you, then get some good converters. It will make all your decisions better while recording and mixing. I've noticed that my better converters casue everything I do to sound better.

[ISedlacek]

Originally Posted by jslevin

The difference between relatively good converters — an RME Fireface and an Apogee Rosetta, just to throw out two — is still very significant, and any of my engineers could have told you the differences between an AD-16x and a Mytek 8X96 after we switched. Beyond that, however, the difference between mediocre conversion and outstanding conversion is absolutely enormous in its own right — regardless of whether acoustical issues are enormous at a different and larger scale — and absolutely worth addressing.

Amen ...

If you have perfectly treated room, great mics and preamps and mediocre converters - it will sound OK. But if you switch to absolutely top, outstanding converters (Like Prism or DAD AX-24), the overall sound will make a grand leap to a new level. Everything (including monitoring) will sound A WAY better.

So, to reply to the original question: yes, AD-DA conversion quality is absolutely essential, the same important as everything else in the chain. And without caring of that as well, other changes in the recording chain may not be fully effective

Doubting that is simply denying the truth and practical experience that is easily available to anyone.

[Ethan Winer]

Originally Posted by jslevin

frequency response variation that would be considered quite reasonable or even excellent for an acoustical design would be considered horrendous for a converter.

Exactly! And that's my whole point. Your speakers and room mess up the sound far more than conversion. Now, I do agree with you that your monitoring affects only what you hear, where conversion affects what everyone hears. Once it's recorded, there's no undoing bad sound.

I guess my main objection is all the opinions I see tossed around about how incredibly important converter quality is. In my experience the differences between working (not broken) converters are far smaller than many people report. In another Gearslutz thread going on right now (sorry, too lazy to track it down) converters were compared in identical situations, and most of the replies used words such as "very subtle" versus the words I see when converters are not compared in a controlled environment. This is why again and again I harp on the importance of highly controlled listening tests, rather than anecdotal reports comparing different performances or even different recording sessions.

But that doesn't mean that conversion issues aren't important, tangible, measurable and worth addressing — even if they measure out to being far smaller in absolute terms than acoustical issues.

I agree fully. Where our opinions part is exactly how wide a range there is between good and bad converters. But we don't have to agree on this, and hopefully we can still be friends.

--Ethan

[Lupo]

Originally Posted by Ethan Winer

Your speakers and room mess up the sound far more than conversion.

Hmmm.. Does the room response affect -100dB or smaller signals?

[Ethan Winer]

Originally Posted by Lupo

Hmmm.. Does the room response affect -100dB or smaller signals?

Great point. Room acoustics are linear, but anything that soft will be under the ambient noise floor and thus inaudible. For the same reason, converter artifacts that soft are also inaudible unless you raise the volume unnaturally on playback.

Just because we can measure something doesn't mean it is audible or matters.

--Ethan

[manning1]

ethan.
ive always had lots of respect for your positions on various subjects.
frankly for the longest time ive been trying to figure out just how much one should spend on quality ADA.
(my degrees in physics...so i'm a bit of a sceptic bout various things.)
thus my question.

for someone like myself with a small studio, that wants quality without spending ones retirement fund.
what would YOU say is good quality...without spending lots.
echo audiofire ? that some people on GS seem happy with...?
would you say its the same situation as with hi fi systems in the 70's..?
ie...50 buks...not too good...500 buks...normally good...
but to get that extra 5 per cent one has to spend 5k ?


i'm curious on your opinion. is an emu 1616m good enough ?
for example ?


its a pity noone has done a test whereby a vocal or a stereo orchestra
recording was sent to 3 different computers and recorded.
one with low end ADA, one with middling, and one with hi end ADA...
so that people could compare the wave file playback of each and assess for themselves objectively.

this is a very difficult topic for lots of folks...ie...how much does one need to spend to get
quality. mebe top audio engrs can notice subtleties in ADA,,,,but i'm not writing songs for
other audio engrs...but for "joe consumer".
so the uppermost question in my mind is will joe consumer notice much difference from
say your delta 66 used to record and ADA costing a couple of g's.
your input is very welcome on this.

[jslevin]

Originally Posted by Ethan Winer

I agree fully. Where our opinions part is exactly how wide a range there is between good and bad converters. But we don't have to agree on this, and hopefully we can still be friends.

Oh.

Okay.

JSL

[fova]

For a long time I was using in my studio M-Audio Audiophile 192 and was happy with it because I thought that AD/DA converter is not as important as room treatment, microphones, preamp and of course the interpreter :-)

I solved everything other - created a nice collection of decent microphones, bought a middle class preamp, built up bass traps but still was using A192 card.
Till the moment when got an opportunity to buy Lavry AD10 converter for a good price.
When I tested the sound of both units (A192 and AD10) with the same microphone and preamp I was shocked how big was the difference in sound. AD10 brought all small details such as sound of the PC cooler and the overall performance was full and mighty. On the othe hand A192s sound was flat and lacking all those details.

I placed records of both units on one czech musical forum and users there confirmed that there is a difference between them that is comparable to difference between two microphones!

I had to change my mind comletely in this issue. Choice of AD/DA converter is definitely a very important part of the chain.

[Lupo]

Hi Ethan!

Originally Posted by Ethan Winer

Great point. Room acoustics are linear, but anything that soft will be under the ambient noise floor and thus inaudible. For the same reason, converter artifacts that soft are also inaudible unless you raise the volume unnaturally on playback.

What I meant is - if a sound leaves the speaker at very low level - will the room acoustics make that sound less audible?

We've been looping around this theme for months now and we've both exclaimed our stance on that one. Let me repeat my train of thought regarding the audability of -100dBFS artifacts.

Maximum peak level is never attained in a recording. We don't listen to sine waves, and even those are 3dB lower than the peak(the crest factor). Most music have peaks at least 10dB below the ceiling. In soft music and quiet parts, which is where we can expect to hear any differences between gear, the peaks are often at about -50(measured on an FFT). That only leaves 50dB's to go down to -100 (FFT) where the artifacts starts to appear in normal converters! A 50dB dynamic range should be perfectly possible for anyone to hear..

If I recall correct from Moore's "psychology of hearing", the simultaneous dynamic range of human hearing is about 80dB. Meaning we can hear sounds being 80dB softer than another sound being present. The use of "filter bands" and limited crossmasking between those bands explains a lot of this. With FFT peaks at -50 for normal music, that means that we can hear all the way down to artifacts at -130 (FFT). If the gain is raised 10-20dB, which we may well expect someone to do with converters used for tracking and mixing (far from full scale) - the FFT measurements needs to be clean down to -150 to not have any audible effect..


So as you can see from the numbers above, my estimates for what constitues "clean" audio have only gotten worse since we started discussing this.


Best regards,

Andreas Nordenstam

[Ethan Winer]

Originally Posted by manning1

what would YOU say is good quality...without spending lots.

There are so many brands and models I don't even try to keep up. So I'm not in a position to pass judgment on specific models I don't actually own and use. I have mostly modest gear that has served me well, but I've heard plenty of really cheap stuff and plenty of very high-end gear. 30 or 40 years ago cheap stuff really was not very good, and I remember paying thousands of 1970s dollars for each piece of pro quality gear I owned.

The past 10 to 20 years things have change a lot, and what cost $2,000 in 1978 might be available today for less than $100. Life is good for recording engineers! There is no longer a reason you can't create a hit record using "prosumer" gear. What matters most these days, and still varies from model to model, are loudspeakers and microphones. Of course the room you record and mix in varies even more.

I'm a "meter reader" type, so I believe that as long as a vendor does not lie (which is not a given) there's no reason you can't buy based on specs. Obviously build quality and durability matter. I'd rather pay $250 for a DI box that I know will work when I need it, versus a Behringer for $30 that may sound okay (never tried one) but from all reports I've seen is not built to last. And you have to assess all the specs. As the Mercenary guys always point out, headroom is where a lot of low-end gear falls short. And output level is often conveniently omitted in many distortion specs.

its a pity noone has done a test whereby a vocal or a stereo orchestra recording was sent to 3 different computers and recorded.

The closest I've come to such a test is comparing a $25 Soundblaster card with a $6,000 Apogee as reported HERE.

The above test was flawed by a slight hum on the Apogee due to a ground loop, but I think it's still revealing. If you listen very carefully you may hear a slight difference between these devices. But in the large picture the differences are slight, and it's clear to me one could do professional quality work with either. My friend Grekim who owns the Apogee tested above plans to do a more thorough comparison soon with more A/D/A devices and a more expensive preamp, and I hope to be present so I can report the results.

Finally, I'd like to make an important point about my posts on this topic. I do not post just to make trouble or be contrary, even though many here seem to think that's my motive. My only goal is to get people to examine their own prejudices, and also to understand why anecdotal reports are inadequate for assessing gear quality.

--Ethan

[Ethan Winer]

Originally Posted by Lupo

What I meant is - if a sound leaves the speaker at very low level - will the room acoustics make that sound less audible?

Again, room acoustics are linear, so the only time something in a room becomes less audible after leaving the speaker is if it falls below the ambient room noise. Or if it happens to fall in one of the many null locations. This is why I tell people to test their rooms at fairly high SPL levels. Nulls can be 30 dB deep or even worse, so you have to be at least 30 dB higher than normal to measure (or hear) a frequency that is nulled by the room at that location.

Meaning we can hear sounds being 80dB softer than another sound being present.

Yes, but this is only under the most ideal circumstances! With most pop music you'll be lucky to pick out a soft closed hi-hat only 40 dB below the rest of the track if it plays on the beat.

--Ethan

[manning1]

Ethan.
thanks for the reference to the 3D thread.("here".).
very interesting read.
kudos to you.
given me lots of thinking to do and pause for reflection mate.