converter specs for all to read

tubedude

tubedude

New member
Heres some specs that may help you decide what you need to get to upgrade your sound.
By far the best deal I've seen is the M-audio Delta 1010.
Don't confuse the Delta 1010 converters with the other Delta cards... no comparison.
Best specs of all of these seems to be the Lucid, of course.. they're dedicated converters.
Aardvark 2496
a/d 100db... fair
d/a 110db... nice

Aardvark LX6.. same

Aardvark Q10.. no specs avail.

Aardvark Aark 24.. same again
All Aardvark cards are shielded heavily to avoid picking up noise. They seem to be the only brand to bother.

M-audio Delta 1010
D/A 108db... good
A/D 109db... good
excellent buy for the money, you get 8 converters

Delta 66
A/D 99db
D/A 103

Delta 44
Same as Delta 66

Delta DiO 2496
D/A 101
no A/D

M-Audio Quatro
No specs posted. If you encounter them, let me know.

M-Audio Audiophile 2496
A/D 99db
D/A 103

Lucid
A/D 2496 (A/D only)
114db... very nice

Lucid D/A 2496 (D/A only)
114db... very nice
Word has it that the Lucid are VERY top of the line.

Thats all for now, these are the important ones, in my opinion. I'll post more later when I have more time.
Peace,
Paul
 
cool man,

I also wanted to chip in that music recorded on the delta 1010 does seem to have this nice sound to it.

You can verify in the mp3 mixing clinic.
MrLipp
Emeric
Jon-x
Cyanjaguar
and many others

The converters just sound so sweet.so present

NOt to say that I wont be jumping on the lucid a/d as soon as I can afford it
 
Tubedude, is the difference between the delta 66 and 1010 the fact that the 1010 has the convertors outside the computer? Or are they actually totally different convertors?

I was sure they were the same convertors, but now, I'm not so sure.

Also, were those comparisons done at equal settings.....24bit, 96kHz......on all tested cards?

Romeo
 
Thanks for the info, but S/N ratio should not be your only concern. I've been looking at converters myself and more often than not it's what companies aren't specing that has me more concerned. You can have S/N ratio of 117db, Harmonic Distortion of 0.001%, but still have significant coloration of the sound. If the balanced input passes through a transformer, or capacitor, or if the analog logic is not well designed its possible for there to be distortion of frequency response and/or phase of the incoming signal.

For example, the Lucid AD2496 Users Guide, which has some pretty impressive specs says that frequency response of the converter is 20Hz-40Khz +/- 1db. That +/- 1db could mean there is a 1db drop in response at the extreeme frequencies, but it also leaves the door open for having a response curve is not completely flat. It could mean it has a 2db boost at 1 Khs for example. Most converters don't list clock jitter as a spec, they often don't qualify levels or frequencies at which the specs were measured, and the list of things not covered goes on.

One company that seem to be extreemely thorough in its specs is Apogee. The user guide for the PSX-100 provides the following specs. I've even edited this down quite a bit:

Dynamic range:
–60 dB, Unweighted 116 dB –60 dB, A-weighted 119 dB
Relative THD+N:
-0.1 dBFS –110 dB (unweighted) –112 dB (A-weighted)
Frequency Response (10 Hz–20 kHz):
Gain ±0.025 dB
Phase < 0.01 degrees
Passband Ripple: ±0.001 dB
Channel Separation: Left/right 120 dB (worst case)
Crystal Oscillator accuracy ±50 ppm
Clock Jitter, 32kHz–106 kHz << 22 psec

----------
Compare that to MOTU who provides as little info as possible (its 24 bits what else do you need to know). You can't download user manuals for their stuff online. I went looking for 1296 information all I could find on their web site was A-weighted dynamic range of 117db. Thanks, but that doesn't tell me much.

-----
Lucid is not too bad (both in price/performance and in the specs they provide), they lie somewhere inbetween Apogee and MOTU. They provide the following in their AD2496 users manual.

Dynamic Range: > 115dB, A-weighted, 48kHz sampling rate
THD+Noise: < 0.002% (-1dBFS out, 1kHz)
Frequency Response:
For sample rates of 88.2kHz and 96kHz 20Hz - 40kHz (+/-1dB)

And for the DA2496:

Dynamic Range > 114dB, A-weighted
THD+Noise < 0.002% (-1dBFS in, 1kHz)
Frequency Response
For sample rates of 88.2kHz and 96kHz 20Hz - 40kHz (+/-1dB)

-----
M-Audio Delta 1010

Dynamic Range:
Outputs: 108dB (a-weighted),
Inputs: 109dB (a-weighted)
THD (at 0dBFS):
Outputs: less than 0.0015%,
Inputs: less than 0.001%
Frequency Response: 22Hz - 22kHz, -0.3,-0.2dB

-----
Aardvark Ark 24

Dynamic Range: 100 dB A/D; 110 dB D/A A-weighted
THD+N: .002%, 20Hz-22KHz, A-weighted
Frequency Response: 7Hz-22KHz, +/- .5 dB

-----
Aardvark Pro 24/96

Dynamic Range: D/A 110dB, A/D 100dB
THD+N: .002% @ 1 kHz
Frequency Response: 7Hz-44KHz, +/- .5 dB at 96 kHz


----
Here are a few high end converters for comparison...

For your drooling pleasure (most of us can't afford $8000 just for a sterio AD), the award for most rediculously high S/N ratio goes to...
db Technologies AD122-96. Their web page lists the specs as.

Noise (true 21-bit noise floor):
-127 dBFS (10Hz to 20kHz unweighted)
-129 dBFS (A weighted)
-123 dBFS (10Hz to 48kHz unweighted)
THD+N over 20Hz to 20kHz:
121 dBFS (.00005%) for -40dBFS or lower
108 dBFS (.0004%) for -20dBFS signals
100 dBFS (.001%) for full scale signals
Bandwidth: 43kHz at 96kHz sample rate
Channel Separation: 120 dB at 1kHz
Flatness Response: 0.05dB (0Hz to 20kHz)
Phase linearity: 0.1 degree (0Hz to 20kHz)

-----
Prism Sound Dream ADA-8

Transformerless electronically-balanced XLR inputs and outputs
Input Dynamic range:
112dB typical, rms unweighted, measured at -60dBFS
Input THD+n:
-105dB (0.0004%) typical, rms unweighted at -1.0dBFS
Over-killer function provides soft-knee limiting preventing converter overload

Output Dynamic range:
10dB typical, rms unweighted, measured at -60dBFS
Output THD+n:
-104dB (0.0004%) typical, rms unweighted at -1.0dBFS

-----
Mytek Digital 8x96

Dynamic Range: 120dB A-weighted, 117dB Total
THD+Noise: -105dB (<0.0005%)
Internal clock jitter: <10picoseconds

-----
Swissonic AD96

Frequency response
Passband: 2–22100 Hz (FS= 48 Khz)
passband scales with sample rate
Passband ripple: 0,01 dB
Group delay variation with freq: 0 linear phase
SNR & dynamic range: 115 dB (A-weighted)
THD+noise: -100 dB (@-1 dBFS)
THD: 0,003 % (@-1 dBFS)
Inter-chanel phase deviation: 0,1 degrees
Inter-channel isolation: -110 dB

-----
I've run out of time or I'd have included more, perhaps later....
 
123

Ah, good posts... I like to get people thinking. Lets get this one to the point that people can reference it and save themselves days of work. :)
Paul
 
Also, everyone need to be reminded that even if you have the card with the best specs, going in with a Radio Shack mic thru the preamps of your old 4 track with crappy ass cables wont sound all that good...every link in your signal chain counts, down to the cables......
 
True Gidge, but the recording of that Radio Shack mic through high end gear elsewhere in the system is going to far more accurate then it would with a Neuman U87 being recorded on a Radio Shack mini cassette recorder.

Microphones are about color, more so then ANY preamp is.

Now, about this subject.

If you all are believing specs from the manufacture, you are crazy as hell!

I did a little mastering job using the Delta 1010 converters for the D/A. I was sorely disappointed because when I took the .wav files home and listened to them on my Lynx One card, I found out that the audio was fully -1dB down from what it could have been, even though it was peaking on the Delta converters!!!

To explain further.

While mastering, I was applying the Waves L1 Ultramaximizer to the mix. While I reviewed the metering, I was noticing that the sound started to distort a bit lower then I was accustomed to, but I was HEARING distortion, and regardless of what I seen, I have to go with my ears right? Of course.

I notice on playing the CD in my truck that the audio is not distorting by a long shot, and I compared the product to stuff I mastered using my Lynx Card as D/A. Guess what, the stuff mastered using the Delta converters was noticeably quiter.

I get home and transfer the .wav files onto my computer and take a listen. FAR from distorting! In fact, I did a Peak Search using Wavelab and found that I NEVER hit digital 0 once in the mix!!! I look at the metering on the Lynx One meters and it shows that I am peaking at around -1dB at best!

My oh my........

I proceed to load up the unmastered .wavs and apply the plug in's I used again, this time using the Lynx One card as the D/A. I couldn't get EXACT settings because I saved those on the other computer and didn't think to copy them to the DATA disk. Anyway, I was still pretty fresh from the session, and was able to replicate the eq curves. Where the big difference existed was in the overall level I was able to achieve before distortion ensued on the Lynx One card.

The resulting .wav file was much hotter and sounded much more like what I was expecting it too.

I would not consider the Delta 1010 to be in league with professional sound cards (professional meaning what a discriminating professional would tolorate!). If a certain member on this BBS chooses to chime in here, he will attest that his Delta 1010 is not exactly "all that". He is claiming to hear a difference between monitoring a mix before recording and playing the mix back after recording. That is something I only experienced with cheaper converters and haven't had to deal with for quite some time now because regardless of the "specs" on the Lynx One card, it is a fine sounding card.

About Apogee. Not all would agree that they sound as good as Lucent, and the reverse applies. Both are very nice chipsets. I have not worked with the Lucent converters to comment, but some very trusting ears have given them very high marks indeed! I have used Apogee, and would not discount that purchase either. Before spending that kind of money though, I would owe it to myself to listen to both Lucent and Apogee and see which flavor suits me. After doing some work on Apogee's, I am STILL open to another converter.

Don't believe spec's unless they come from independent sources. Manufactures are usually not in the business of showing off weak areas of their product. And when it comes right down to it, specs mean little in audio once a certain minimum level has been reached.

Ed
 
Ed, that is exactly why the Lynx card goes for 3 times more $$$ than a M-Audio card......
 
the fact stands. 8 A/D, 8 D/A, the BEST headroom in prosumer/homerecording. i noticed the clipping deal as well. my delta is very sensitive to exagerating peaks. how to solve this? believe your ears not the meters.

the other solution? i have 4 channels of compression. you can take away all the peaks and it will not be "compressed" per se, and then you bost the output. its quite simple. and it works like MAGIC. now, you have 16 total channels that are -1dB down (peak) from what the would if you could afford 4 lynx one card. but when you mix 16 channels, the -1dB doesnt matter as much because of how signal levels increase with the adition of channels.

just use 2 channels of the 1010 and yes, especially if you arent sending it though a limiter, or something to tame the peaks then it wont compete. but layer upon layers? and it defintely WILL compete.

somusman. i know you work for money in your studio/live. but if you made a trial with the 1010 and recorded a 16+ track song on a computer. i think you would find youself please with the results. HALF the price of XTs. lets us spend money on a mic, a pre, monitors, etc.

me with an SSL and a 2" studer vs. You name any pro and make them use my 1604VLZ and delta 1010. who do you think will make a better sounding record?
(hint - not me)
 
Specs are just specs and are totally manipulated by manufacturers.

M-Audio and Aardvark, as well as 90% of other soundcard manufacturers use the SAME converters, mostly AKM4524s. Some manufactures, do not overblow their specs others do. It should all be taken with a big grain of salt. Dont base your buying decisions on specs.
 
kristian -

Please re-read the post. I didn't RECORD anything via the Delta converters. I only played back a song that was already on the hard drive to master it. The problem that was showing is that even though the meters were showing more headroom, the SOUND on the D/A converters didn't have anymore. Switching over to a better D/A converter cleared this up!

If you don't think that the accumilation of headroom in your recorded tracks is NOT going to make a difference in the outcome, well........

This is problematic. If the D/A converters are crapping out before digital zero, that IS definately a problem. Use any perverse thinking you want to account for this, but it is still problematic on a number of fronts, the first being the age ol' questions that pops up on here at least once a week of "My recordings don't sound as loud as commercial recordings, why?". I don't see cheap A/D/A converters being any help here. Do you? ;) Can you see yourself listening to distortion while making critical decisions because you know that distortion will not show up when you play a CD actually burned from the source? No....you will turn it DOWN.

I don't suspect that the A/D converters on this card are going to perform all that much better, especially in light of the complaint another person on this bbs, who's ear I do trust, told me about them. I don't have that first hand knowledge, so at this point, I am just relaying what someone else said about the card used fully in it's intended use.

Another point. By the time you bought your Delta 1010 AND the compressors (which I tend to not favor while tracking if I can get away with....) you probably spent close you what two Lynx One cards would cost to only solve converters issues. You really didn't even solve the REAL issue here, that the converters are NOT what they claim to be, regardless of the specs, and now to try to get volume levels in the decent range, you HAVE to run a compressor in front of it. I don't get the thought process here at all!!!! I suspect that you would have foregone the extra compressors had you had a quality chipset in the first place.

ET, thanks for the detail of which chipset most manufactures use. This is a fact! The Lynx uses Cirrus chipsets! Whatever they are using, this card has been a tremendous investment for me. $400 for the capabilities and quality it offers is damn cheap in my book! It has allowed me to do far less quessing and compromising.

Ed
 
Another issue is just generalized specsmanship. My Audiophile 2496 is spec'd as having a D/A dynamic range of 104dB(A). This pretty well implies that the noise floor should be at -104dBFS. The converters used are all from Crystal, and really ought to be able to do those numbers very easily. However, in my real-world system, the *actual* noise floor is more like -80dBFS. The noise floor is entirely dominated by EMI hash resulting from living inside the computer's box, and there is a lot of it, and you can't easily ignore it. It also masks the bottom maybe 8 bits of my data.

I have no doubt that the 2496 design could achieve a usable dynamic range of 104dB(A): on an extender card, or more likely with the host computer halted (clocks stopped), just driving its own SPDIF outputs. Or maybe it was just installed in a _really quiet_ computer, eh?

The same converter devices from Crystal are used by the same company in a different product, the 24-bit Flying Cow standalone converter box. I have one on order, after a lot of study. Here's one independent review of the thing:

www.midiman.de/files/cowrev.pdf

I'll bet a beer that the A/D, D/A circuitry in the Cow is _identical_ with what's on the 2496. However, the EMI shielding and isolation provided by being in a separate box is critical to actually achieving the promise in the real world.

When you're getting down into the -100dBFS range, you are going to have a damned hard time getting world-class performance to happen inside a computer case. In a breakout box, maybe, *IF* the converters and their power supply filtering and regulation are all in there too- and if the grounding is handled very well indeed.

If some manufacturer casually specs that they can do a noise floor at -113dBFS with a card in a PCI bus cage next to forty-leven SCSI disks, a frigabyte of memory, and a Pentium 9 overclocked to somewhere near daylight: feel free to laugh at them. Doing wide-dynamic-range analog circuit design is _hard_, and doing it in the hellish EMI environment of a computer case is nigh-unto-impossible.
 
how much are 2 lynx one cards then? i thik 2 of those are stilll only 4 channels. i have 8 channels. still isnt competing. if you can afford to pay $1600 for 8 channels you can. but where i pay half of that and use peak limiting to help get hotter signals without the worry of clipping, the delta 1010 is great.

i have heard better converters. are they better? yes, can i afford them, no. another thing. i havent once had the problem of "why isnt it loud". i imagine the people asking that are the people doing it wrong.
 
I have to agree with Kristian....follow the formula of performance divides by cost and the Delta wins...I needed a card that had stereo ins/outs only and S/PDIF and Midi in/out.....Lynx One and Delta Audiophile were almost identical in features...True, the Lynx One is a step up in quality...Ive even helped talk RuskyK into buying one based on HIS needs....but for MY needs, I needed to pay $149, not over $400.....BUT, you wont here me crying later about why my recordings arent this and why they arent that etc. etc......
 
delta 1010 converters. All that and more

Hey,

has anyone seen the new aadvark q10. It looks like a solid beast.
Man, that unit is spec'd out.

Ok. Back to the delta 1010. Nobody will argue that they are the best value for money on the market right now.

As far as converters go, I did a lot of research before getting the delta 1010, and stuff like "clear as a bell", "what goes in comes out", "best soundcard I have ever used", "no ick layer" etc kept coming up.

Before I go on, most soundcards do use AKM converters, but I believe that the delta 1010 uses AK5383 A-d and 4393 D-a.
on the box, m-audio claims to print real world results of the whole unit, not the converters, cos the AK4393 converters claim to have a noise floor of -120 dbs.

I read this thread last night, but could not reply cos I was at a friends, and I wanted to do some tests.

the test was a 1000hz sine wav a various decibels in soundforge.
The meters in soundforge and in the delta 1010 control panel were identical at every mark, even to -0.5 dbs. Even at unity, there was no distortion on the sine wav.

I think the problem sonusman was having was the drivers for the delta 1010 he was using. M-audio has pulled drivers off their webpage more than once and gone back to using old drivers. I had to switch from version 29 back to version 27 because my asio was freaking up, and m-audio had pulled 29 from the board. They recently did that with version 31 as well. As far as I know, v27 is still the most stable.

That settled, I proceeded to testing the a-d , d-a converters themselves.

I am running from the delta 1010 through 10 awg cable into an Alesis ra100 amp, into Alesis monitor twos. Not professional, but quite accurate.

I ran my jv1010(sound module) directly into the Amp, and compared it with running through the delta 1010. I must say that I percieved a difference while playing the demo. I could hear all details clearly in both cases(snares hihats kicks etc) but the jv1010 seemed a phantom bit colored and a litlle less loud.

So I say that the delta 1010 is all that and more. Why?
considering that it sounded practically the same even though I was not using full 24 bits( The jv1010 demos peak at about -12 dbs).

Also considering that I am using $3 cable to the ad and from the d/a,.

I bet that I wont hear a single difference if I was going through canare cable or moster, and I had a source that peaked closer to zero.


I read about an a-b comparison between the delta 66 and a lynx one. In my opinion, these are two fantastic cards, but the reviewer thought that the lynxone sounded SLIGHTLY better.
I am sure the delta 1010 converters are better than the lynx converters simply because they are better than the delta66 and are housed in a breakout box.

. Sonusman, you said that the audio was not distorting by a "long shot". was the output set at the right level(settable at the back of the box) cos if they are wrong, you will have a signal that is approx 11 decibels hotter going into the amp.That will definitely produce that distortion.

One last thing. I think the delta 1010 is a professional soundcard.
Everything better is a dedicated converter and not a soundcard.

All this is not to say that I won't get the lucid when I can afford it though :)

peace everyone.
 
o crap!

"In Use

Even on initial auditioning with 16-bit playback, the converters of the Delta 1010 sounded noticeably sweeter and more focused at the top end than those of my 20-bit Gina card. Measuring RMS background noise levels at 44.1kHz and 48kHz using Wavelab gave values of –93.4dB at 16-bit, and –109dB at 24-bit. Increasing the sample rate to 96kHz didn't change noise levels at 16-bit, but noise levels at 24-bit increased to about –102dB. This isn't so surprising when you consider that the noise bandwidth has doubled! I should point out that both Wavelab and Sound Forge use the convention that 0dB RMS is the largest possible square wave signal, so a sine wave with a peak level of 0dB will actually measure –3dB RMS. Because of this, the actual signal-to-noise ratios (measured relative to a sine wave at full level) are 3dB smaller than the figures I have quoted. "

I just saw this on the net. I should have used a square wave not a sine wav. Does this mean that a square wave at zero will be +3 to the delta 1010? I have to go back and test this.
 
CJ - You can be assured that I would have heard a level capatibility issue and checked that out. The Lynx One simply has far more headroom.

To all - Have your Delta 1010's. If it works for you, great. I have heard all I need of it to base it's use for my needs, and it didn't cut it. If it does for you, great. But, don't let "spec's" guide that purchase. Spec's lie more times then not.

Ed
 
i dont use specs.

i dont compress to disk. i just limit peaks.

but when an completely uncompressed stereo render in vegas cannot be limited a single dB without sounding bad, im believing im getting great levels out of my card.

" I have heard all I need of it to base it's use for my needs, and it didn't cut it. " - you only used the D/As.

you're right most of the time, but not always. sorry.
 
Im not even gonna attempt to say the Deltas converters are better than the Lynx One's converters.....
 
sonusman said:


I did a little mastering job using the Delta 1010 converters for the D/A. I was sorely disappointed because when I took the .wav files home and listened to them on my Lynx One card, I found out that the audio was fully -1dB down from what it could have been, even though it was peaking on the Delta converters!!!

To explain further.

While mastering, I was applying the Waves L1 Ultramaximizer to the mix. While I reviewed the metering, I was noticing that the sound started to distort a bit lower then I was accustomed to, but I was HEARING distortion, and regardless of what I seen, I have to go with my ears right? Of course.

I notice on playing the CD in my truck that the audio is not distorting by a long shot, and I compared the product to stuff I mastered using my Lynx Card as D/A. Guess what, the stuff mastered using the Delta converters was noticeably quiter.

I get home and transfer the .wav files onto my computer and take a listen. FAR from distorting! In fact, I did a Peak Search using Wavelab and found that I NEVER hit digital 0 once in the mix!!! I look at the metering on the Lynx One meters and it shows that I am peaking at around -1dB at best!

My oh my........

I proceed to load up the unmastered .wavs and apply the plug in's I used again, this time using the Lynx One card as the D/A. I couldn't get EXACT settings because I saved those on the other computer and didn't think to copy them to the DATA disk. Anyway, I was still pretty fresh from the session, and was able to replicate the eq curves. Where the big difference existed was in the overall level I was able to achieve before distortion ensued on the Lynx One card.

The resulting .wav file was much hotter and sounded much more like what I was expecting it too.




Ed
Click to expand...


Hey Ed, I'm just a humble guitarist turned live engineer so the fine points of digital recording aren't my strong suit, so can you clarify something for me:
If the 1010's DA converters are distorting before digital zero doesn't that mean that any modern commercial CD played back through the 1010 should distort the converters?

Also you said that you used the Waves L1 Ultramaximizer, was it set to Analog or Digital? I think that if it is set to Digital that it will pass peaks higher than the specified max.

I am not disputing that the Lynx's converters are better and have more headroom, I just find it unbelievable that a card that distorts before zero would be released commercially.



Just looked up the Waves manual, this is what I was talking about;

"Digital/Analog Domain

In almost all cases, leave this button in the Analog Domain position.
When in the Digital Domain position, absolutely no sample will be over the Out Ceiling value. However, after analog conversion, it is possible to have peaks higher than in the digital domain. This is due to very complex digital audio issues involving peaks "between the samples". Almost all quality-made digital-to-analog converters have at least 3 dB headroom to allow for these peaks.

You would want to use the Analog Domain position is when you wish to have absolute control over any peak that occurs in both the analog and digital domains. Some examples: you wish to accommodate poorly designed DACs; the file will undergo further modification, such as ADPCM data-reduction; or, you wish to have a signal that can be broadcast without further peak controlling. In these cases, brick wall limiting is desired in both domains, and you should put L1 in the Analog Domain mode."
 
Last edited:
You must log in or register to reply here.
Back
Top