Become a Site Supporter and Never see Ads again!

Author Topic: True comparison/analysis of A>D Converters  (Read 4141 times)

0 Members and 1 Guest are viewing this topic.

Offline heath

  • Laugh it up, Fuzzball...
  • Trade Count: (4)
  • Needs to get out more...
  • *****
  • Posts: 24817
  • Gender: Male
  • I'm score!!!!!!
    • The Upstream Mend
True comparison/analysis of A>D Converters
« on: June 21, 2007, 02:45:32 PM »
SO I'm thinking about actual analysis of what an A>D does to an incoming signal.  There has to be a qualitative method of measuring/examing how a given converter "colors" a signal.  Originally, I thought I could use spectral analysis/transfer function in Spectrafoo.   I would split the output of my Studer and run analog in to the computer (for spectrafoo analysis) and also run into my converter which is also feeding spectrafoo.  Great, I can use the transfer function to see the difference between the 2 signals.  But no, I completely spaced on the fact that my "analog in" to the computer for analysis is in no way an analog in which gets analyzed on an analog level.  It is digitized by the computer/card/DAW and that has an impact on my analysis thus making my results inaccurate. 

I want to find a method of "viewing" what an A>D actually does to the signal.  I can look at specs all I want (freq response, THD, Dynamic Range, etc) but specs don't mean anything...I want some concrete evidence that a "transparent" A>D really is transparent.  Not sure if it's even possible.  The only device I've found that might help me out is an Audio Precision 2722 which claims it can simultaneously analyze in both the analog and digital realms.  Only drawback is that the box is approx $28,000  ;) 

Any tech heads have ideas on how to truly test an A>D?  Listening tests are subjective, and thus are out of the question...

h
And the Sultans... yeah the Sultans play creole

 The Upstream Mend

Offline Church-Audio

  • Trade Count: (44)
  • Needs to get out more...
  • *****
  • Posts: 7571
  • Gender: Male
Re: True comparison/analysis of A>D Converters
« Reply #1 on: June 21, 2007, 02:50:06 PM »
SO I'm thinking about actual analysis of what an A>D does to an incoming signal.  There has to be a qualitative method of measuring/examing how a given converter "colors" a signal.  Originally, I thought I could use spectral analysis/transfer function in Spectrafoo.   I would split the output of my Studer and run analog in to the computer (for spectrafoo analysis) and also run into my converter which is also feeding spectrafoo.  Great, I can use the transfer function to see the difference between the 2 signals.  But no, I completely spaced on the fact that my "analog in" to the computer for analysis is in no way an analog in which gets analyzed on an analog level.  It is digitized by the computer/card/DAW and that has an impact on my analysis thus making my results inaccurate. 

I want to find a method of "viewing" what an A>D actually does to the signal.  I can look at specs all I want (freq response, THD, Dynamic Range, etc) but specs don't mean anything...I want some concrete evidence that a "transparent" A>D really is transparent.  Not sure if it's even possible.  The only device I've found that might help me out is an Audio Precision 2722 which claims it can simultaneously analyze in both the analog and digital realms.  Only drawback is that the box is approx $28,000  ;) 

Any tech heads have ideas on how to truly test an A>D?  Listening tests are subjective, and thus are out of the question...

h

Tests dont show everything.. I use winaudiomls.. But even that does not really tell you the differences in digital vrs analog.. I think in the end your ears are the best judge. With gear its more then just the deviation of frequency response that effects sound.. I would say if you really want to test something do a blind comparison... with the same input source and try it on a bunch of converters until you like one.
for warranty returns email me at
EMAIL Sales@church-audio.com

Offline heath

  • Laugh it up, Fuzzball...
  • Trade Count: (4)
  • Needs to get out more...
  • *****
  • Posts: 24817
  • Gender: Male
  • I'm score!!!!!!
    • The Upstream Mend
Re: True comparison/analysis of A>D Converters
« Reply #2 on: June 21, 2007, 02:54:50 PM »
I completely agree with you, Chris.  However, I am trying to come up with a method of showing a potential client (typically a library or archive) that our converters are transparent, as opposed to simply saying to them that "we spent $10k on converters and they are transparent, so trust me".  I want some way of assuring them that their "priceless" asset is accurately represented in a digital file.  I don't know if it's even possible. 

I think I'm rambling.  I've been tossing this issue around in my head for a few days, so I'm kind of out there right now....
And the Sultans... yeah the Sultans play creole

 The Upstream Mend

Offline Church-Audio

  • Trade Count: (44)
  • Needs to get out more...
  • *****
  • Posts: 7571
  • Gender: Male
Re: True comparison/analysis of A>D Converters
« Reply #3 on: June 21, 2007, 02:59:50 PM »
I completely agree with you, Chris.  However, I am trying to come up with a method of showing a potential client (typically a library or archive) that our converters are transparent, as opposed to simply saying to them that "we spent $10k on converters and they are transparent, so trust me".  I want some way of assuring them that their "priceless" asset is accurately represented in a digital file.  I don't know if it's even possible. 

I think I'm rambling.  I've been tossing this issue around in my head for a few days, so I'm kind of out there right now....


FFT from 0hz to 100khz using winaudio MLS lab version with a very high end card the name escapes me but the sampling rate is 100mhz.. and the bandwidth is 0hz to 400khz.. flat.... You can measure distortion with this you can also measure frequency response and phase response and impulse response the whole rig will cost about $2k or you can rent an AP rig there are a few places that rent Audio Precision gear do a Google you can find them pretty easy.. But both systems have a learning curve. There are many other types of systems out there but this is the cheapest method I know of. The AP will probably rent for about $500 a day + a left nut for a deposit.

Chris
for warranty returns email me at
EMAIL Sales@church-audio.com

Roving Sign

  • Guest
  • Trade Count: (0)
Re: True comparison/analysis of A>D Converters
« Reply #4 on: June 21, 2007, 03:00:57 PM »
I completely agree with you, Chris.  However, I am trying to come up with a method of showing a potential client (typically a library or archive) that our converters are transparent, as opposed to simply saying to them that "we spent $10k on converters and they are transparent, so trust me".  I want some way of assuring them that their "priceless" asset is accurately represented in a digital file.  I don't know if it's even possible. 

I think I'm rambling.  I've been tossing this issue around in my head for a few days, so I'm kind of out there right now....

How much of an understanding of the technology does the client have? Have they requested a certain spec?

Offline heath

  • Laugh it up, Fuzzball...
  • Trade Count: (4)
  • Needs to get out more...
  • *****
  • Posts: 24817
  • Gender: Male
  • I'm score!!!!!!
    • The Upstream Mend
Re: True comparison/analysis of A>D Converters
« Reply #5 on: June 21, 2007, 03:09:17 PM »
clients have little understanding (typically) of the digitization process.  They just want some sort of assurance that what we are doing is as accurate as possible...

I have spoken with a rep for AP, and I might be able to get a demo unit flown out here to play with before I decide if the lab needs to buy one ;)

And the Sultans... yeah the Sultans play creole

 The Upstream Mend

cshepherd

  • Guest
  • Trade Count: (0)
Re: True comparison/analysis of A>D Converters
« Reply #6 on: June 21, 2007, 07:56:25 PM »
Would published industry reviews of a/d converters have this sort of information?  Consistently favorable reviews on the a/d converters might make a bigger impression on your clients than a technical readout regarding the accuracy of a digital conversion.  It's just 1's and 0's anyway.  What's the big deal?    ;D

Chris

Offline heath

  • Laugh it up, Fuzzball...
  • Trade Count: (4)
  • Needs to get out more...
  • *****
  • Posts: 24817
  • Gender: Male
  • I'm score!!!!!!
    • The Upstream Mend
Re: True comparison/analysis of A>D Converters
« Reply #7 on: June 22, 2007, 08:36:06 AM »
Some say 24 bit PCM is junk so maybe you should sell that junk gear and buy a Korg to do DSD :P


the converters i am acquiring all do dsd (one of the does dxd!!!!).  SO should the need arise, I have no issues with moving away from pcm ;)
And the Sultans... yeah the Sultans play creole

 The Upstream Mend

Offline DSatz

  • Site Supporter
  • Trade Count: (35)
  • Needs to get out more...
  • *
  • Posts: 3347
  • Gender: Male
Re: True comparison/analysis of A>D Converters
« Reply #8 on: June 22, 2007, 11:41:47 PM »
Some things you'd probably want to test about an A/D converter if you could:

[1] Frequency response
[2] Distortion and noise
[3] Linearity at different signal levels (from just below full scale to just above the noise floor)
[4] Clock accuracy and stability
[5] Accuracy of level metering and/or overload indicators

With certain software one can feed analog test signals into the converter and analyze the digital output directly in order to measure these things; it isn't necessary to convert the output back to analog (which in effect would be testing two converters at once; if you found troublesome non-linearities, you wouldn't know right away whether they were in the A/D or the D/A).

The Audio Precision test sets can run a suite of tests like this on A/D converters; unfortunately they cost around $5000 for the simplest model ...

--best regards
music > microphones > a recorder of some sort

Offline SmokinJoe

  • Trade Count: (63)
  • Needs to get out more...
  • *****
  • Posts: 4210
  • Gender: Male
  • "75 and sunny"... life is so much simpler.
    • uploads to archive.org
Re: True comparison/analysis of A>D Converters
« Reply #9 on: June 24, 2007, 12:10:26 AM »
I'm not sure what the business relationship is here... but if you are talking about a library or "non profit" archive, emphasize that and people may start to get cooperative.  I can see involving the EE department of you local University.  Sounds like a good lab project for them, maybe even a graduate thesis.  They probably have some sort of precise signal generator, or if not, find your local rep for HP/Agilent, etc and see if you can loan you/the college a piece of demo hardware for a few days.  I'm kind of out the loop on these modern test instruments, but I expect you can program them to put out some specific signal like 1V at 10hz, and it will do exactly that, and then whip through some sort of programmed cycle to 100khz.  "Sound" is just a signal with a wavelength in the audible range.  Standard engineering practices apply.

Once you do this work for a non-profit, and let them benefit by making an appropriate buying decision (they buy the $100 one), it's legitimate to take what you have learned and present that data in a profitable situation (you convince them to buy the $100k one).
Mics: Schoeps MK4 & CMC5's / Gefell M200's & M210's / ADK-TL / DPA4061's
Pres: V3 / ST9100
Decks: Oade Concert Mod R4Pro / R09 / R05
Photo: Nikon D700's, 2.8 Zooms, and Zeiss primes
Playback: Raspberry Pi > Modi2 Uber > Magni2 > HD650

Offline heath

  • Laugh it up, Fuzzball...
  • Trade Count: (4)
  • Needs to get out more...
  • *****
  • Posts: 24817
  • Gender: Male
  • I'm score!!!!!!
    • The Upstream Mend
Re: True comparison/analysis of A>D Converters
« Reply #10 on: June 26, 2007, 09:24:15 AM »
hmmm...interesting idea.  I hadn't even thought about asking local universities :)  would be right up Carnegie Mellon's alley ;)
And the Sultans... yeah the Sultans play creole

 The Upstream Mend

 

RSS | Mobile
Page created in 0.115 seconds with 40 queries.
© 2002-2024 Taperssection.com
Powered by SMF