JPF Home Page Forums Musician Gear and Techniques Recording Studio and Gear Forum Tech Query about analog to digital conversion

Hi All,

This question is for the technical experts.......

Is there a difference in analog to digital conversion between brands or types of interfaces, or are they all doing exactly the same thing when they convert an analog signal from a mic or instrument into a digital signal for recording?

Thank you!

Colin

While I'm not savvy enough to go into it in any depth, there is a difference in the algorhithms used by equipment/interface manufacturers to convert analog to digital and a coinciding difference in conversion quality.

I suspect the device is much the same for all brands. There may be older converters that are still used but you can't hear the difference unless something has failed. Remember Analog when converting to digital for making a CD for instance has to be converted back to analog for you to hear it.

Hi Colin,

I'm not a "techie," but I will tell you that there is a big difference between brands and models of A/D converters.

To begin with, there's a lot of room for difference on the analog end of things. The brand and grade of input, the wiring and the soldering, whether or not there is a preamp, the use of phantom power...all these things make a huge difference in the analog sound as it is going from the microphone to the actual converter. Then, on the digital end, a lot depends on the sampling rate and whether that's the rate your recording program uses. Some programs work in 48k and benefit from dithering before coming back to 16k for recording to standard audio cds. If you are working in 48k, then you'll want to stay in 48k from the analog stage through the mastering stage before converting back. One of the techies can explain why better than I can, but put simply, each step that changes rates removes some of the definition. It's better to start out in hi def, then move to the lower def with a dither.

Apogee is known as a high-end A/D converter. I use the converters on my EMU box, they are the same that are used by ProTools. I also have kept two of my old ADAT machines in case I ever need to record an extra 16 tracks in one session (I haven't yet, but since their value has dropped, it's worth it to me to keep them "just in case.") Of course, the ADATs are older models, and they will stick me to 44.1k. I'll have to dither.

Hope this helps until one of the techies comes along.

Mike

Thanks for the responses.

Mike, which EMU do you use? I like the looks of the small USB models but some of the reviews are not too good - but I don't know how much that means. I have not found any brand that has consistently good reviews in the less expensive models. Probably a lot of operator problems involved.

I am using an external sound card from Creative Labs and a mixing board with pre-amps. I get decent sound, but I wonder if it could be better without spending a fortune.

Colin

Being one of the resident tech weenies, I'll throw my $0.02 in...

The short answer is "YES". There is a difference in converters. Most people will hear the difference between a $99 stereo a/d converter and the $7000-$9000 Prism or Weiss stereo converters. The $2000-3000 converters such as those by Apogee, Lynx, etc are also considered very high quality. Typical comments on quality converters are that they sound more "lifelike", "bigger", "deeper", "more spacious", "more natural", etc.

I did an informal test a few years ago between my Digi001 (fairly low quality converter) and an Apogee Rosetta (fairly high quality). I simply recorded my acoustic guitar with both using the same mic and mic preamp. Subjectively, the Digi001 produced a very clean recording of the guitar. However, the Apogee recording sounded less like a cleanly RECORDED guitar, and more like a real guitar in the room. It simply sounded more real.

For the explanation below, I want to exclude mic preamps from the equation as they add a whole OTHER layer of quality variables. Let's just break it down to the process of converting a line-level signal to a digital bitstream.

There are several factors that play a role:

1. The quality of the analog components and design (it DOES start with an analog signal before sending it to the digital converter chip). This has to do with any active (analog) component quality such as resistors and capacitors, wiring quality, shielding, power supply, etc. This also includes the filter, which is especially important in reducing the frequency response to 20K hz (the reason for this frequency response reduction is a whole other theoretical explanation). Cheap filters will color the sound and produce aliasing effects, which are non-harmonic "undertones" and originally gave digital audio a bad name in that they make the sound "hard" (for lack of a better term). Cheaper components essentially color the sound and will be a bit noisier.

2. The quality of the converter chip itself. Many brands of a/d converters now use the same Analog Devices AD1971 chip whether they are MOTU, M-Audio, or several other brands. This is much less a differentiator than several years ago when the actual converter chips varied in quality more. As a general rule, if a converter chip can convert to 192K, it will be of higher quality than a chip that is only able to convert to 96k - all other factors being the same.

3. Clock - all digital devices need a timing source so that the bits all line up nicely once the analog signal has been converted. This is probably a bigger differentiator between the "openness" of various brands of converters more so than the type of converter chip used. The steadier the clock, the more transparent and "bigger" the sound becomes. You may have heard of the term "jitter". Jitter results when the clock is less than perfect (which it ALWAYS is). The bits arrive at the destination just a little "off-time" from a theoretical perfect placement in time. These fluctuations are called jitter. It's probably easier to understand what jitter will do to audio if you can imagine a computer monitor having tiny little "shakes" in the video stream which end up making the letters harder to read. If you reduced the jitter, the text on the screen would be clearer and have more contrast, and photos would have more realism and depth. Jitter effects audio in a similar way.

4. Dithering - this was probably more a factor several years ago than today since the only time this enters into the equation is when reducing the bit depth from 24 to 16 bits. Dither is simply noise added to help the bit reduction sound a little smoother on very quiet audio such as reverb tails and other very low level signals. Since almost everyone records at 24-bits these days, dither only enters in when either mastering, or when converting your mix to a format for CD (which must use only 16-bit audio).

In summary, the quality of the analog components (including the filter) and the steadiness of the timing source (clock) probably influence the sound quality of a converter more so than the type or brand of converter chip.

There are always other factors that can influence the sound such as whether the device has any built-in limiting - of which there can be huge quality differences, whether you're using a built-in mic preamp (a whole new set of potential color), and the sampling rate (44.1k, 48k, 96k, etc), all can influence the limitations and the overall sound quality of the device.

Hopefully, that will clear things up a little. If anyone has other questions, I can get down to the bit level if needed and can reference several white papers on the subject.

Hi Larry,

That is exactly the explanation I was looking for. Thank you.

I am considering upgrading from an older Creative Labs Soundcard to the EMU 0404 which I think is made by the same company. It seems to score well on the criteria you mention.

I appreciate the help.

Colin

Hi Colin there are better more expensive interfaces out there BUT the EMU 404 is pretty good for the bucks. You will not go far wrong if you get it.

I like Larry's explanation a lot.
It's funny, but many of the same component chips are used in the less-expensive converters as in the more expensive boutique stuff.
Where the difference lies is in the design, construction, and layout of the circuit boards, the quality and ability to pump out current of the power supplies, the quality of the other electronic parts used, the analog input/output stages, and on and on.
"Get the best you can afford" is the best advice I think.

Learn somethin new everyday!!!

I decided to try a PCI board style AD converter and ordered an Audiophile 2496.

i will report back whether it makes a noticeable difference when I receive it.

FYI: PCI cards generally have the lowest latency and are often the most reliable (no cabled connections to go intermittent). Firewire is a little higher latency (but can get pretty low). USB interfaces have the highest latency.

Thanks Larry very informative. If anyone uses Cubase SX here is a link giving some tutorial info and tips on how to overcome latency probs plus tips and hints on sorting other probs.

http://www.thewhippinpost.co.uk/cubase/cubase-sx-tutorial.htm