This is the test.

I took a single track out of an old Nuendo project via lightpipe at 44.1/24 and ran it D/A out of a Creamware A16 Ultra converter, then back into the A/D on the same unit.

I pumped the track through once, making track 1.

Then I sent that track back through the same signal path, and continued sending the last converted file back through until I had 20 tracks.

Track #20, if you do the math, is 210 instances of A/D and D/A conversion.

The files had to be normalized, as each conversion lost about maybe .2dB of volume. Normalizing was necessary so that differences in volume wouldn't be perceived as "better". But, normalizing any way you cut it isn't going to clean up these tracks or change the noise floor etc.

I then converted the .wav files to 320k .mp3 which I admit ain't ideal but it's the best .mp3 has to offer. The converter was www.dbpoweramp.com.

http://www.nowhereradio.com/artists/album....=2079&alid=1201

I plucked tracks 1, 5, 12 and 20 and posted them up.

Would you use track #20 in a mix?

This is NOT a scientific test, but in the world of tracking and mixing mainly pop and rock music you may consider it good enough to get an idea of what I'm talking about. In my studio, through Lucid D/A monitoring and Mackie HR824's I am not hearing a crazy difference. I think 210 instances of conversion is a few more than most folks would put them through even mixing in analog.

The Lucid stuff to my ears sounds more open, and the headroom is excellent and so is the metering. The Creamware doesn't have useable metering and isn't quite where the Lucid is sonically. I'm really kind of throwing this up here for the sake of argument, how picky can we get about converters?

War

I seem to recall at one time guys doing similar sorts of tests with mini disc systems.

I guess what I'm saying is, yes there's a difference but when I see guys who have 1 mic, 1 preamp and now they want to spend $2000 on conversion it makes my skin crawl. Wrong place to sink the money at that point in the rig.

These Creamware converters are probably leaps and bounds better than anything that was available through most of the 90's. Yet think back to some nice sounding music you heard recorded from the time, it's there and it didn't stop a good album from being made.

War

210??? isn't it 21???

Cool test. I thought that the 20th track would have been OK in a mix with the exception of it being the featured or solo instrument. There was only minor degradation to my ears.

You probably got saved by using 24 bits instead of 16 in this particular test. There is always some noise/loss in a conversion but in a well designed converter it is less than 1/2 of an LSB. This noise would be additive and doing 21 conversions could lose you as much as 10 bits! This would make the converter basically sound like a 14 bit converter. At 14 bits you still have very good signal to noise, ~84dB. Even 13 bits would give you 78dB. I think anything better than 72dB is usable for most applications.

Well the way I did it was I kept converting the prior track (just converted again) so you have to add in all the conversions before it. So the math is:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20 +
---
=210

For instance, if I would have done this 3 times (more than most folks would ever A/D and D/A a track) you would have:

1
2
3 +
---
=6

War

What I'd like to hear would be this same experiment on analog tape, high output or low output. You'd have one hell of a mess at the end of the "conversions". Can you say noise floor??

War

Which Lucid do you have? I’m currently running an Aardvark (now defunct) Q10 and am considering stepping up to something like a RME Multiface (should have got that to begin with ) if I decide that the difference is significant enough.

Flatpicker, we just became a Lucid dealer. I have been using the AD9624 and the DA9624 for a little while now, along with their clock distro and the quality pushed me into contacting them about selling their product.

We're a Creamware dealer also (listed on their site but we don't have it on ours yet!) but this was the first time I was able to use their converters.

The Multiface is a decent unit anyhow. I guess you'd need to look at budget vs. i/o requirements. The Multiface also has other digital connectivity and who knows if you need all that.

I really do like the Lucid stuff though, the quality is there. But what I'm doing is using the A16 for drums etc. when I need a lot of conversion and my Lucid 2 channels of A/D D/A for everything else. It's always a good idea in a studio to have 2 killer channels from the mic, the cables, the preamp, the cables and the converter!

War

I don't think that's right.

Let's say the original track is A (A/D converted during recording)

you convert A and make A' (D/A -> A/D) so A' has been converted 1.5 times
(assuming once all the around is 1)

then take A' and make A" (another D/A->A/D)

now you have 2.5 loops, or 5 total conversions, not 6.

if you go again you get A'" ... 3.5 loops, 7 total conversions, not 10.

I don't think it's valid to keep re-counting previous conversions.

Hmm, well I don't see how simply adding those up won't do it. You're still hearing conversions in addition every time, you must add them all together.

Either way this file has been put through hell and back.

War