Computer Interface vs. Standalone CD-R - I'm Perplexed (Archiving Analog Audio)

OK. So I finally got my whole setup going to archive the reel to reels. I'm using a Technics 1500, which is considered one of the better R2R decks made. The goal was to archive at 192/24 or 96/24, and then burn to DVD-A (when the new Bronze program comes out), or if I wanted to, downsample to 44/16 for CD if I wanted.

I have a Pentium IV 2.25 GHz, and I have a dedicated Seagate Barracuda drive (200 GB) for this project. I'm using XP and for my audio software, WaveLab 4.0.

I just purchased the E-MU 1820 as the audio interface. It's an external breakout box attached to a PCI card. It did 192/24, 96/24, and everything else. Considering the number of ins/outs, it was a good bargain. For the same money is the M-Audio Firewire 410 (I think that's the model) which is only 96/24.

Once I figured out the ins and outs of using the 1820 with WaveLab, I thought I was set. No. I then found I needed to set my ASIO latency to like 40 ms (a good discussion on latency can be found here http://soundwave.com/Htm/Articles/April/Audio_Latency.htm). Much lower, and I got pops and clicks no matter what recording resolution. Which is weird, becuase my I/O path is ATA/100, I have 1 GB of memory, and have the fast processor. So that's problem #1. 5 ms or less of latency is considered acceptable in most cases.

I overlooked latency and set it at 40 ms to get stuff recorded. That seemed to work fine. I didn't notice any audio glitches, nor did I percieve anything bad to the sound. I didn't add any plugins or anything, so there was no real time or post-recording processing of the audio in the computer.

For grins and yucks, I decided to run the audio in parallel direct to my Marants CD-630, a pro CD-R machine. I recorded the same stretch of music as I did into the computer into WaveLab and at 96/24. Then I dithered down the 96/24 to 44/16 and burned it

What was the result? The Marantz beat the pants off of the dithered 96/24. It sounded more like the reel. The one from the computer (burned at 2x BTW) had much less low end and sounded like it had a veil in front of it.

Comparing the waves in Wavelab (I imported the Marantz track), there are slight dB differences, but nothing drastic. And what I'm hearing does not seem to be loudness related at all. The standalone CD-R had a more pleasing sound to it is the bottom line. The computer track sounded kinda lifeless.

I am stumped.

Here are my theories of
1. Is it related to whatever is not allowing me 5 ms latency? (i.e. less latency, less quality since there are less samples) Since the 1820 supports 96 and 192, you'd think latency wouldn't be an issue at all.
2. Is it that the A/D converter in the E-MU is not as good as the one in the Marantz, even at 96 or 192?
3. A combo of #1 and #2?
4. Is it that the E-MU sucks? (LOL)
5. Is it something in the dithereing process to 44/16?
6. None of the above?

Any help would be appreciated.

Allan

 

Those would be my two guesses...and the downsampling/dithering could very well be the culprit. Do you have any way to compare your CD to the 96/24 directly?

 

Not easily. The way the ASIO driver works in WL, you need to change it every time you change resolution. I was going to try that, but if it still doesn't sound as good, I'll be disappointed that even at 96/24 I can't get as good a sound as 44/16. Then again, it is all in the A/D converters, but still. For $400, I was expecting to have better quality.

What I'm trying to avoid is:
1. Having to buy something like the MOTU 828, which is good but expensive.
2. Just doing 44/16 for obvious reasons.

 

StyxCollector asked:

" ...is it related to whatever is not allowing me 5 ms latency? (i.e. less latency, less quality since there are less samples)"

I don't believe this is correct. Latency has NOTHING to do with the SOUND QUALITY of a recording, and should not really be a factor in this case...

Latency is only a measure of the TIME it takes for an analog signal to be converted to digital, recorded to the hard drive, and then re-converted to analog so that you can monitor the recording.

Somewhat similar to a good-quality 3-head open-reel or cassette deck, in that you will hear a delay between recording and playback, depending on which head you are monitoring from.

In the case of the analog machine, the delay is related to the tape speed and the actual physical space between the recording and playback heads...

While in the case of the computer the delay (latency) is related to the hard-drive speed (at least 7200 rpm for audio recording) as well as the processor speed, amount of RAM available, etc, etc. ("Damn it, Jim, I'm an audio engineer, not a computer scientist.")

As I said, latency should not be a factor in this case, unless you need to "overdub," or record some tracks "in sync" between the reel and the computer. Otherwise, just monitor your recorded signal from the computer.

hth,

 

A betterv way to determine this is to compare the Wavelab 96k/42-bit to the Marantz CD-R. That will help you determine if it is the E-MU. If you can do that, and they wind up sounding the same, then your problem is the dithering.

Not all dithering is created equal. Perhaps you should try another program for dithering. I personally use Adobe Audition/Cool Edit for this.

 

Any elaboration would be crapping the thread so I'll do as SC and quote Pirsig again:

"If the machine produces tranquillity, it's right."

 

I would agree that the dithering algorithm used will impact what happens.

What I wound up doing is A/Bing the Marantz vs. the E-MU by ripping the Marantz track into WaveLab. There was a difference, but what I then did was alter the buffers and the latency, and I got a sound that was close to, if not equal, the Marantz sound.

My problem now is that I still get the occasional pop and crack at a lower latency. Which is annoying, so it's like going back to square one.

BTW, I posted over at the Steinberg/Cubase forums, and someone had an interesting reply which I thought I'd cross post here (and ignore the DVD-A comment ... this is not a SACD vs. DVD-A thread; it's just part of his original post)

So based on this, if I do 48/24 or 44.1/24 and not 96/24 or 192/24, and then do a parallel to the Marantz, I'll have good CD master as well as one for DVD-A methinks ... (unless Sony wants to lend me a Sonoma LOL)

Thoughts? Comments?

BTW, I will call E-MU support tomorrow and see if I can't straighten this out - they were closed on the weekend.

Allan

 

Hi Allen,

I have dubbed many of my Reel tape masters directly to CD-R using a HHB Pro CD-Recorder in the analog mode, of course. I have been quite satisfied with the results. I have no experience with any attempts at hi-rez transfers, however...

Good luck!

Bob

 

I think I decided to just use the Marantz to do the basic stuff and assemble it in WaveLab later. I am still going to try to work out my 96/24 192/24 issues, but I can do that stuff at a later date.

 

Or, why not just record the tapes directly into Wavelab at 16-bit/44.1k if you plan on doing nothing more than a bit of editing and assembly? That way you don't have any dither or sample rate conversion issues, and you don't have to go an extra two steps.

I only recommend recording at higher bit-depths and sampling rates if you are either going to process the sound or archive to DVD.

 

You could even bypass your sound card entirely--record the tapes straight to CD on the Marantz, then rip those into the computer for editing at a later time. I used to do this when I needed a quick and dirty way to get music into my computer.

Or if you can spare a disc to experiment with, try recording through your computer setup at 44.1/16 and compare to a disc made on the Marantz. If they're close, then maybe it is your Wavelab's downsampling that is causing the difference.

 

That's what I am going to do (see post above). I usually do that because it's always easier than firing up the computer.

After I iron out the issues with the E-MU - I'm calling them in the AM. Until then, it's not a fair test.

 

If you read the thread, right now I'm having E-MU issues and am not only concerned about the quality of the A-D converters, but pops and clicks. Plus, to be dead honest, it's much easier to do it on the Marantz (A-D converter quality aside) and then just rip it into the computer. It saves many steps - much more time efficient than using the computer actually.

The goal was to burn them to DVD-A when Bronze is released later in the month, so I had either a 96/24 or 192/24 version. That was the original plan.

 

Yeah, I did read all of that earlier. I forgot, that's all.

Definitely call them up! Maybe the E-MU is the wrong card to use.

 

Update for you guys:
I'm still having some latency issues, but I got it down to 20 ms at 96/24, which is much more acceptable.

Two links for all you PC-based musicians that may help you:
http://www.musicxp.net/
http://www.tascamgiga.com/pdf/optimizing-xp-and-2k.pdf

These are excellent resoruces for anyone using a PC for music.

I made sure some of the tweaks were applied (as many as I could within reason AND were not done already), and I went to the 20 ms latency. In the process also made sure that DMA was enabled in BIOS and to make sure, and I disabled the soundcard built into the mother board in BIOS as well.

E-MU support so far has been ... umm ... shall we say less than helpful. I called them once and they said they tested the card with Cubase, but not WaveLab, and tried to say that the ASIO implementation in WL sucked. Yeah, right. I'm importing stereo, not doing multi-tracking. Using Cubase (which I do have) would be like trying to drive a tractor trailer as a golf cart - it's overkill LOL

I found all of this stuff in the middle of the night last night. I still plan on running a parallel 44.1/16 to the Marantz at some point to compare a dithered 96/24, but at least now I can do my redbook as well as a copy for DVD-A with no problem.

I'll post an update if something goes south or changes ...