Transferring Multitrack Reels Into Computer?

The space between the noise floor and clipping in digital is so vast that there's no point trying to keep levels on the edge of maximum. You could transfer with peaks at -46 dBFS and the tape noise will likely still be 20+ dB over the digital noise floor. And you can gain it up with essentially no consequences.

If your tracks are all hot, you're going to run out of headroom on the main bus anyway and have to run your faders way below 0 or have to gain down the clips. In digital, -18 (ish) dBFS is the new 0 VU.
 
Yes, it does sound counter intuitive if you are coming from tape, but it's a matter of perspective. You are putting a 50 ft ladder in a 120 ft room. There's lots of room above and below if you put things in the middle.

If you start doing digital recording, you begin to realize how much headroom you have when the ambient noise floor in your quiet room is 40dB louder than you system's noise. Effectively, you have dead silence from the equipment. Some systems are going to 32 bit floating point which gives and incredible 1500dB of range. While there's nothing made that can handle that range, it is simply a situation where the math is easy for a computer to handle. It effectively rules out any possibility of not being able to handle the sound levels.

Re: the Teac/Tascam branding, there are brochures on the HifiEngine site that put the 3440 in with the 22-4 and 80-8 Tascam line, even though it says Teac on the label. Its kinda like a Lexus is a Toyota underneath. Same basic design and parts, but targetted to a different audience.
 
Last edited:
The thing that goofs people up is the difference in metering between analog and digital. Depending on your interface, 0db VU in the analog realm equals about -18dbfs in the digital realm. To confuse things further, vu meters are slow and read average level while digital meters are fast and read peak levels.

So, with proper gain staging, transferring from tape at 0dbvu would probably give you peak levels of around -6dbfs on the digital side.
 
bmg said:
Maybe it would be easier to change the numbers on the meters - replace -18 with 0.
Problem solved.
Click to expand...
Except for the average level vs. Peak level thing.

In analog, it is important to keep the average level at a certain point, the peaks kind of take care of themselves. Since tape in non linear, the goal is to get the signal as high above the noise floor as you can without distorting too much of the peaks. So average level is what you need to keep track of.

In digital, it is important that the transient peaks don't hit the digital ceiling. Digital is linear and the noise floor is so low (with 24 bit) that the peak level is the only thing that matters.

That's the reason for the different kinds of meters and the different scales.

Also, don't misunderstand, on a digital meter you are trying to put the meat of the signal (the average level) around -18dbfs (not the peak level) while keeping the peaks below 0dbfs.

The reason to keep the average level at -18dbfs is because you have analog gear feeding the digital converters. If you push the analog gear hard enough to get a non-percussive signal (violin, etc...) to get close to 0dbfs, you will likely be distorting the analog signal path.
 
bmg said:
Maybe it would be easier to change the numbers on the meters - replace -18 with 0.
Problem solved.
Click to expand...
They're addressing different technical limitations of each system. With tape, you've got a narrow window between noise and saturation, and the saturation is not instantaneous. You need a really clear target level to keep above the noise without getting too far into saturation. In digital, you've got a massively wide window, but the top limit is a hard boundary that you need to stay well away from.
 
Is this something of concern?
alt=
 
bmg said:
Is this something of concern?
View attachment 116499
Click to expand...

You recorded at 48kHz sampling rate but your audio interface is now set to a different sampling rate. Go to File->Project Settings and check the playback and render resample mode. Check these are set to R8Brain which is almost transparent. The default used to be a fairly low quality setting which made everything sound a bit fuzzy but they may have changed that recently. If you don't want to resample then you could also tick the Project Sample Rate box in Project Settings and choose 48000 on the dropdown box.
 
bmg said:
Is this something of concern?
View attachment 116499
Click to expand...
It's not the worst thing that could happen, but it's not ideal and it should be fixable or preventable.

Did it happen immediately on recording or after opening the file in a new project?

Normally, you'd pick a sample rate for a project and set the interface to match. In most cases with basic systems, the interface detects the project settings and matches itself, but occasionally there's a lapse of communication between the DAW and the interface. For example, if the interface only goes to 48 kHz but you have a project set to 192, I could see something like that happening (though I would normally expect an error message saying the interface couldn't match the project settings).
 
These are the settings:
alt=

but I'm wondering can this be the result of not having the interface connected, and listening through just the mac?
 
Yes, if Reaper is set up to read the settings from the audio device, then if you unplug the interface and go to the internal audio chip, and the two have different settings, then it will start up in a different profile. In any case, I have never had an issue with Reaper's resampling.

Also, Reaper does non-destructive editing, so if the original file is 44K and you fire it up into 48K, it doesn't change the original data. It just remembers everything you did up to the current point (as far as editing, eq, etc). Ultimately, when you render the file it will set to whatever values you pick.

For archival purposes, I always capture to higher sample and bit rates, and then when I mix down to 44K, I dither the mix. Then it can be burned directly to a CD. The Clarett+ can sample up to 192K, I believe. My Tascam does 96k, but I usually do everything at 88.2k which is double the CD standard.
 
bmg said:
These are the settings:
View attachment 116502
but I'm wondering can this be the result of not having the interface connected, and listening through just the mac?
Click to expand...
Yes, the Mac is probably working at 44.1kHz while your interface was working at 48kHz. Whatever you do you MUST change those resample mode settings because those settings will sound pretty bad. Make sure you also change them on the default template and any other templates that you use.
 
Yes - this icon disappeared when I connected the interface.
As for the resample modes - these should be set to the highest quality?
 
As far as sample rates, 44.1 kHz is probably adequate. I don't go in for extremely high rates, but I do record at 48 kHz for a couple of reasons. One, I do a lot of video and 48 k is the standard rate for audio in video. Two, if I'm going to resample to the delivery format, I'd prefer to start higher. A sample rate of 48 kHz accommodates both preferences without being too greedy for storage and processing.

But absolutely do record to 24 bit. You might be able to record to 32 bit floating point, but I think that's overkill.
 
You must log in or register to reply here.
Back
Top