Hi all,
I’m currently using version 3 on a Pent 3 550 processor. Anyone successfully using version 4 on something this “slow”?
Also, if I’m mixing down to 16 bit CDs, is the 24 bit version of N-track something that’s going to make any (hearable) difference on the end CDs? I’m using an Audigy 2 platinum sound card that is capable of 24 bit.
Thanks, James
24 bit is a pretty significant difference really. If you have the capability, I would use it personally. I went from 16 to 24 bit a few years ago and I never want to go back. It is just so much easier to work with even if the audio improvement isn’t significant (which it was in my case). The fact you aren’t worried about clipping all the time etc. makes it worth it to me. In addition, if your convertors are decent enough, it WILL sound better… double bonus.
Thanks for your reply. I’m not clear on convertors. I thought I could just record and work with the files in 24 bit and then convert in N-track to 16 bit files before I burn to CD. Are there hardware convertors in the soundcard? (I’ve been doing this for a few years but a lot of times feel like I’m destined to be an eternal “newbie”.) Also, what type of processor are you using?
Thanks again, James
I wouldn’t try N-track V4 on your machine. Stick with V3.
And I agree with Bubba about 24-bit vs. 16-bit. There’s really no point in not using 24-bit. Even on a slow machine – it doesn’t affect CPU usage.
Yes, you record in 24 bits and do all your work in 32-bit floating point (n-Track’s internal format). If at any time you save a mix and then move it to a new song file (which many of us do – more on that in a moment), save it as 32-bit.
Then, when everything else is done, convert to 16 bit, with dithering. n-Track does a fine job of this, but folks with very discriminating ears say that other programs can do it better and I’d take their word for it. (Different kinds of “noise shaping” is the only difference between programs. Your mix has to be pretty fantastic for it to make much of a difference.)
The reason we save 32-bit mixes is that work is usually divided into two stages: mixing and mastering. Mastering is a big subject, but suffice it to say that it’s applying compression, eq, and reverb on the whole mix, and balancing the different cuts on a CD so they fit well together in a collection. So, when we’re happy with the mix, we save it as a 32-bit file and open a new song file with every cut on it, so we can easily compare tracks while mastering.
The other reason we separate mixing from mastering is that it’s usually best to have another person do the mastering, for objectivity. Lots of folks have professional mastering houses do it.
Thanks for all the information! I’m going to stay with version 3 and just upgrade to the 24 bit. It’s been over a year since I’ve been on this forum and now I remember what a supportive and freely giving group of people you are. I hope I get an opportunity to pass along the help to others sometime.
James
| Quote (James927 @ April 01 2006,13:32) |
| Thanks for your reply. I’m not clear on convertors. I thought I could just record and work with the files in 24 bit and then convert in N-track to 16 bit files before I burn to CD. Are there hardware convertors in the soundcard? (I’ve been doing this for a few years but a lot of times feel like I’m destined to be an eternal “newbie”.) Also, what type of processor are you using? Thanks again, James |
Okay, let’s not confuse “convertors”. When someone refers to convertors in this game they are referring to the piece of hardware that converts an analog signal to a digital signal. The convertor is usually on your soundcard, but more expensive convertors come as separate units. So there si a huge range in quality of convertors. Anything that goes from digital to analog or vice versa have a convertor in it… everything… sound cards, CD players, DVD players, Nintendos, XBoxs, PlayStations, those crappy birthday cards that talk. Obviously convertors are everywhere and range in quality from absolute crap (birthday cards and answering machines) to very high quality (those used in professional film and music production).
Now, converting from say 24 or 32 bit to 16 it is another process. In high end applications this is done again with a hardware device… but most folsk liek us will use some sort of software utility like r8brain or a wave editor like GoldWave. Of course, depending on the algorithm the programmer uses, the sound will be different (unless you are using a direct truncation). In this process is what is known as dither. I would try to explain that, but Ozone has an excellent paper that describes better than I ever could. Ozone Dither Guide
As for processor, I am using a P4.
BTW, keep in mind that a P3 is considerably faster than a P4 of the same MHz. P4’s were a big disappointment. Modern designs (Centrino) derive a lot more from P3 than P4. So, think of your 550 MHz P3 as more like a 750 MHz P4, maybe faster.
I did a lot of nTracking using a P3 750 MHz and it worked great as long as I didn’t use too many plugins, and if necessary you can always manually “freeze” portions of the mix (mixdown, import the result, and mute those tracks while working on other parts).
BTW, I suspect that memory is as much of a factor as CPU for n-Track V4, maybe more. I suggest 512M minimum.
Thanks Bubba and Jeff. I’ll check out ozone’s info on dithering.
Freezing - what a great idea! Kinda reminds me of bouncing tracks on my analog 4 track recorder (but for extra proc/memory power as opposed to extra tracks). I’ve got 512 megs of memory. Thanks also for the info on P4 procs. Makes me feel a little better about my P3!
Thanks again, James
<!–QuoteBegin>
| Quote |
| And I agree with Bubba about 24-bit vs. 16-bit. There’s really no point in not using 24-bit. Even on a slow machine – it doesn’t affect CPU usage. |
So we can still upgrade 3.3 to 24 bit eh?
<!–QuoteBegin>
| Quote |
| It is just so much easier to work with even if the audio improvement isn’t significant (which it was in my case). The fact you aren’t worried about clipping all the time etc. makes it worth it to me |
Why is 24 bit less problematic-less clipping? if you don’t mind explaining that.
I am still using 3.3 16 bit with my PIII 800Mhz and 512 Meg of RAM and am pretty happy with the stability thing.
Yes, less clipping. That’s the short answer. Here’s the long one: 
The difference is the noise floor is lower, which means you can leave more headroom.
First, we think of digital as having a zero noise floor, and that’s true that the silence is dead silence. But whenever we have a signal recorded, it’s quantized to however many bits we use (16 or 24). That quantization means the signal isn’t exact, thus it’s noise. How much? Well, with 16 bits, remembering that each bit is worth about 6 dB, it’s about 96 dB. Well not quite that much for technical reasons, so let’s call it roughly -93 “dBFS” (93 dB below full scale). That’s pretty darn good, considering that we’ve used tape with a noise level of -70dB and been pretty happy. Well, there’s noise and then there’s noise. Tape hiss is pretty much white noise, which the brain is a master at ignoring (filtering out). Quantization noise is a kind of noise the brain is very good at hearing, and it’s a nasty and nonmusical noise. So, it’s not nearly as much better as the numbers would indicate (still better, just not that much better).
I’m sure you know how, when recording on tape, how important it is to get a good beefy signal, so that the hiss isn’t too loud on playback. The same is true of 16-bit digital (though not quite as critical) – you want to record a signal that peaks as near to 0dBFS, because lots of the signal – e.g., the tails of notes, will be well below that, at closer to -18 dB or lower, and it still needs to sound good. Here’s another spot where tape helps and digital doesn’t: as you know, if you peak a bit with tape, it tends to compress the signal and almost sweeten it. Not so for digital, you get complete crap if you go over 0dBFS – sounds terrible to everyone except a few metalheads that like nasty sounding stuff.
So, when recording using 16 bits, we sweat the levels, and make sure each track, when recorded, peaks at or above (say) -6dB. Meaning we toss a lot of otherwise good performances because we played too loud in one spot, or didn’t have the level set high enough. (Note that in a pro studio this would be frowned upon! “What? I gotta do it again? But that was perfect!”)
Another thing about 16 bits is, if you’re serious about getting the best recordings, you have to very carefully calibrate your system so that the 93 dB signal range of the digital gear is aligned with the sweet spot in your analog signal chain. So that, when you’re hitting near 0 dBFS in digital, you’re hitting say +6dBVU on the analog gear. Any discrepancy in settings will cause you to lose signal clarity, either from the digital gear or the analog gear. You have to sweat levels in two domains at once: analog and digital.
Well, when using 24 bits, that quantization noise floor is down at about -141 dBFS. It’s wide, plenty of room. This allows you to focus on the sweet spot of your analog gear, with plenty of digital headroom, so that if you peak digitally, the signal was already crapped all over by your analog gear. For example, setting it so that +18 dBVU on analog is 0dBFS on digital. Now, you watch your analog levels and don’t worry about the digital: if the analog gear gets it, the digital has it nailed.
This means that you can set your levels lower (digitally), so if there’s an unexpected peak (but not so high as to sound like crap in the analog gear), you recorded it faithfully and it’s good. Most of the signal is now well down from 0 0dBFS – you might plan to peak over -12dB or even lower and not worry about the noise floor – at least, not the digital one.
So, bottom line: easier not to screw up. Leave more headroom for safety without losing clarity. Less sensitive to being perfectly tuned to your analog gear chain.
Finally, the accepted wisdom is that even though few people (maybe nobody) can tell a well-recorded 16-bit track from a 24-bit one, the mixes that result are distinguishable by trained ears, as a subtle clarity. I have my own theory about this, but I won’t muddy the water about it today.
Now, where is my bow tie?
<!–QuoteBegin>
| Quote |
| Finally, the accepted wisdom is that even though few people (maybe nobody) can tell a well-recorded 16-bit track from a 24-bit one, the mixes that result are distinguishable by trained ears, as a subtle clarity. I have my own theory about this, but I won’t muddy the water about it today. Now, where is my bow tie? |
And your slide rule?..
learjeff,
Thanks for the explanation. I may have to read it a couple of times. I haven’t recorded 24 bit so I don’t have any comparisons. If I can upgrade 3.3 to 24 bit I will probably do it. I don’t want to overtax my PIII with any of the version 4 builds. Don’t think that I have ever listened to a side by side comparison or 24 bits / 16 bits and don’t know if my over 50 ears could hear the difference but if recording 24 bits would make recording easier, I am all for it.
Thanks again,
Doug W
It’s technically possible (to upgrade to 24 bits), just a matter of whether Flavio is willing. If you can’t order it on the website, just send him an email and ask: I bet he’d be willing to (if he still can generate the numbers) since you have a reasonable excuse for running older code. (Plus Flavio might say that V4 doesn’t take more CPU than V3. I tried a few early versions that did bog down, but that might not be meaningful.)
In any case, moving to 24 bits doesn’t increase CPU, just adds 50% to disk usage but that shouldn’t be significant in most cases. I was able to run a dozen or so 16-bit tracks on a 750MHz P3 laptop with 4500 RPM hard drive, so unless you’re on a laptop or run a very large number of tracks, you should be fine.
Most of our recordings are at most 4 vocal tracks and 3 to 4 instrument tracks, with a couple of mini-tracks to cover mistakes-I made one once. I throw a couple VST effects in on the AUX channels and 1 or 2 on individual tracks, so normally I don’t stress my recording 'puter too much.
PIII 800 Mhz
512 Meg RAM
2 7200 RPM hard disks
Doug,
Flavio is still offering version 3 upgrades to 24 bits - I got mine today. Happy recording all!
James