16/44 vs 24/48

  • Thread starter Thread starter dobro
  • Start date Start date
dobro

dobro

Well-known member
I'm mixing an album of songs right now. I recorded all but two at 24/48, but because of problems I was having both with my computer and my brain at the time, I tracked two songs at 16/44.

These two sound different. It's harder to make them sound good. It's like the sound is raw. It reminds me of photos I used to take with 400 ASA film - lower resolution.

It might have been due to sloppy mic technique, but I don't think so. I think it's the difference in resolution. And if *I* can hear it, anybody can - my ears are at the other end of the scale from golden ears. What's that called? Tin ears?
 
Not tin ears- just human, and a realistic human at that. It takes a lot of careful listening to hear some of these effects, and some of them are really more the voices of the angels (like the difference between the high-zoot speaker wires and regular old 14AWG power cable) than actual concrete "that-sounds-bad" cruft.

I've spent a lot of time trying to hear a lot of these things, and I certainly *can* hear some of them repeatably. 16 vs 24 bit is easy to hear, especially if you use a lot of DSP in a DAW. There's no question that that effect is real: it is a grainy, nasty effect, especially out in the reverb tails. Worse yet, the more you mung the original tracks with plugins and the like, the worse it usually gets (from DSP resolution and roundoff errors).

Truncation from 24 to 16 bits is equally nasty. Bad dither and sample-rate conversion algorithms are not difficult to hear as well. Where my perceptual apparatus falls apart is hearing the difference between 44.1 and 48 kHz sample rates on the oversampling converters on my own rig: I just can't do it. I _think_ there are differences, but could I consistently identify the one versus the other in a blind test? Nope. I'd be lying if I said that it was that stark, to my aging ears.

Which is why I stopped cranking the resolution at 24/44.1. My hardware can do 48 or 96, but if I can't hear it, I don't want to pay for the additional storage needed...

Nope, You don't gots tin ears. Just ears that are used to optimizing the mix to what is available at the moment. Present them something better, and they adapt quite quickly, don't they?
 
Cheers, Skippy. Everyone I talk to except about one person says the same thing about the difference between 48 and 44. I'm looking forward to listening to the difference between 96 and 44.
 
I agree that 24 / 96 sounds much better, but unfortunately we still dither down to 44.1/ 16 for CD audio. There isn't an exact equasion that selects the unwanted frequencies when you dither, it just pulls them out at random. Until we have affordable media that handles 24 / 96 I perfer to save hard drive space and master at the resolution the I intend my final product to be at.
 
I can still hear a very distinct difference between 44.1 and 48KHz sampling rates. 48 sound much more open in the top end. But, the effects of software sample rate conversion ALSO make a very distinct difference too!!! Not a good one either. This is a reversal of what I said over a year ago so go ahead and sue me! :)

Jumping from a 16 to 24 bit CONVERTER made a difference, although, I can hear NO difference between recording at 16 when using a 24 converter if I do nothing to the audio afterwards compared to recording at 24 bit. Most preamps don't have 96dB of s/n ration anyway....:) But, because I usually will be messing with the audio in the computer, I record at 24 bit.

It is when I need to mess with level and eq that the 24 bit file starts showing it's advantage. DSP applied over a 24 bit file just plain sounds better. Much more smooth and pleasing to the ear. Much more predictable too.

Recording at 24/44.1 is probably the best way to go if you are going to apply DSP. This gives you the added resolution to move quantiniztion errors beow what your ears can hear while applying DSP, and it means you don't need to SRC with a crappy converter in software.

I DO NOT recommend recording at 16 bit if you can record at a higher bit depth if you are going to mess with levels and eq with DSP.

Ed
 
sonusman said:
I DO NOT recommend recording at 16 bit if you can record at a higher bit depth if you are going to mess with levels and eq with DSP.
Click to expand...
Ed,

So, does this mean there is no advantage to converting a 16 bit file to 24 bit when applying DSP? How about if you added low level noise? Wouldn't this be similar to recording a 96 dB source in 24 bit?

I use a lot of samples from CD's and was hoping I could improve the processed sound.

barefoot
 
I have noticed a bit of improvement in the DSP when I "cheat" up a 16 bit recording to 24 bit.

Ed
 
Why not 88.2 instead of 96kHz??

I agree that 24 / 96 sounds much better
Click to expand...

Why is it that everyone always talks about 96kHz as the upgrade from 44.1 or 48kHz???

My PC isn't powerful enough at the moment, but if I could I would record at 24/88.2 and NOT 24/96...Correct me if I'm wrong but I believe you get a better result downsampling from 88.2kHz (since most of us don't have expensive outboard downsamplers), no???

I mean can anybody really hear a difference between a file at 88.2kHz and one at 96kHz????

So why not go with 88.2kHz instead of 96kHz?? Any specific reasons???:confused:
 
Two Against Nature

My understanding is that this Steely Dan album was tracked at 16 bit 44.1Khz. Would you guess that all DSP was done outside of the DAW, and brought back in for final mix/master?

Albert
 
Who knows. If the CD mixed via analog processing, an excellent sounding A/D/A converter at 16 bit 44.1 sampling rate would be just fine.

Ed
 
Well, I'm glad I'm not crazy. Or at least I have company.

On my gear - which maxes out at 24/48 - I did some test recording early on. Maybe it was the material I used, but I could hear no difference between 44.1 and 48. But I heard a significant difference between 16 and 24 bit.

I now record everything at 24/44.1. I figured the odds of introducing an error during the inevitable 48 to 44.1 conversion outweigh the inaudable sound difference.
 
Think the main difference you heard was the difference between the noise floor of 16 bit and 24 bit converters. Also, better chipsets have fallen in price, and dithering while doing the A/D process have improved. This would account more for a difference in most cases between 16 and 24 bit.

A great way to check this out:

Use a 24 bit converter and record the same source, preferrably something predictable in amplitude like a distorted guitar, with one recording being stored at 16 bit, and the other at 24 bit. Same converter, just two different storage depths. I BET you wont' hear a difference if you do nothing to either file afterwards. Play them side by side and they will sound identical. Of course they will, they both enjoyed a much lower noise floor, and I doubt most on here even have monitoring systems that they can hear more than 96dB cleanly anyway.

Now, compare the 16 bit version recorded with a 24 bit converter to a 16 bit recording recorded with a 16 bit converter. You WILL hear a difference there!

While the difference between 44.1 and 48KHz sampling rates are slight, I still hear a difference. You really start to hear a difference when you stack up the tracks! But, I think that bad SRC is worse than starting out with a lower sample rate. But, this may also be offset again by the fact that a 48KHz file would contain more samples for DSP to work with. This will of course help the DSP do a more effective job.

Probably the best thing would be a outstanding SRC unit. sjoko2 swears by the Lucid SRC he owns. I will not dispute his claims that it offers transparent SRC. He also agrees that working with higher sampling rates with DSP makes the DSP work better. He can enjoy that luxury with a great SRC unit though. I can't afford the unit he is using myself though....:(

In the multitrack environment, having higher bit depth means less noise. Higher sampling rate means a better detail in the sound, especially in the higher frequencies. Great and all good if you run those tracks out to a analog console that is better than a Behringer of a Mackie. Yes, even analog gear has limitations in the higher frequency is will pass through that is uneffected by phase shift (distortion). Better unit in analog pass significantly higher frequencies than the cheap gear.

Oh blah blah blah.....This gets involved! Not up for it today....

Ed
 
You must log in or register to reply here.
Back
Top