why bother with 24 bit

[boogle]

I wish this myth would die already. This isn't true. Any time something is resampled, it is upsampled into the MHz range and then downsampled. You've got to remember that your converters are actually sampling a lot faster than the sample rate (oversampling). That gets paired down to the sample rate that gets stored on the computer. There is a theory that super high sample rates (192k) are actually less accurate because each sample is based on less information.

Thanks, I am still learning as well.

 

[Nick The Man]

would 48 be an ideal sample rate? i was reading my "home recording for dummies book" .. haha, and Jeff Strong (the author) said that a standard is 24 Bit 48Khz

 

[Farview]

would 48 be an ideal sample rate? i was reading my "home recording for dummies book" .. haha, and Jeff Strong (the author) said that a standard is 24 Bit 48Khz

For CD, 44.1k would be the way to go. 48k is for video.

 

[IronFlippy]

I prefer 1-bit PWM. But most mediums don't use that type of digital audio, so I use PCM 24-bit/48kHz. Alright, I'm being a smart ass tech geek, I apologize.

In short, recording at 24 bits and dithering down to 16 yields a better result than recording at 16. Plus, when you record at 24 bits, you don't have to worry as much about recording at too low a level, which can cause errors in 16 bits. In the digital world, you don't want to be anywhere near 0db. The lower you can record it, the more you can play with the sound during mixing.

I wish I knew more about this than I do, I'm brand new to all this.

 

[cusebassman]

You guys are making me feel bad about my 16-bit digital tape system, even though it sounds good to me!

...I should feel bad... I can't wait to upgrade to an HD24

 

[MCI2424]

whats the point of recording in 24 bit? when you put it on a CD it becomes 16 bit anyways so why waste space on your computer and i think its hogging soem of my processor speed too.

Same with sample rates higher than 44.1. Why bother? when its on a CD its reduced to 44.1. the human ear can only hear up to 20khz anyways.

i must be missing something?

the reason i bring this all up is because ive noticed that when i record at 24 bit my computer seems like a slug and i get many pops. if someone thinks they can help me, ill go into better detail on what my computer actually does.

24 bit is cooler. 16 bit is run-of-the-mill and we all want a bit more. 2track went to 4 track, then 8 track,16 track and finally 24 track. So, bits is like tracks these days. Everyone wants more it seems. Next will be 32 bit and then 64. Is there no end to the snobbery? More resolution means more analog type gooyiness slobbered all over a luscious, dense mix. Makes your mouth water, no?

 

[boogle]

I searched around on the forum and could not find a direct answer to this question. The point of oversampling was brought up earlier as contradiction to the concept that 88.2 to 44.1 was simply the decimator dividing by two for the sample rate. My confusion rest in the fact that if I understand correctly oversampling is usually for a/d and d/a conversion (primarily to alleviate stress on the filter) . For internal downsampling where the conversion is simply d/d why is oversampling used, and if it is not than why does the 88.2 to 44.1 divide by 2 rule not apply.

 

[boogle]

24 bit is cooler. 16 bit is run-of-the-mill and we all want a bit more. 2track went to 4 track, then 8 track,16 track and finally 24 track. So, bits is like tracks these days. Everyone wants more it seems. Next will be 32 bit and then 64. Is there no end to the snobbery? More resolution means more analog type gooyiness slobbered all over a luscious, dense mix. Makes your mouth water, no?

By the way I do not agree with that.

 

[mohthom]

If you've not come across it, there's a really useful article I read a while back that has some really great explainations of samplerate and bit-depth stuff.

It's called The Practise of Mastering in Electroacoustic Music (not as in electroacoustic guitars, more as in acousmatic [Schaeffer, Dhomont] music) and it was written by Dominic Basaal, an ME based in Montreal (I think).

It's on the web if you're wanting to find it, and it's a bit of a bible. Essential reading, and I think it'll cover your question. It's around page 28 if I remember correctly, or that might be the samplerate stuff.

Hope that's of use,

Thom

 

[Farview]

For internal downsampling where the conversion is simply d/d why is oversampling used, and if it is not than why does the 88.2 to 44.1 divide by 2 rule not apply.

Because it is mathmatically more accurate if you upsample then downsample.

 

[chessrock]

At one point in time, a long long time ago
...

1/8" Cassette tape was the most popular and prefered final medium for music.

But for some odd reason , music was still initially tracked using 2" tape. You could pose the question: "Why not just track everything to 1/8" cassette, if that is ultimately going to be the final medium anyway?"

The reason that practice wasn't adopted is because, well, it sounded like ass. And it's really the same way with digital. If each track isn't tracked at the highest possible resolution / fidelity, then that lack of quality is simply multiplied by the total track count. On the other hand, if every track is of the highest possible recording quality to begin with, then the final mix will be far superior, regardless of what format or resolution the final master is ultimately delivered.
.

 

[boogle]

I completely understand the argument of track at the best quality. But if you upsample and then downsample by say 128 or some other fixed number doesnt the divide by 2 rule still apply to 88.2 down to 44.1 as being an easier calculation for the converter?

 

[Farview]

I completely understand the argument of track at the best quality. But if you upsample and then downsample by say 128 or some other fixed number doesnt the divide by 2 rule still apply to 88.2 down to 44.1 as being an easier calculation for the converter?

You are forgetting about the interpolation on the upsample. It isn't just multiplying by a certain number and dividing again. The process 'connects the dots' between the original samples and then averages them to create the target samples. How it connects the dots is why some sample rate converters sound better than others. Or more accurately, sound different.

 

[NYMorningstar]

new problem that ive run into .. i dont think its anything to do with bit rate and such like that but here it is.

when i pan things there is a problem, its loud and normal in the center. but then i pan it 100% left and it is all in the left but its really quiet... and then like 50% left its like centered but just quieter .. i can figure it out.. any ideas?

What? No ones going to answer his stereo panning question?

 

[SouthSIDE Glen]

What? No ones going to answer his stereo panning question?

He took that to another thread. Case already closed

G.

 

[boogle]

You are forgetting about the interpolation on the upsample. It isn't just multiplying by a certain number and dividing again. The process 'connects the dots' between the original samples and then averages them to create the target samples. How it connects the dots is why some sample rate converters sound better than others. Or more accurately, sound different.

Ah hah, that makes sense, I will do some more research on that but thank you very much, that has been eating at me all day.

 

[dgatwood]

I wish this myth would die already. This isn't true. Any time something is resampled, it is upsampled into the MHz range and then downsampled.

You're right that it is not as simple as throwing out every other sample because of the need to avoid aliasing (Shannon-Nyquist limit). However, you don't have to upsample to downsample unless the old sample rate is not evenly divisible by the original rate.

Downsampling to an uneven rate is done by adding a low-pass filter to remove aliasing, then boosting the rate to a higher frequency (probably doing interpolation to generate intermediate sample points), then downsampling it. Downsampling to a rate that is an even multiple can skip that second step entirely, and thus is as simple as applying a low-pass filter and then throwing out every other sample. If you upsample to any frequency that is an even multiple of both sample rates, you will always get the same results back as though you just threw out every other sample.

If you'd prefer a more accurate set of values, you could weight the sample as 1/2 S[k] + 1/4 S[k-1] + 1/4 S[k+1] to smooth things. It probably isn't a bad idea to do that, really. Either way, the math is a lot simpler than doing an uneven frequency reduction.

If you have a theoretically perfect interpolation, then an uneven division is just as good as an even division. In reality, an equal division is slightly more precise, though the difference is so minimal that it would be way below the noise floor of even the best converters. If you did it a few thousand times, you might notice a difference....

You've got to remember that your converters are actually sampling a lot faster than the sample rate (oversampling). That gets paired down to the sample rate that gets stored on the computer. There is a theory that super high sample rates (192k) are actually less accurate because each sample is based on less information.

If that's really a theory, then it is complete crap. The 192 kHz sampling is the average across a shorter period of time, and is thus, by definition, a more precise approximation of the value at a single point in time.

Consider a survey conducted by random telephone polling. You want to know the typical opinion of a person in San Francisco on some subject. You interview people in SF, San jose, and Sacramento and average the results. This might be a more accurate representation of the opinion of Northern California or it might be less accurate if Sac and SJ are not representative of other areas. In any case, it is a less accurate representation of the opinion of San Francisco.

 

[Tim Gillett]

Interesting someone asks on another thread for more info on the Miles "Kind of Blue" sessions from '59.
From what I know that would have been tracked on a recorder with maybe 65db signal to noise.
What does that equate to today? Way below 16 bits, and on the tracking tape.

Sure, they would have somewhat compressed the dynamics. Sure it's a compromise recording but, hey, it sounds great musically and otherwise.

Tracking even at 16 bits gives options that were only a dream back in '59.

If we cant make a classic recording tracking at 16 bits (s/n 96db!) is there possibly something lacking in our technique?

Tim

 

[cjacek]

One can't compare analog and digital in the way you have, Tim. That tape from '59 surely had much better resolution and sound stage in comparison to a 16 bit CD, despite it [the CD] having better S/N ratio. Specs mean little in a real world listening environment.

 

[brendandwyer]

well i just got a great deal on a radar 1 16 bit 24 track recorder.

i'll report whether i notice difference, although, that might point to my own inadequacies more than that of the recorder