Bit Depth/Sample Rate Comparisons...

BigRay

BigRay

New member
Its all personal preference.

I am a location recordist(classical mainly)

I always do two copies of the material

one at 24/96(which i turn into DVD-Audio Discs)

and one at 16/44.1 (for CD of course)

Teddy
 
Well in a 24-bit system, you can work with volume with virtually very minimal change in sound quality when you mix at a low level. In other words, you don't have to worry about losing sonic clarity of instruments at very low levels that we can hear.

I couldn't tell you the specifics, I have to look into my books again.


However, some people believe a 48-bit depth is just around the corner. So I assume that's going to be useful to some people.

In tracking and mixng, a sample rate of 48 khz to 96khz will give you efficient and convincing results at a time when audio interfaces are pushing out samplerates higher than 96khz.

For example, I was told that tracking at anything higher than 96khz is really unessessary. Overkill if you will.

It's always been my practice to record rock material at 24/48 or 96khz and cleaner essembles (jazz, classical, salsa, etc) at 48.

Extremely high sample rates are something better left for mastering.
 
Interesting..

I always record at 24/96 since I do classical...I dont do any mixing much..

essentially I record , open up in Samplitude, add track marks, (cut out long silences)

save the tracked 24 bit version, burn to DVD-A

then I resample, dither (I always use POWR-3 dither), open the saved cue sheet from the 24 bit version, then save those 16/44.1 tracks, and burn to CD.

I have heard that it is better to select sample rates that are multiples of 44.1 if you are going back down to redbook(88.2)(since it is a multiple, less mathematical conversions in the SRC)
 
BigRay said:
I have heard that it is better to select sample rates that are multiples of 44.1 if you are going back down to redbook(88.2)(since it is a multiple, less mathematical conversions in the SRC)
Click to expand...
That is a myth. SRC's upsample before they downsample so it makes no difference. SRC is SRC, the only way to avoid it is to record at the target rate. (impossible for projects that are destined for multiple formats)
 
Yeah>??

Michal , the head engineer at Mytek Digital told me the 88.2 deal...but it was WAAAAY over my head so it sort of went through one ear and out the other.

I just caught "best to record at multiples of 44.1 for later src"
 
BigRay said:
I have heard that it is better to select sample rates that are multiples of 44.1 if you are going back down to redbook(88.2)(since it is a multiple, less mathematical conversions in the SRC)
Click to expand...


Traditionally, you select sample rates based on the frequency range of the content you're recording. So for example, if you're going to record a classical ensemble and you know the instruments you are recording can't produce anything higher than 22,000 hertz (for the sake of discussion), then you double that to get 44,000 hertz. Thats part of the Nyquist theory and preventing frequency fold over.

So technically, you can even go with 44.1khz and still be safe.

Of course depending on your mics (some record well over 20khz) and your source material, you can then choose your samplerates more productively.

That's why it's always good to know the dynamic ballparks of the sources you're recording.
 
BigRay said:
Yeah>??

Michal , the head engineer at Mytek Digital told me the 88.2 deal...but it was WAAAAY over my head so it sort of went through one ear and out the other.

I just caught "best to record at multiples of 44.1 for later src"
Click to expand...

It's old advice from a few years back when SRCs sucked. Check Lavry's board on PSW for the up-to-date deal, he's discussed that a few times.

Having said that if you don't have a good SRC routine available, record at 24 bits and the target rate. If you are recording for CD and DVD-A, I'd use 96.

48 bits? I hope not. That is completely silly. We don't even have converters with 22 bit dynamic range. Even if the laws of physics didn't prevent a minimum noise floor below around -144dB or so, who is recording sounds louder than that? A sonic boom is 192dBSPL, 48 bit is a theoretical 288dB range.

Seems like a manufacturer conspiracy to drive upgrades, just like 192kHz.
 
^^^ 48 bits at tracking doesn't make much of a sence. However higher bitrates are desirable within the DAW during processing/mixing/submixing to introduce as little rounding errors as possible.
 
Farview said:
SRC's upsample before they downsample so it makes no difference.
Click to expand...

is there a link somewhere, or can you explain, why they upsample before downsampling?

Traditionally, you select sample rates based on the frequency range of the content you're recording. So for example, if you're going to record a classical ensemble and you know the instruments you are recording can't produce anything higher than 22,000 hertz (for the sake of discussion), then you double that to get 44,000 hertz. Thats part of the Nyquist theory and preventing frequency fold over.

So technically, you can even go with 44.1khz and still be safe.

Of course depending on your mics (some record well over 20khz) and your source material, you can then choose your samplerates more productively.

That's why it's always good to know the dynamic ballparks of the sources you're recording.
Click to expand...

can you point out some regularly used mics that record well above 20kHz? Similarly, how about some headphones or speakers that consumers or audiophiles are using to listen back on that reproduce up to 48kHz and above for the higher sampling rates?
 
bennychico11 said:
is there a link somewhere, or can you explain, why they upsample before downsampling?



can you point out some regularly used mics that record well above 20kHz? Similarly, how about some headphones or speakers that consumers or audiophiles are using to listen back on that reproduce up to 48kHz and above for the higher sampling rates?
Click to expand...


Well one that I know of is the Manley Gold Reference mic. It records up to 30khz. In fact, I think they have a spot for it in the current sweetwater catalog.

I've never used it, but I've heard insane things about it. It will pick up things you can't even hear sometimes.

As for the headphones and speakers, wow, couldn't tell ya. Although, there is one company that makes some serious mastering speakers that do go up to 30khz.

I personally can hear up to 18khz. So I really couldn't find a purpose to use all that stuff unless I just want to "feel" it.

http://www.wegg3.com/

thats the website.
 
On the subject of high sample rates, it's interesting to have finally learned the skinny on why high sample rates are excessive.

A big factor being mathematical compromises in the design of digital software and gear. More mathematically intensive units mean larger expenses, therefore, quality is compromised to save money. This is especially inherent in cheap digital filtering.

This explains why the Waves Linear Band EQ is so expensive while simpler plug-ins are not. Some even state the Lin EQ sounds almost analog in nature. However, I just feel it sounds unique to itself, not quite analog.


So unless plug-ins become alot better in the next years, high samplerates won't be as beneficial to the mixing and tracking engineer.

However, now I wonder if high sample rates along with good analog outboard gear are a better match.
 
LeeRosario said:
Well one that I know of is the Manley Gold Reference mic. It records up to 30khz. In fact, I think they have a spot for it in the current sweetwater catalog.

I've never used it, but I've heard insane things about it. It will pick up things you can't even hear sometimes.

As for the headphones and speakers, wow, couldn't tell ya. Although, there is one company that makes some serious mastering speakers that do go up to 30khz.

I personally can hear up to 18khz. So I really couldn't find a purpose to use all that stuff unless I just want to "feel" it.

http://www.wegg3.com/

thats the website.
Click to expand...

basically my point was, the Nyquist theory doesn't really matter above normal CD sampling rate. The average microphone or speaker is not going to reproduce the frequency. Which, I know all of this was stated above...but I was just reiterating that no one is going to be able to hear or feel those frequencies
 
noisewreck said:
^^^ 48 bits at tracking doesn't make much of a sence. However higher bitrates are desirable within the DAW during processing/mixing/submixing to introduce as little rounding errors as possible.
Click to expand...

Yes. But that is already implemented. In fact my very first DAW ran internally at 32 bit fixed, and that was five years ago.

It's when somebody tells me I'll need 32 bit converters that I want to knee them in the groin :eek: ;)
 
LeeRosario said:
On the subject of high sample rates, it's interesting to have finally learned the skinny on why high sample rates are excessive.
Click to expand...

The biggest piece is that the sample rate is really just the data rate. A given converter running at 192 or 96 is actually running the same physical sampling rate, something north of 2mHz. So to construct a higher data rate, the converter is relying upon fewer data points for each sample. Thus, higher sample rates produce higher bandwidth, but at the expense of sample accuracy.

When that bandwidth is totally unnecessary, it's just a waste.
 
You must log in or register to reply here.
Back
Top