Sample Rate Question.
 

I'm wondering what Sample Rate most people on the AGF set to record acoustic guitar. My understanding is that for CDs 44.1 kHz works and for film it's best to jump up to 48 kHz. Beyond that is there a qualitative reason to go higher? I have a vague awareness of the Nyquist - Shannon Sampling Theorem but given that the frequency range of a guitar is roughly 80 Hz up to 800 Hz it's hard to see the advantage of a higher Sample Rate for our instrument.
With that out of the way there are several guitarists who regularly contribute to this part of the forum and make great sounding recordings. I know that the Sample Rate is just one factor among a many but would still appreciate hearing what setting folks here prefer.
FWIW - I'm recording with an Apogee Quartet > MBP 2015, 15" > Logic Pro X (10.1.1)
Thanks,
Trevor

There is no scientific evidence anyone can reliably pick out a higher sample rate and bit depth audio file from 44.1/16 bit.

You should record at 24 bit because it allows you to record at a more relaxed audio level, allows you to give yourself headroom with no audio penalty.

You will hear some people suggest they can hear the difference. They will tell you about openness and cymbals and definition and other non measurable criteria. I personally cannot say for sure they aren't hearing things, but I do know when I did blind abx testing I could find no difference. Zero. In casual non ab testing I certainly THOUGHT I could hear a difference, but.. Expectation bias I guess.

There are lots of places you can go go read about this online as well.

The benefit to higher sampling rate audio is more high frequency extension, which is above the ability of humans to hear. Your dog might prefer higher sample rate audio though. :)

The 48/44.1 question is simple too, just IMO record at 44.1 if you are going to release your music as music. If you are writing for video then go to 48.

The one real world advantage to higher sample rate audio is it allows the low pass filter to be at a less severe slope, but today's audio converters (combined win oversampling) are so good I have never personally heard a difference and I suspect there might not be a difference that humans (or at least most humans) can hear.

Oh, and there is a free abx testing app on both windows and Mac. So you could do your own testing. If you do, report back!

Albeit slightly off topic it is worth mentioning. Televion and film production uses 48k strictly for an age old Avid compatibility issue. I don't think anyone in post production is arguing sonic quality in that specific situation.

Okay. So do you agree with Psalad that there's no discernible difference between a Sample Rate of 44.1 kHz and a higher rate when recording acoustic guitar?

I've never read of anyone being able to discern a difference between 44 and 48 but there are plenty of people who say they can between 44/48 and 88/96.

I've never heard of anyone who could reliably pick out 96/88 vs 48/44 in a double blind test either. So either there IS something there that we can't pick up in ABX testing but we can in casual listening... Or those casual listening differences can be chalked up to expectation bias.

You might be right about the source of the preference, but today everyone I know in video (including the video team who work for me) records at 48. Just sort of the standard.

Here's a premise that may help clarify my question. I sometimes hear guitar recordings (a few from AGF members) that have an open or airy sounding top end that most acoustic guitar recordings don't have. So just assuming for now that all the other elements are equal (room size and treatment, mics, quality of guitars and playing skill etc, etc,) is it possible that higher a Sample Rate captures some sort of high end room ambience?

Of course it will capture it, if the microphone captures audio above 20k and if you are recording at a higher sample rate. There is no question it captures it. There is no question an acoustic guitar will have harmonics above 20k.

The question is whether you can hear it, or more specifically, if there is any magic at frequencies you can hear due to frequencies you cannot hear. Science says no (at least the science I've read). Some audio people say yes.

In the case of recordings posted here, I doubt very much if people are posting 96 or 88 MP3s. So I'm pretty sure the difference you are hearing has nothing to do with frequencies above which humans can hear.

I hear some jaw dropping recordings of acoustic guitar on CD or even MP3. It's about the player, the placement of the mic, the room, the microphone, and the knowledge of the engineer to process the recording.

Good answer; much appreciated!

I used to be able to hear the extended HF at 48kHz, but it wasn't really a game changer since the CD release format (for however long we may have it) is 44.1 kHz. I haven't tested myself recently.

Regards,

Ty Ford

Fukushima A, Yagi R, Kawai N, Honda M, Nishina E, Oohashi T (2014) Frequencies of Inaudible High-Frequency Sounds Differentially Affect Brain Activity: Positive and Negative Hypersonic Effects. PLoS ONE 9(4): e95464. doi:10.1371/journal.pone.0095464.

The conclusion to this study was:


Obviously one study this small is not a game changer but it is indeed an interesting study.

I record at 24 bit, 88,200 hertz.

24 bit for the headroom to prevent overs

88,200 probably for no particular reason with my current high quality gear. However with some A/D D/A converters the filters are not audibly transparent at 44,100 hertz. For final release, resampling from 88,200 to 44,100 for a CD is more precise than sampling down from 96,000.

That said you can probably do totally fine recording in 16 bit at 44,100 hertz. The file sizes are relatively small, and especially with a slower computer the latencies may be noticeably less and digital editing may go faster.

Hi Trevor.
If you'd like to understand both sample rate and bit depth I have a relatively easy to understand explanation on my "Simple Homestyle Recording" page:

http://www.bluestemstrings.com/pageRecording1.html

This has been the topic of much discussion and debate particularly since computer performance and storage has become so high.

And interestingly enough there is very little if any "scientific research" on this at all.


So while it is probsbly accurate to say (as Psalad) pointed out
"There is no scientific evidence anyone can reliably pick out a higher sample rate and bit depth audio file from 44.1/16 bit."

It is also probably accurate to say:
There is no scientific evidence that recording at higher sample rates does not make any kind of audible difference.

I do remember several discussions on Gear Slutz, that got very technical and (I do not remember the details but to paraphrase very simplistically) it had to do with plugins functioning better or more efficiently, at higher sample rates.
and as Rick pointed out some converters operating better at higher than 44.1k

Also it should be noted that many if not most engineers who are the ones producing the music that is at the top levels of commercial productions, getting grammies, and recognized as the shakers and movers etc. Are recoding at higher sample rates.

Also I think Rick hit the nail on the head. Today it probably depends much more on the spec's of the particular computer. If it is newer and has the speed and storage to handle 88k or 96 k files there arguably no reason not to


As far as the argument about anything above human hearing range "having no audible effect". I think the jury is still out on that one. I think much more research specifically involving music as opposed to say (for example) single sign wave tests, would have to be done.

From my own anecdotal experience in blind testing (and I know it is not really exactly the same ) But---- I was involved in some testing while taking critical listening classes at Berklee Collage of music . In the single sign wave tests I was unable to detect a signal above 14.7 kHz (I was 63) . However I was consistently able to detect both a boost or cut of 16kHz in pink noise tests and music .

I was dumbfounded and no one had an definite explanation as to why, other than to theorize that perhaps even though I could not audibly hear the signal I was able to detect that it was having an effect on the frequency range I could hear.