Questions on audio set up

[dak]

The manual for the Korg says "48 KHz, 32 bit, Stereo".

The 32 bit is a bit of a waste of storage space (24bit is already more than you'll ever need). 48kHz stereo is fine. What would make people's mouths water is being able to get a separate track for left and right hand rather than left and right ear. If you really want to push it, a separate track for each different instrument in play. But that would cause varying data rates and could lead to hickups. If you really want to go that route, you probably should just record the MIDI, use a MIDI splitter, and then pull the individual tracks through separately through the recorder.

But all-in-all: good news.

 

[debra]

The 32 bit is a bit of a waste of storage space (24bit is already more than you'll ever need). 48kHz stereo is fine. ...

But all-in-all: good news.

I think that 32-bit means 32-bit float numbers, which contain 24 bit "precision" data and 8 bit to indicate the "exponent". It allows for more volume adjustments while processing without causing clipping (to maximum or rounding to zero).
As I do make my recordings for putting them on a CD (despite some people considering this medium obsolete) I have started to record in 44.1kHz/24bit. I'm not sure that first recording at 48kHz and then converting will be "just as good"... (It depends on whether the recorder can really use 44.1kHz internally or records at 48kHz and then converts internally.) My audio mixing program (n-Track) uses 32-bit float internally or even 64-bit float which is overkill.

 

[dak]

As I do make my recordings for putting them on a CD (despite some people considering this medium obsolete) I have started to record in 44.1kHz/24bit. I'm not sure that first recording at 48kHz and then converting will be "just as good"...

If you are doing any non-linear processing or equalization, there may be a point to use 88.2kHz instead. It gives those kinds of processing more leeway to work with, and 88.2kHz still has good downsampling to 44.1kHz. And downsampling requires anti-aliasing filtering which has to make a tradeoff between latency and quality. An audio interface will care a lot about latency; an export from an audio mixer does not need to care about latency at all and can do the downsampling in best quality.

And going down from 88.2kHz to 48kHz (for online use) may not be optimal but beats going from 44.1kHz to 48kHz.

 

[JerryPH]

The manual for the Korg says "48 KHz, 32 bit, Stereo".

Cool, that would need to be tested, of course, but it is a bit strange... very high bit rate but mediocre/average kHz rate. I'd love to know their logic. Sound quality would have been much better at a standard 24-bit/96kHz... but I am no sound technician.

 

[dak]

Cool, that would need to be tested, of course, but it is a bit strange... very high bit rate

That is not a bit rate. It is a bit depth. And as Paul said, it is the floating point version of 24bit depth, avoiding clipping artifacts from processing.

but mediocre/average kHz rate. I'd love to know their logic. Sound quality would have been much better at a standard 24-bit/96kHz... but I am no sound technician.

Well, I do have a degree in signal processing. 96kHz sample frequency tends to be a better match for intermediate processing (as it reduces the likelihood for several processing artifacts and allows the use of minimum-phase IIR filters with low frequency warping), particularly starting from analog sources. As an end and storage format, 48kHz is just fine. Good 48kHz D/A converters introduce more latency than good 96kHz D/A converters, but the lag introduced in that manner is less than the dead time for starting a reed.

So in my book, 48kHz here is perfectly fine. Indeed, the high-quality Opus codec for lossy audio compression (used by video platforms and high-level video formats like newer webm standards) refuses to go anywhere beyond 48kHz because its standard bearers do not consider frequencies above 24kHz as audible.

People who claim to hear the difference between 96kHz sample frequency and 48kHz are much more likely to actually hear the difference between the artifacts of their specific audio hardware at 96kHz sample frequency compared to 48kHz sample frequency.

48kHz sampling frequency is just fine for processed audio. I wouldn't use it for sample synthesis of analog-modeling synthesizers (for example) because operations like squaring a signal cause significant foldaround of frequencies and higher oversampling helps in keeping most artifacts out of the audible range. The results of squaring an 48kHz sampled sine signal and of upsampling that signal to 192kHz (with suitable anti-aliasing filtering), squaring then, and then downsampling (with suitable anti-aliasing filtering) to 48kHz are different. And both are still different from just squaring the continuous analog signal, but the 192kHz sampling shows less difference.

You just have to know when higher sample rates count, and when not. And in the case of the Korg Fisa's canned sampled instruments, they count not.

 

[KeysFla]

The 32 bit is a bit of a waste of storage space (24bit is already more than you'll ever need). 48kHz stereo is fine. What would make people's mouths water is being able to get a separate track for left and right hand rather than left and right ear. If you really want to push it, a separate track for each different instrument in play. But that would cause varying data rates and could lead to hickups. If you really want to go that route, you probably should just record the MIDI, use a MIDI splitter, and then pull the individual tracks through separately through the recorder.

But all-in-all: good news.

That would be the next generation. A midi sequencer digital accordion?

 

[KeysFla]

The iPhone has auto gain, I believe. I never had to worry about doing a video and getting clipping. When connected to an audio interface, I get good sounding video….

 

[JerryPH]

Auto gain has it’s good side just don’t play a song with dynamics, loud parts get compressed, quiet parts get raised and adds noise. Great for voice, mediocre dynamic range for music… but this is NOT a dis on the comment, it’s a fair indicator that different people see things differently.

Also explains why I don’t normally capture on my android outside of pro mode where auto gain is disabled. Cannot do this on any iPhone.

 

[breezybellows]

Auto gain has it’s good side just don’t play a song with dynamics, loud parts get compressed, quiet parts get raised and adds noise. Great for voice, mediocre dynamic range for music… but this is NOT a dis on the comment, it’s a fair indicator that different people see things differently.

Also explains why I don’t normally capture on my android outside of pro mode where auto gain is disabled. Cannot do this on any iPhone.

What's pro mode?

 

[JerryPH]

What's pro mode?

There is a mode on my android phone that removes all automation (chocie of auto/manual exposure, choice of audio source, capture from external source in stereo and choice of auto/manual gain control) on my old Samsung S21 Ultra cell phone.


In the fist pic you can see that all options for video are adjustable manually if desired, or left in auto. In the 2nd pic we see that it captures 2 channels (stereo), and you can choose audio source and at what gain levels. 0 means no change over what comes in, which if you properly gain staged the audio interface, will be at whatever levels you set that to.

I really still prefer a "proper camera" for public videos, but it's fun to know that I can walk this path if I want to.

Apple has done a ton of audio/video improvements on their iPhone 13-16 in the last few years, unfortunately, they insist on maintaining control over anything audio and don't offer that choice at this time. I was close to going iPhone 17 (love the prores raw video option), but I have a few back door connections with them, and was told that there is no change in the audio department... so I am sticking with what I have for now.

 

[KeysFla]

Auto gain has it’s good side just don’t play a song with dynamics, loud parts get compressed, quiet parts get raised and adds noise. Great for voice, mediocre dynamic range for music… but this is NOT a dis on the comment, it’s a fair indicator that different people see things differently.

Also explains why I don’t normally capture on my android outside of pro mode where auto gain is disabled. Cannot do this on any iPhone.

iPhone gives me the quality I need. Not sure if it has Auto gain? but I do not want to go through a whole recording and find I have to do over because of clipping. I just like simplicity. Quick recording is all I need.
Although, if the input to the audio interface is good, then it should be ok without any adjustment to the camera…

 

[kep]

I really still prefer a "proper camera" for public videos, but it's fun to know that I can walk this path if I want to.

Apple has done a ton of audio/video improvements on their iPhone 13-16 in the last few years, unfortunately, they insist on maintaining control over anything audio and don't offer that choice at this time. I was close to going iPhone 17 (love the prores raw video option), but I have a few back door connections with them, and was told that there is no change in the audio department... so I am sticking with what I have for now.

You can check the BlackmagicCamera app for iPhones, it has some tweaks you would like (apparently not the level directly, but a lot of setting to tweak):

 

[JerryPH]

You can check the BlackmagicCamera app for iPhones, it has some tweaks you would like (apparently not the level directly, but a lot of setting to tweak):

Thanks for bringing it up... almost forgot about it. Yes it has SOME great options, but doesn't disable auto gain on Apple iphones (works well on my ipad and S21 Ultra but also not ablle to disable autogain on either), but the video options are where this app truly shines! If I had no "pro mode" to lean on, this is what I would be using on my cellphone.