What sample rate do you record at by default?

[ComposerLDG]

Hi all,

I usually record at 44.1, but a library I work with wants their .wav masters at 48. Would it be better to just record at 48 for everything by default?

Thank,

-Loren

[andygabrys]

Up to you.

Hardly anybody can hear the difference.

Since I do primarily music for TV / film and the standard is 48 kHz I record at 48 khz.

If I ever need to supply 44.1 kHz I convert down for those occasions (and again, hardly anyone can hear the difference in a blind listening test).

[stevebarden]

Presuming your audio is intended for video production you should always deliver 48khz audio. Video utilizes 48khz audio so if you deliver 44.1 then the audio will have to be re-sampled and you risk introducing artifacts.

When you record your audio in your DAW you should record 24-bit audio, even if you ultimately deliver 16-bit audio to the client. Most of my clients have always required 16-bit, however, I recently had a client request 24-bit.

Ideally you should render 16/44.1, 16/48, 24/44.1, and 24/48 versions just to have yourself covered. You may not need those versions but it's good to be prepared.

I honestly don't know why a client would request 44.1 audio unless it was intended for audio-only products, not video.

[ComposerLDG]

Andy and Steve, thanks for the info! I knew you two would know.

Thanks again,

-Loren

[Casey H]

I honestly don't know why a client would request 44.1 audio unless it was intended for audio-only products, not video.

I have some libraries that request 44.1K, others that want 48K, and some that don't care. Most are OK with 16 bit, occasionally one asks for 24 bit but I don't get that request too often.

My opinion is record at 48K 24 bit if you can. That way your source material is at the highest resolution anyone will ask for (nowadays). From your 48K 24 bit you can down convert to 48K/16, 44.1K/24, or 44.1/16 as needed for different libraries.

Best
Casey

[Kolstad]

I use 48k as default. I record a lot of acoustic guitar, and did a test if I could hear the difference between 44.1k and 48k on acoustic guitar, and to me the differenece was not subtle on my setup. Theres more depth and overtones in the 48k recording. If you dont do video nor acoustic instruments, Im not sure if the differences can be heard, though.

In session, I record software instruments in 96k, as I've found the same to be true in terms of depth and overtones for them in the digital domain. Software just sounds better at higher resolutions, imo. But it is somewhat idealistic, as an effort to capture the absolutely most pristine tonality that I can with the equipment I use. For many genres like blues, indie pop/rock and such, it absolutely doesn't matter, and most often there you dont want a pristine sound.

[Len911]

Like others have said the difference between 44.1 and 48 is negligible.

In this instance it's more about convenience. Especially if you are using audio loops that are at a different sampling rate than the project sampling rate and if you have to manually convert them.

[andygabrys]

I use 48k as default. I record a lot of acoustic guitar, and did a test if I could hear the difference between 44.1k and 48k on acoustic guitar, and to me the differenece was not subtle on my setup. Theres more depth and overtones in the 48k recording. If you dont do video nor acoustic instruments, Im not sure if the differences can be heard, though.

In session, I record software instruments in 96k, as I've found the same to be true in terms of depth and overtones for them in the digital domain. Software just sounds better at higher resolutions, imo. But it is somewhat idealistic, as an effort to capture the absolutely most pristine tonality that I can with the equipment I use. For many genres like blues, indie pop/rock and such, it absolutely doesn't matter, and most often there you dont want a pristine sound.

Interesting but unless the software instruments were sampled at 96khz you are not capturing anything beyond really big file sizes.

Very difficult to do but have you set up a double blind type test to make sure your bias of knowing what you are listening for isn’t making up your mind for you? No offense intended but you are talking about a maximum 48/44.1 difference (~8%).

That’s less of a difference between 16 bit 44.1k wav and a 192 kB/s MP3. And the difference between those is pretty hard to discern - and I have never been able to tell definitively in a blind test the difference between those. Not saying that your ears aren’t better than mine (hopefully it’s true - don’t play in rock bands kids) but it’s honestly a pretty small difference.

[Casey H]

I have to agree with Andy. Few humans on Earth can hear the difference between 44.1K and 48K and in a DOUBLE BLIND test, it would be obvious.

And recording at 96K is a massive waste of disk space. Any small improvement in audio quality is so small to the human ear that the tradeoff of using 2x the disk space is not worth it.

Most people can barely tell a 192K mp3 from a 44.1K/16 wav!! The human ear is just not that good. (That was an example, I know there is a difference there)

Casey

[Kolstad]

I use 48k as default. I record a lot of acoustic guitar, and did a test if I could hear the difference between 44.1k and 48k on acoustic guitar, and to me the differenece was not subtle on my setup. Theres more depth and overtones in the 48k recording. If you dont do video nor acoustic instruments, Im not sure if the differences can be heard, though.

In session, I record software instruments in 96k, as I've found the same to be true in terms of depth and overtones for them in the digital domain. Software just sounds better at higher resolutions, imo. But it is somewhat idealistic, as an effort to capture the absolutely most pristine tonality that I can with the equipment I use. For many genres like blues, indie pop/rock and such, it absolutely doesn't matter, and most often there you dont want a pristine sound.

Interesting but unless the software instruments were sampled at 96khz you are not capturing anything beyond really big file sizes.

Very difficult to do but have you set up a double blind type test to make sure your bias of knowing what you are listening for isn’t making up your mind for you? No offense intended but you are talking about a maximum 48/44.1 difference (~8%).

That’s less of a difference between 16 bit 44.1k wav and a 192 kB/s MP3. And the difference between those is pretty hard to discern - and I have never been able to tell definitively in a blind test the difference between those. Not saying that your ears aren’t better than mine (hopefully it’s true - don’t play in rock bands kids) but it’s honestly a pretty small difference.

You are probably right, Andy (and Casey), and I have to question what I'm hearing and the environment Im hearing it in. What I think, though, is to have the best session files I can, if I decide to remix (or degrade) the recordings at a later stage. I bounce all midi to audio for the same reason. Everything goes so fast these days, so it's an old school approach in the digital domain, to go for the best sound possible (at least to my artistic ears) and archive it properly.

I just don't rely on technical theory only for artistic measurement, so I also like to go with my ears as well. I guess that kind of idealism (which is absent from many other domains in life these days) contributes to the fun for me, fwiw.