Hi -

Reading a recent sounding off in SOS made me realize I wasn't really quite sure about the whys and wherefores of 16 bit vs 24 bit recording. I found this SOS article by Hugh: http://www.soundonsound.com/sos/jun08/articles/qa0608_2.htm which I understood fairly well, except the part I've bolded:

Quote:

---

...Having said all that, you are right in that for most people recording at home, the recording noise floor is almost always defined by the ambient room noise rather than the digital system noise. Therefore, you can usually work with reasonable headroom margins at 16-bit word lengths while maintaining the ambient noise floor comfortably above the system noise floor. In other words, the dynamic range of the source recordings is typically far less than the capability of a 16-bit recording system. Where that changes is if you record electrical instruments such as guitars and keyboards via DIs. In these cases, the noise floor is often far lower than that possible when recording with microphones in a home studio, so the dynamic range of a 16-bit system may start to become a limitation.

---

I think I understand why, when using DIs at home, the noise floor is often far lower than that possible when recording with microphones (because the noise floor is the very low electronic noise floor, not the higher ambient room noise floor?), but why would that make " ...the dynamic range of a 16-bit system ... to become (more of) a limitation"? Using a mic, Hugh described "the dynamic range of the source recordings is typically far less than the capability of a 16-bit recording system". Does that change somehow with DI'd signals?

Thanks much for any thoughts!

--------------------
Alexis -Cubase 6.5.0/SX3.1.1.944, XP SP2, 4GB RAM (1GB not accessible, but used just to balance the computer so it doesn't tip over); Delta 66 in Omni i/O Studio; Motif8; UAD-1

For the microphone case the dynamic range is likely to be the highest sound level at the microphone minus the room noise, for a DI it it the highest level minus the electrical noise.

Thus in a city centre spare bedroom, you might have say 120dB from the guitar amp at its mic, with a room noise floor at the same mic of say 50db, giving a dynamic range of 70dB, well within the capability of a 16 bit recording chain even with say 15dB of headroom left.

For a DI in the same situation, the electrical noise floor might be over 100dB below the loudest peak, so you are looking for rather more range then a 16 bit chain can deliver.

Of course on playback you are still limited by what the room noise floor will let you hear!

Regards, Dan.

--------------------
Audiophiles use phono leads because they are unbalanced people!

Quote dmills:

---

For the microphone case the dynamic range is likely to be the highest sound level at the microphone minus the room noise, for a DI it it the highest level minus the electrical noise.

Thus in a city centre spare bedroom, you might have say 120dB from the guitar amp at its mic, with a room noise floor at the same mic of say 50db, giving a dynamic range of 70dB, well within the capability of a 16 bit recording chain even with say 15dB of headroom left.

For a DI in the same situation, the electrical noise floor might be over 100dB below the loudest peak, so you are looking for rather more range then a 16 bit chain can deliver.

Of course on playback you are still limited by what the room noise floor will let you hear!

Regards, Dan.

---

Thank you Dan, for a crystal clear (24-bit!) explanation!

I know the recommendation is to record with more headroom on a 24-bit system than a 16-bit system. Is that because the [signal]-to-[noise floor] ratio at, e.g., -12dBFS, is much higher with the 24-bit system? (And therefore there's less risk of ruining a take by going "over" on a loud passage)?

--------------------
Alexis -Cubase 6.5.0/SX3.1.1.944, XP SP2, 4GB RAM (1GB not accessible, but used just to balance the computer so it doesn't tip over); Delta 66 in Omni i/O Studio; Motif8; UAD-1

Yea, a 24 bit system (usually really about 19 - 20 bits effective) has the dynamic range to allow you to leave ample headroom for unexpected peaks without seriously compromising the overall noise level.

With a 16 bit system your margin is smaller, at least if you don't know (and you never really do in advance) what the peak level is going to be.

Thus a 16 bit system takes somewhat more care with levels then a 24 bit one does, sufficient reason to use 24 bit for capture and mixing.

Interestingly once you have completed all the capture and mixing, your need for headroom goes away and as a DISTRIBUTION format 16 bits is entirely satisfactory.

Regards, Dan.

--------------------
Audiophiles use phono leads because they are unbalanced people!

I'm feeling strangely redundant!

Thanks for delivering such nice clear answers, Dan.

Hugh

--------------------
Technical Editor, Sound On Sound

Quote Hugh Robjohns:

---

I'm feeling strangely redundant! ...

Hugh

---

Does she mind?

--------------------
Alexis -Cubase 6.5.0/SX3.1.1.944, XP SP2, 4GB RAM (1GB not accessible, but used just to balance the computer so it doesn't tip over); Delta 66 in Omni i/O Studio; Motif8; UAD-1