SD 633 set up. Does this seem reasonable?

[Steven Digges]

Richard,

That is an interesting way to look at it. I understand that higher bit rate recordings are more forgiving of post production processing but to me the difference is pretty small. I would not count on it enough to change my gain staging decisions. I will use 24 bit for highly dynamic audio situations or audio critical recordings like music or performance sound tracks. For human voice interviews with a couple of lavs and a boom mic I use 48K 16bit recording to avoid extra post processing. I consider all of my audio critical and it gets sweetened in post to some degree but my goal is always to avoid unnecessary processing.

I am surprised you see enough of a difference between 16 and 24 to effect your gain staging. Out of curiosity how many DB lower do you think you can go with 24 vs 16? If that is what your saying? Also, for an average voice only interview (I know there is no such thing) what is your preferred gain setting? I like to run a little hotter than some guys are comfortable with. -12 seems to be the most common standard, I know some guys that will record as low as -20 (that's way too low for my taste), I am usually at -12 or above if it is a comfortable situation. The reason I like it hot is because I can hear everything I am recording in my cans. Low levels don't always let you hear some background noises until they are boosted in post and then it's too late. But that is just my style.

Steve

[Richard Crowley]

Originally Posted by Jim Michael

Interesting. I always took bit depth to define the number of discrete steps between the minimum and maximum levels of whatever was being measured.

And that is exactly correct. You have accurately DEFINED what bit-depth/dynamic range is.
I was stating the practical applications of USING a greater dynamic range in the Real World.

[Richard Crowley]

Originally Posted by Steven Digges

That is an interesting way to look at it. I understand that higher bit rate recordings are more forgiving of post production processing but to me the difference is pretty small.

I will assume that you meant greater sample-depth. "Rate" refers to the sampling interval, and is completely independent of the sample-depth.

I would not count on it enough to change my gain staging decisions. I will use 24 bit for highly dynamic audio situations or audio critical recordings like music or performance sound tracks.

But greater greater dynamic range and less-critical gain staging decisions is the PRIMARY reason to USE 24-bit sample depth.

For human voice interviews with a couple of lavs and a boom mic I use 48K 16bit recording to avoid extra post processing. I consider all of my audio critical and it gets sweetened in post to some degree but my goal is always to avoid unnecessary processing.

What extra post processing? Most editing software applications actually use 24-bit (or even 32-bit, or even floating-point) arithmetic internally even if you are only feeding it 16-bit data, and extracting 16-bit data. There is no perceptible "extra post processing" that I have ever noticed.

To be sure, the main disadvantage of using 24-bit (vs. 16-bit) is that larger sample-depth data simply takes more space to store. Good recorders will "pack" the data so that the 24-bit file is only 1.5x larger than the equivalent 16-bit file. But sometimes the data is stored "unpacked" and the file turns out to be 2x larger. But in these days of 32GB and even 64GB SDHC cards the size of your thumb-nail, and when I can buy 3~5 Terabyte hard drives at my membership discount warehouse, that hardly seems like a practical concern.

I am surprised you see enough of a difference between 16 and 24 to effect your gain staging. Out of curiosity how many DB lower do you think you can go with 24 vs 16? If that is what your saying?

I don't recall ever consciously calculating "well, since I am using 24-bit, I can set my levels 6 dB lower", but thinking back over it, maybe 6~9 dB lower. Or maybe zero lower, all depending on the specific situation.

Also, for an average voice only interview (I know there is no such thing) what is your preferred gain setting?

Again, I don't think about it consciously, but set it depending on the nature of the subject, how good the mic coverage is, what the ambient noise is, etc. Some subjects are very soft-spoken, and others are highly animated. Clearly you can get away with needing less headroom for the soft-spoken subject than you would for politician at a rally or a wrestler in front of the crowd.

I like to run a little hotter than some guys are comfortable with. -12 seems to be the most common standard, I know some guys that will record as low as -20 (that's way too low for my taste), I am usually at -12 or above if it is a comfortable situation. The reason I like it hot is because I can hear everything I am recording in my cans. Low levels don't always let you hear some background noises until they are boosted in post and then it's too late. But that is just my style.

But setting record levels just so that you can hear anomalies in your headphones seems like exactly the wrong motivation. If you have to boost record levels in order to hear adequately in your headphones, then get a headphone booster amp (or more sensitive headphones). I agree completely that you must be able to hear what the microphone is picking up, ESPECIALLY the unwanted noises. But setting the recording level should be independent of the monitoring setup.

[Jim Michael]

Originally Posted by Richard Crowley

And that is exactly correct. You have accurately DEFINED what bit-depth/dynamic range is.
I was stating the practical applications of USING a greater dynamic range in the Real World.

I see what you're alluding with respect to headroom.Thanks.

[Steven Digges]

Richard,

I am not disagreeing with anything you said, I am learning from it. I understand the difference between bit depth and sample rate, mostly. I certainly never thought of it in terms of gain staging though. I am trying to get my head wrapped around that. I thought of high bit depth as producing higher quality sound and more dynamic range because every time it samples it uses more bits per sample and creates a more accurate sample. Having more data per sample to work with also means it is more forgiving of adjustments made in post because of all that dynamic range and data. Am I wrong?

I don't need better headphones or more volume (I use Sennheiser HD280s). What I hear is not how I select a gain stage, it is a small part of it. But I am saying that when monitoring an anemic signal on set you might miss something compared to sitting in front of my near field monitors in the editing bay after the signal has been properly boosted. I like to record strong signals so I am hearing what I get, that is just one of the reasons why. I too, was recording in the analog days when we watched a needle bounce. Even though we are digital now, I still think old school. The acquisition phase and post production of recording video and audio are two separate events to me. I try very hard to make the original recordings as close to perfect as I can. I am not one of those guys that believes everything can be fixed in post. Garbage in, garbage out. I back up audio almost all of the time. I think I will start recording those back ups at 24 bit 48 hz more often and get more experience working with it. I must be missing something.

Steve

[Richard Crowley]

There was very recently over on another forum, yet another discussion about 16-bit vs. 24-bit and 44.1 (or 48) KHz vs. 96 or even 192KHz sample rate as a RELEASE format. (SACD, etc.) But the fact remains that almost NONE of us live in even a 16-bit world, sound-wise. In our normal, everyday environments, even if we work in a sound studio, few (any?) of us encounter even a 96dB range between the quietest and loudest sounds we hear every day. In fact, most music here in 2015 is probably played from MP3 compressed files through cheap earbuds while the listener is riding the subway to work, etc. And TV is still probably listened to on the little "speakers" in that ultra-thin LCD screen on the wall.

So even 16-bit sample-depth is more than adequate as a RELEASE format. Except maybe in the very highest-end theaters on the planet.

The problem for PRODUCTION, however, is very different. There, even just for dialog, you must be equipped to accurately capture a wide dynamic range of possible sounds from people whispering behind the hedge, to people yelling to their troupes from horseback in the middle of a period battlefield.

The issue isn't exactly "higher quality sound and more dynamic range because every time it samples it uses more bits per sample and creates a more accurate sample." The issue is the ability to accommodate the sound while preserving it from being stuck in the mud at the bottom, or having its head chopped off in clipping.

Now, there are probably production sound people with decades of experience who know their gear intimately who can set a channel gain and mix level, and even set a recording level based on what they are expecting from the scene (after reading the script and observing the setup). And they probably use high-end microphones (Schoeps, Sennheiser, DPA, et.al.) and high-end mixers (Sound Devices, Nagra, Aaton, Zaxcom, et.al.) and high-end recorders (Sound Devices, Nagra, et.al.) And to be sure, companies like SD produce gear (mixers/recorders) with mic preamps specifically designed to behave nicely even in harsh environments with a wide dynamic range of signals.

Alas, I am not in that league. I can't afford SD gear for the kinds of things that I do, so I have to learn how to use more plebeian gear, and use techniques to compensate for its limitations. And one of the great benefits is the advent of affordable 24-bit recording. It makes field recording SO much less stressful because you don't have to be as accurate in predicting peak levels.

The whole job of a production audio person is to accurately CAPTURE the dialog (and ambient "tone" and SFX and Foley, etc.) and bring it home safely. THEN it becomes the job of the post-production mixer to take those clean tracks, and properly combine them into the final mix. Certainly, the mixer doesn't leave the sound down there at the conservative levels it was recorded. Their job is to set levels, pan, effects, etc. to properly construct the final audio product.

Now, back on the set, establishing levels and monitoring signals in the mixer is rather a different thing. Certainly, you should be able to set channel gains and mix levels, etc. properly as a part of gain-staging regardless of what is downstream, whether it is a 16-bit recorder or a 24- or even 32-bit recorder, or no recorder at all (like a live event and/or a broadcast show, etc.)

I just look at it like this. If you have something fragile, and you have a box just barely big enough to hold it, even in the box you still must be very careful not to break it. But if you have a somewhat larger box, you have some extra space in there to put some padding around it to give some extra protection against hazards. That is how I see using 24-bit recording vs. 16-bit.

[Steven Digges]

Johnathan,

Congrats on your SD 633! It is an amazing piece of kit and still on my wish list. In case we went to far off track talking about data (and the concept of digital headroom, unusual) I want to say something about your post on "headroom". The limb you went out on broke ;)

As Richard said, one of the primary challenges of recording good audio is setting your gains properly. Headroom is the difference between your base db setting and the cliff of absolute zero you can not go over. I'm sure you read his later explanation and it was right on. I would encourage you to learn and master the limiters on your SD 633. Think of them as a safety net you don't want to use. When properly set they can avoid disaster but they come with a price. You do not want your gain setting so hot you are engaging your limiters often. But they can save you from clipping unexpected peaks.

I read all of your posts because you ask good questions. I usually stay out of them because Richard and Rick answer them better than I can. You might ask them to explain noise floor. You were close but not quite there.

Steve

[Steven Digges]

Thanks for taking the time to write that Richard. Your post and mine hit at the same time. Johnathan, limiters would be like padding in the box!

Steve

[Jonathan Levin]

Thanks Steven. And thank you all for your help.