Need advice on audio solutions

[Seth Williams]

So first the basics: I'm going to be making an indy drama film using a Sony HVR-Z1U camera. I'm looking to spend as little as possible, but I'm willing to spend as much as it takes to get adequate results.

For a primary microphone, I'm told that Sennheiser is probably the best way to go... So I'll stick with that suggestion. The specific product I was recommended was the Sennheiser 416, but as I did research I also discovered the 418. What exactly is the difference between these (I know the Sennheiser has a "figure 8" thing, but that doesn't mean much to me), and which do you think would be better for my purposes?

Also, The Z1 camera has XLR inputs... Can all audio go directly into the camera and yield good results, or would I still have to use a separate DAT and/or mixer?

Thanks in advance!

[Glenn Chan]

A quick Google search shows the 418 is a stereo shotgun microphone, while the 416 is a mono shotgun microphone. The 418 has two extra microphones on the side for the mid-side stereo setup. For your purposes, it's unlikely you'll need a stereo microphone.

2- You could save money by going with something cheaper than a top-of-the-line microphone. There's lots of threads on which shotgun microphone to get... the Audio Technica 897 and 4073a are popular recommendations, although there are also other good microphones. There are also clips comparing the various microphones against each other. If you do a search you'll find most of them.

Some of the are hosted over at
http://www.dvfreelancer.com/forum/vi...8f681b7f92b896

3- How are you going to mount/use the microphone?

The best sounding setup would be a boom-mounted microphone. You need a boom and shockmount. If outdoors you'll need a windscreen too. You should budget for these items.

You may also want a field mixer and closed/in-ear headphones.

4- As an alternative, you could just find a sound recordist with equipment. A good one will come with experience, which'll really help in getting good sound. You can look for people like this on mandy.com, local film/video associations, etc.

5- If you want to do things yourself, I suggest you get Jay Rose's book "Producing Great Sound for Digital Video". See dplay.com for how to get it for $30.
http://www.dplay.com/book/index.html

[Ty Ford]

If you'll be shooting inside spaces like rooms in normal houses or offices, you don't want a shotgun mic at all. You want a hypercardioid. Best is Schoeps cmc641. AT 4053 costs less, is less. Sennheiser makes the mkh 50 or 60 that's also a very good hyper.

As to the camera audio. The audio recorded on HDV is data compressed MPEG. It is not linear PCM. There have been cautionary statements about using that camera's audio for production. Word is, if you want great audio, you must double record.

I haven't done the tests, I have read the comments.

Regards,

Ty Ford

[Douglas Spotted Eagle]

if you are doing dialog, and you know how to set levels, the Z1's audio is just fine. You'll not notice the compression at all. It's extreme range with tails that you'll notice it on, ie; cymbals that have long decays. You'll hear aliasing as the decay gets near silent.
But the rumors of it not being usable are more BS from folks who simply don't like compression, or don't know how to set levels, one or the other. We did a blind comparison of a Z1 vs a PD170, couldn't tell which was which. But set up of level is very important. You can't recover as easily from the MPEG compression like you semi-can with PCM audio.

[Ty Ford]

About the audio compression.

My concern is based on what happens to the audio AFTER the compressed audio is captured. There are many places downstream of the original capture that MAY also use some sort of data compression. If that will never happen with your product, you should be OK.

BUT....When you data compress audio that has already been data compressed, (by the same or another compression algorithm) you run a very high risk of audible degradation.

I don't know what your plans for postproduction and distribution are, but if they include formats which apply more data compression, both your video and audio may suffer.

Here's what I DO know about what happens after one compression pass to 128kbps mono audio. (BTW, about 750 kbps is the size of a stream of uncompressed 44.1 KHz, 16-bit, mono) After one stage of this compression, EQ does not work the same way. You reach for some 6 kHz to brighten things up and it just doesn't sound right.

In a quote from a story I wrote recently which was a compendium of video editors comments, one said, "I'm seeing a 50/50 mix between DVD and VHS for client copies, but not for broadcast. The DVD looks better but doesn't have time code control. Not all DVD recorders and players are fully compatible. There are some DVD/mpeg compression problems that don't look so great on NTSC; not that that seems to stop anyone from broadcasting compression. You see it all the time on now."

Steve Mullens who writes for video systems has also mentioned in his articles that, in regards to the HDV format, if you really want good audio, you should record it on a separate system.

Will you notice problems? I don't know. I would probably hear the differences before a lot of other people because audio is what I do and I know what to listen for.

Regards,

Ty Ford

[Douglas Spotted Eagle]

I can't begin to agree with Mullen's comment. I read that too. It's ridiculous.
In nearly every HDV editing situation, one of 3 things will occur when you capture:
1. You convert to an intermediary. (DI) The audio is then converted to PCM format.

2. You transcode via HD/SDI and capture to whatever codec or uncompressed stream. Audio is still converted to a much more usable format.

3. Your NLE will demux the stream and give you a PCM or other uncompressed conversion format.

At no time in the normal editing process, unless you specify it to do so, will the audio be recompressed.

As I've said here and in many other spaces, the ONLY time you need a second audio capture source is when you're doing music or highly detailed, high frequency audio captures/events that have decay. MPEG 1/Layer II audio is exceptionally good in the scheme of things at bitrates of 384Kbps. I've done a wide variety of tests with this format using all sorts of plugs to check decay times, S/R release noise, dropped long verbs from WAVES on it to check to see how it resamples/aliases, and it's very, very good. But it does need to be converted before processing, and every NLE system I'm aware of does exactly that, excepting Pinnacle Liquid Edition.

[Ty Ford]

That's OK Doug. I like you even though we disagree on this.

Ty

[Douglas Spotted Eagle]

Ty, how can you disagree if you've not tested it? I'd respect your opinion if I knew you'd been working with the audio. On page 38 of my HDV book, the waveforms are there from both formats, PCM and MPEG. It's pretty clear.

I know, I keep hearing about the math of it. It bugs me to read those positions because they simply ignore the technology in favor of math based on what we've known in the past. Encode/decode is the magic, not compression. If you base things on numbers and past knowledge alone, then of course you/Mullen are right. If Boeing had based airplane design on past knowledge and numbers we'd not be flying today.

Maybe there is a better way of monitoring this than through a set of Hothouse 4's in a very tightly designed room. Maybe someone's got a much better room that they can hear better in than I can. But it sounds great.

Have you actually tested properly recorded audio from the Z1? I've got over 40 hours of tape in a variety of circumstances, music, dialog, football games, wild audio in a helicopter.... And it all sounds great.

Yeah, you pushed a button because I'm pretty tired of reading all the parroting on this format that is happening.
I like you too, Ty...even if you are repeating what someone else has wrongfully assumed.

[Ty Ford]

<<<-- Originally posted by Douglas Spotted Eagle : I've got over 40 hours of tape in a variety of circumstances, music, dialog, football games, wild audio in a helicopter.... And it all sounds great.
. -->>>

have you sent it through other compression algorithms which may typically be encountered downstream?

Ty

[Douglas Spotted Eagle]

Yes, I have. I've rendered to wma, MP3, AC3, QT at various bitrates, OGG, REAL, Scott Studio, and B-wave, these were all done post DI and editing.
I've already delivered HDV footage in final HD for worldwide distribution, and delivered in AC3 format for audio, and delivered in multiple languages.
Vegas has a batch renderer, and I've got a test file that I regularly use for testing this stuff both for Beta and new products.
It's a very fast chase to test this stuff with the batching tools.

[Ty Ford]

allow me to clarify..


multiple algorithms, as in more than one for the same piece of audio in series so that different algorithms are applied to the same piece of audio, one after the other. During that process, have you attempted and=y DSP, editing, EQ, stuff like that?

BTW, as you may know, I'm not alone in my concern here. Many others are equally concerned that this is a slippery slope.

Regards,

Ty Ford