Posted by Moving Vision on 09/20/05 10:45

In article <TYOdnUNG848qqrLeRVn-uA@comcast.com>, Ty Ford
<tyreeford@comcast.net> writes
>On Mon, 19 Sep 2005 14:26:21 -0400, Moving Vision wrote
>(in article <lSuiwHKNLwLDFwKt@movingvision.demon.co.uk>):
>
>> As for the MPEG 2 audio the key is to digitise it into the system as 48K
>> PCM before doing anything with it. Once it's PCM you can muck about with
>> it 'till the cow come home and it won't be subject to further
>> compression issues.
>
>Until you make a new DVD master and then you do what? .....Oh, right,
>compress the audio you have already reduced by 75% again. So what's left is
>25% of 25%.
>
>The first generation MPEG 2 recorded to tape on such
>> as the Z1 is quite good enough for most purposes and actually better
>> then most other compact DV's because of the broadcast standard preamps.
>
>Whose broadcast standard preamps? I'm feeding from a Sound Devices 442. You
>use the preamps on the cameras????
>
>Guess what. Go for it. However, I won't let you or anyone else go on about it
>in ignorance.
>
>Ty Ford
>
>
>
>-- Ty Ford's equipment reviews, audio samples, rates and other audiocentric
>stuff are at www.tyford.com
>


If anyone really wants to get into the issues relating to compressed HDV
audio I'd suggest going to the site that Mike Kujbida posted earlier;


http://www.sundancemediagroup.com/articles/HDV_Audio_comparisons.ht

It's a very thorough and scientific article, with various wave form
charts that pretty much covers the issues. As far as removing 75% of the
data and then removing it again is concerned, well it doesn't really
work quite as badly as that assertion might indicate. The 'missing' data
is not really equivalent to losing 75% of sound itself, it's a codec
that cleverly reduces the to the most essential bits of data. Once the
first generation of compressed sound is converted to uncompressed PCM it
is a complete sound recording, whether one feels its good enough or not
is subjective but look at all the other MPEG offerings such as MP3, DVD
and MD.

In your scenario you suggest that when the audio is recompressed back to
DVD it's the same as throwing away 75% of the bit data twice, well I
don't even think you believe that, you're just trying to be sensational
in order to support your now well know opinion on the subject.

But for those who can't be bothered to go to the web site linked above,
here's a few shamelessly copied paragraphs from the article written by
Douglas Spotted Eagle/VASST Instructor ;

HDV audio is based around a somewhat new audio compression spec known as
MPEG 1, Layer II. The bitrate is 384Kpbs. There are those that dislike
this compression format; others that have no issue with it. The crux of
the issue from the negative side, is that some audio professionals feel
that compressed audio is bad. On the other hand, many audio pro's would
suggest that compression isn't bad, it's just not as good as PCM/linear
audio. (uncompressed)
The differences that need to be examined are:
1. What can the ear hear different, if anything, in a compressed audio
format.
2. How well can the compressed format be processed in an NLE or DAW?
Another statement to consider is the sometimes-heard reseller hype that
HDV audio is CD quality.
It's not.
This doesn't mean HDV audio is bad, It just means that it's not optimal
when compared to PCM audio. PCM audio comes in various flavors of "good"
as well. Is it 12bit, 22KHz audio? Or on the other side of the question,
is it 24 bit, 192KHz? (no camera records this format)

This very brief examination was inspired by postings found on the web,
suggesting that HDV-based productions should "double-record" audio to an
MD player. If it's not an HDMD player, this suggestion is simply absurd,
as you can easily surmise from the images below. Recording to a device
like an Edirol R4, an HDMD player in PCM mode, a DAT machine, or other
linear audio device is obviously of a higher quality, but also of a
bigger production load and another button to push. (In the case of the
Edirol, the LANC controller on HDV cameras will start/stop the unit
along with the camera)

However, for general purpose/dialog audio, HDV audio has proven to be
perfectly acceptable for both big screen and broadcast use. Below are
images of a test performed using Mackie 626 speakers playing back a
popular and fairly common rock tune. (Dire Straights "I Want My MTV) The
audio is recorded from a mixer feeding an HDMD device recording in PCM,
and the same audio signal split to the HDV-Z1U HDV camera going into the
balanced inputs. The split was accomplished using a pair of Radial J3
splitters. (Jensen transformers inside, for optimal quality) The same
recording was done using the line in of an Aiwa MD player recording with
the ATRACS compression scheme, split off a bus on the mix desk. This was
done as many people have recommended using an MD player to record with
and I wanted to demonstrate the reason this is a very bad idea when
compared to HDV audio. Not only is the HDV audio superior, but also
doesn't require another device, and another button to push.

HDV audio is far from perfect. Frankly, so is linear, PCM based audio if
you're using a DV camcorder to capture the audio. However, for dialog
purposes, HDV audio is not only perfectly acceptable, it's preferable
when considering the need for carrying double audio recording gear. If
you were recording a John Prine concert, or a symphony, I'd definitely
carry a DAT machine or other high end recording equipment such as the
Edirol R4 device. And I'd suggest the same if you were recording these
sorts of events on a DV camcorder too. Double record if audio range is
critical. However, do not be concerned about the quality of HDV audio
based on anti-HDV commentary found on various websites. These charts,
and your own ears, bear out the quality of HDV audio and where it is
quite useful and amenable to audio production.
Ironically, while listener acumen isn't a test of "what's good enough,"
most people are very happy with their 128Kpbs MP3 files on their
personal music devices, and virtually no one ever negatively comments on
AC3 audio, which is generally somewhat poor. Granted, these are delivery
codecs, not acquisition codecs, but the quality of the HDV audio format
is substantially higher than either of these delivery formats.

Keep in mind that the analog to digital conversion process is
substantially more critical than the medium on which it is recorded. HDV
camcorders, particularly the Sony Z1 and A1 camcorders, have outstanding
converters. To get a better conversion, you'd need to step to an
external box such as an EchoFire box from Echo, an Apogee Rosetta (if
you've got serious cash) or one of the many M-Audio devices available.

Avoid recording to MD; you can clearly see that even the frequency range
of spoken word is dramatically impacted. Using an iRiver device is
another great option provided you're recording in PCM mode vs MP3
format.

Although converting the MPEG 1, Layer II stream into PCM audio at point
of capture doesn't reconstitute the frequencies that were compressed,
immediate conversion allows the file to maintain integrity throughout
the editing process. Reverbs, delays, etc will be smooth and clean, once
converted. Compressed audio formats shouldn't be filtered, as the
recompression for final output may have issues. Most NLE's will convert
the compressed audio to uncompressed when the audio is brought to the
timeline, or converted to PCM when an intermediary format is introduced
to the HDV video.

Record your HDV audio at appropriate levels; this will make the biggest
difference of all when shooting with HDV audio. Use good microphones,
just like you'd do with any other recording device, and get it close to
the source. Make sure all gain stages are proper, so that you're
recording clean, with robust levels, and you'll have no issues with
compressed audio at all. This holds true for any audio that is to be
compressed at any level, and for audio that isn't to be compressed. Good
audio techniques are important regardless of the recording format.

Test your HDV camcorder to know how it will handle audio in a variety of
shooting situations. You don't want to find yourself being quite
knowledgeable about the visual aspects of the camera and ignorant of the
audio aspects, do you? Sound is 70% of picture; it should encompass a
substantial bit of your camera knowledge as well. Good sound is the
absence of bad sound; great sound is pre-planned good sound.
Make your audio great.

Happy recording,

Douglas Spotted Eagle/VASST Instructor
--
John Lubran

[Back to original message]