A Few Words About A few words about...™ Casablanca (70th Anniversary) -- in Blu-ray (1 Viewer)

[haineshisway]

I've said this before elsewhere, but it would be fun to sit people in a room and play them some compressed and uncompressed tracks side by side, without telling them which is which. Lossless and lossy are words and when people see those words they immediately "hear" the difference. But my guess is that in a blind test they'd guess wrong more than they'd guess right, especially with older tracks. And it's not just mono tracks from pre-1950s films, it's mono tracks right up through most of the 1970s. In fact, modern sound readers do poorly with those tracks and I have some direct experience with that aspect. But in today's world people read, and then think they are experts, without having any real direct knowledge about sound or image. Whether one agrees with Mr. Harris or not, he's been in the trenches and knows whereof he speaks, while others posting have not been in the trenches, have never sat in a transfer room or handled film and audio - this is not a slam or slap-down of anyone, but the Internet has made experts of everyone.
My question always is do people really think, with an older film, they'd be able to tell the difference in a blind test, audio-wise? I know I probably wouldn't and I HAVE been in the trenches, audio-wise, and have really expert ears.

 

[Robert Harris]

Originally Posted by haineshisway /t/319152/a-few-words-about-casablanca-70th-anniversary-in-blu-ray/60#post_3906221
I've said this before elsewhere, but it would be fun to sit people in a room and play them some compressed and uncompressed tracks side by side, without telling them which is which. Lossless and lossy are words and when people see those words they immediately "hear" the difference. But my guess is that in a blind test they'd guess wrong more than they'd guess right, especially with older tracks. And it's not just mono tracks from pre-1950s films, it's mono tracks right up through most of the 1970s. In fact, modern sound readers do poorly with those tracks and I have some direct experience with that aspect. But in today's world people read, and then think they are experts, without having any real direct knowledge about sound or image. Whether one agrees with Mr. Harris or not, he's been in the trenches and knows whereof he speaks, while others posting have not been in the trenches, have never sat in a transfer room or handled film and audio - this is not a slam or slap-down of anyone, but the Internet has made experts of everyone.
My question always is do people really think, with an older film, they'd be able to tell the difference in a blind test, audio-wise? I know I probably wouldn't and I HAVE been in the trenches, audio-wise, and have really expert ears.

A relatively easy answer. On the average system, I would not!

RAH

 

[Yorkshire]

Yes.  Going back to an original element is sometimes problematic.  And you deal with it, shot by shot, and frame by frame.
An example:  In GF2, during the murder of Fanucci, in a 4k scan of the camera original, tiny wires attached to squibs in the actor's face became visible.  These had to be digitally removed.  They were not visible in final prints.
RAH
 

Cheers Robert, that's very much appreciated.
Steve W

 

[Rob W]

I've said this before elsewhere, but it would be fun to sit people in a room and play them some compressed and uncompressed tracks side by side, without telling them which is which. Lossless and lossy are words and when people see those words they immediately "hear" the difference. But my guess is that in a blind test they'd guess wrong more than they'd guess right, especially with older tracks.

I attended just such a session sponsored by Toshiba in Toronto back when they were making the rounds looking to launch HD-DVD. In fact, it may have been this very forum where I found out about it. At any rate, they ran comparisons of 1080i vs 1080p as well as the audio blind comparison tests you describe. In an audience of approximately 40 'videophiles', there was nothing close to a clear consensus on which formats were which for both audio and video.
And no, I don't remember how many I got 'right', but I certainly didn't ace the test !

 

[haineshisway]

I attended just such a session sponsored by Toshiba in Toronto back when they were making the rounds looking to launch HD-DVD. In fact, it may have been this very forum where I found out about it. At any rate, they ran comparisons of 1080i vs 1080p as well as the audio blind comparison tests you describe. In an audience of approximately 40 'videophiles', there was nothing close to a clear consensus on which formats were which for both audio and video.
And no, I don't remember how many I got 'right', but I certainly didn't ace the test !

Well, there you are

 

[Mike Frezon]

Such anecdotal evidence (along with speculation of how such tests would turn out), while interesting for discussion's sake, is really worthless when it comes to explaining the reasons as to why lossless shouldn't be a standard on all Blu-way releases.

 

Maybe the thinking is that lossless can reveal too many imperfections in the sound, especially on older tracks. Compression can hide some of those imperfections which weren't meant to be heard in the first place.

 

[TonyD]

It does.  But it doesn't really matter that it does.  There is little to be gained.
RAH

The reality is, that precisely the opposite is in play.  Uncompressed audio, which can yield superb results for modern tracks, tends to reveal too many of the deficiencies of older tracks, which can sound superb in a more analogue world.  The perfect and absolute reproduction of zeros and ones is not always the best way to go.
Please keep in mind that older tracks were not created to be heard 1:1 from a master.  Quite the same as classic films were not meant to be viewed from their camera negatives, which in many cases reveal far too much within the image, that would be hidden in original multiple generations.
RAH

Mike,
This is the can or worms that was hashed out when The African Queen was released without an uncompressed soundtrack. Fortunately the studios have listened to the demands of consumers and now just about every new catalog release does have uncompressed sound. I don't see why we need to bicker about sound in this thread because Warner has done the right thing with this new release.

Mark, Robert brought it up. Someone asked if it was there and a "Yes" would have worked well enough but Robert added in the kicker with a so what, it doesn't matter.
It matters to me and with my simple $2000 ten year old speakers and $1500 five year old Denon I think I can tell the difference and would prefer that the studio didnt remove that chance that I might not be able to by not including lossless.
Also from what I can see in this discussion video gets to be "restored" because that's is more important so audio doesn't deserve to get a restoration.
If This movie or any other "older" film has a less than quality audio track tough tooties because it's just sound.
What Robert did say was there is little to be gained.
For me, every little bit counts.

 

[Yorkshire]

I think it's really to do with two things that you might be hearing - the sound that was recorded and the defects in the recording system.
To take this to an extreme, take a wax cylinder recording; I suspect most of us have heard one of these. There's the sound Edison made when he read out Mary Had a Little Lamb, and then there's the noise of the cylinder, which almost drowns out the recording of the voice.
The actual sound it was intended to record - the voice - is very dull, even without the background crackle. Use a lossless track rather than a low bit-rate mp3 track and there'd be very little difference in what you hear of the voice. The better format might be able to pick up subtle nuances in sound that a lossy track can't, but what good is that if those subtle nuances just aren't there.
However, and unfortunately, the lossless track will pick up every last subtle nuance of the crackles.
TonyD, you say every little bit matters, and I take your point. But when the sound you want to hear is improved hardly at all (or not at all), but the sounds you don't want to hear are 'improved' greatly, then surely that's not a good thing.
A good vanilla dts or Dolby Digital track can sound quite wonderful. Whilst it's great to strive for perfection, I think that when we're 99% of the way there we might want to consider just how much angst we want to generate worrying about the remaining 1% instead of just enjoying the film.
That's not being complacent about quality, it's just putting small issues into perspective.
Steve W

 

[Robert Harris]

Originally Posted by TonyD /t/319152/a-few-words-about-casablanca-70th-anniversary-in-blu-ray/60#post_3906506
Also from what I can see in this discussion video gets to be "restored" because that's is more important so audio doesn't deserve to get a restoration.
If This movie or any other "older" film has a less than quality audio track tough tooties because it's just sound.
What Robert did say was there is little to be gained.
For me, every little bit counts.

For product of a certain age and dependent upon extant elements, both image and audio may need restorative efforts. For audio, post facilities such as Chace and several smaller shops with very specific capabilities can do a superb job of this. But as with image, every nuance needs to be carefully honed, less parts of that aural image be lost. Digital clean-up of a track is far less expensive than dealing with image, when image is necessary, but it needs to be done without harming anything.

Far too often in on-line reviews, I see the words "and they've done a beautiful job of restoring the visuals of this almost 20 year old film..." And I wonder what there is to "restore." Restoration is a very unique phrase that is more than oft, mis-used.

If one is returning to the original elements, be they image or audio, for the average 30 year old production, there is little to be done. The only difference that comes into play in preparing a film for Blu-ray is the single decision by the asset protection exec as to what they are desirous of accomplishing.

Is the task to make a pretty Blu-ray, or back up the assets while that Blu-ray has been requested by home video?

If it's the former, you begin with a a timed interpositive, scan it at 4k and immediately down-rez to 2k or HD for any simple clean-up that needs to be performed. Universal went beyond this for Out of Africa, as their goal was NOT simply a Blu-ray release, but the production of elements toward full asset protection. To use that example, for audio, the studio used a 5.1 re-mix produced in 2002, as it could not be bettered. To complete that specific example, Out of Africa was scanned at 6k, with all work being performed at 4.

Go back to the 1930s through the early 1950s, and, presuming an original negative survives, image will need far more love and attention, and yet may still not need restoration. With current abilities to harvest the most complete audio image possible from an optical dupe element, or audio stems, should they survive, the asset protection route would be to perform a digital clean-up of those elements, while working toward a goal of not removing anything that should be heard. It can get sticky here, as one can almost get into a CSI mode, listening, and occasionally finding recorded information that may be in the background, and bringing it to the fore for modern reproduction. Do we want this? Occasionally, but for the most part, doubtful, as the job is to reproduce what was originally heard in theatres.

And there would have been a huge difference between what was heard in theatres and heard in the studio dubbing rooms.

There is a third area that also exists. These are films for which no original negative survives or its whereabouts unknown. These productions go up into the 1970s. If original mag tracks have deteriorated or are lost, one would go to the next best thing, which in some cases, may be a positive optical track print, or worst case, a used 35mm print. For the 1930s or '40s, one might find themselves dropping to a 16mm element for partial audio.

For any modern film, let's say The Descendants, which just arrived on Blu-ray, you'll find a production shot on S35/3, with an image that was scanned and processed at 2k, and a fully digital track. Blu-ray has the capability of reproducing a film of this type with great quality. The 2k scan is ported to HD, and the digital audio track will move over as well, 1:1.

And for this, we would want our Blu-ray to have uncompressed audio, whether we might be able to hear a difference on our personal systems or not.

But take Dracula (1931), or Rear Window (1954) or even the tracks that we created as replacement for junked originals for Vertigo (1958/1998), and it will be found that if the tracks are to be presented in digital form uncompressed vs. compressed, that additional work may be needed to remove certain attributes from the audio image. In the best of analogue worlds, a 70mm presentation will allow the finest reproduction of audio possible. All 70mm prints of Vertigo, except one, were produced with a DTS code track. And those prints sounded as full and rich and complete, when played back on theatrical systems, as the original elements allowed. But something was discovered.

The analogue print sounded better.

And not because there was a problem with the DTS track, but rather, because the DTS track reproduced the audio so perfectly, that it exposed too many of the problems associated with the original elements. Those problems were covered by the simple generational loss of going from the 6 track full-coat master, to the 6-track print master to the print. The final result in the analogue world was a slightly richer, smoother, and more rewarding experience.

Compression works in a similar way. But the word compression has gotten a "bad rap" because of the marketing sizzle of Blu-ray's ability to reproduce uncompressed audio. The reality of the situation is, and I'm guessing here, that 75% or more of the uncompressed tracks found on Blu-ray discs would be indistinguishable from properly compressed counterparts on even high quality reproducing systems in the home.

On the best systems, uncompressed tracks can give us better image placement, and differentiation of the various tonalities of the track, ie. the ability to pick the voice of someone shouting over a huge music and effects tracks, or the tiniest bits of audio energy that might be found in an extremely quiet scene.

But generally, for films pre-1960 or so, with the exception of certain large format multi-track recordings, the difference between compressed and uncompressed will not be heard. Go back another 30 years, and the battle becomes not exposing things not meant to be heard.

Should uncompressed be the standard for Blu-ray?

Certainly.

But in most cases, it doesn't matter, and when it does will only be heard on high end systems.

Those whose viewing environment involves their monitor's built-in speakers need not apply.

RAH

 

[Scott Calvert]

Maybe the thinking is that lossless can reveal too many imperfections in the sound, especially on older tracks. Compression can hide some of those imperfections which weren't meant to be heard in the first place.

The kindest thing I can possibly say about that argument is that it's specious. If I weren't concerned with being kind I would say it's complete nonsense.

 

[Traveling Matt]

Robert, does the home video world distinguish the difference between "uncompressed" and "lossless?" This thread doesn't seem to.

 

[DavidJ]

The kindest thing I can possibly say about that argument is that it's specious. If I weren't concerned with being kind I would say it's complete nonsense.

That is not at all a specious argument and it is certainly not nonsense. You may not agree with the argument and you don't have to listen to what others in this thread are saying, but to me that statement is taking it too far.
Steve and RAH have both made recent post that provide a lot of good context and information regarding this issue.

 

[Robert Harris]

Originally Posted by Traveling Matt /t/319152/a-few-words-about-casablanca-70th-anniversary-in-blu-ray/60#post_3906615
Robert, does the home video world distinguish the difference between "uncompressed" and "lossless?" This thread doesn't seem to.

They should be two different things, with the same final result.

If an algorithm is truly lossless, it should be able to equal uncompressed data 1:1.

Probably the thornier question, for which I don't have the answer, might be, does all hardware / software created with the purpose of decoding compressed "lossless" data, so that in the end it is truly "lossless" do so? This is above my pay grade.

For this discussion, I refer to uncompressed and compressed as 1:1 to the original studio mag masters, and not 1:1 to the original studio mag masters.

RAH

 

[Scott Calvert]

It is specious because lossy data compression does nothing to make a signal sound better, unless you have the ability to hear and actually like the sound of compression artifacts. If that's the case lets just all go with 98kbps MP3s and be done with it.
I have to wonder if some are confusing data compression with dynamic range compression.

 

[John Hodson]

Robert, moving back to picture quality, you noted that:
"...classic films were not meant to be viewed from their camera negatives, which in many cases reveal far too much within the image, that would be hidden in original multiple generations."
If a Blu-ray Disc goes back to the original camera negatives and shows up detail the director never hought would be seen at the cinema (poor make-up, or whatever), do you think the picture quality can ever be too good?
If we can see things which suddenly cause us to stop suspending disbelief, isn't that a bad thing?
Steve W

On that very point, this is from a piece regarding the newly restored A Night To Remember:
...Some of the challenges were the dancing reflections off the water where you can’t use automated dust busting. The grain in some shots needed to be treated with Relativity, and care taken to match the contrast and grain in the dupe negative footage.
Brown notes: “Some of the Titanic model shots, whilst looking okay on the fine-grain, were seen for the first time ‘warts and all’ on the scans from the negative. We were seeing flags and all sorts of things which were all but invisible on the fine-grain due to the much higher contrast, and the fact that the film had only ever been seen previously three generations of laboratory printing further away from the negative than we were seeing for the first time on the five-metre screen at 142.
“On a few of these ‘problem’ shots I was able to draw Garbage Mattes on Resolve and track the film weave with fills and blurs to disguise the bits and pieces that weren't welcome in the final film-out. I felt comfortable in doing this as they were rendered less visible to the original viewing audience due to printing limitations. So, I had to replicate this in a sympathetic way without ‘improvements’ that were never there previously.”

 

[DavidJ]

It is specious because lossy data compression does nothing to make a signal sound better, unless you have the ability to hear and actually like the sound of compression artifacts. If that's the case lets just all go with 98kbps MP3s and be done with it.
I have to wonder if some are confusing data compression with dynamic range compression.

No one has said that it makes it sound better---just that the uncompressed version may be unnecessary and that it might even too easily reveal flaws. You make an interesting point about confusing data compression with dynamic range compression, but I don't think that is what is going on and the compression used on these disks for audio is really quite different from typical data compression since it uses perceptual encoding.
And again, throwing around stuff like "If that's the case lets just all go with 98kbps MP3s and be done with it" is insulting and the people involved in this discussion deserve better.

 

[bigshot]

With the proper settings and codecs, compressed audio should sound identical to lossless. I've done line level matched A/B comparisons myself, and there is a point where all differences become transparent. The issue isn't lossless vs compressed, it's well encoded vs artifact filled sound.

 

[Scott Calvert]

No one has said that it makes it sound better---just that the uncompressed version may be unnecessary and that it might even too easily reveal flaws. You make an interesting point about confusing data compression with dynamic range compression, but I don't think that is what is going on and the compression used on these disks for audio is really quite different from typical data compression since it uses perceptual encoding.
And again, throwing around stuff like "If that's the case lets just all go with 98kbps MP3s and be done with it" is insulting and the people involved in this discussion deserve better.

This is what bugs me to no end. I've seen this bandied about several times and it is complete nonsense.
If you have a lossy compressed signal and it isn't revealing "flaws" in the source that an otherwise losslessly compressed (or uncompressed) signal would reveal, then I would say that's a pretty damn poorly compressed file. The goal of audio data compression is to reduce a file size while at the same time preserving as much of the audio characteristics of the original signal as possible. This is obvious to anyone that knows what they are talking about and I'm surprised on a forum filled with "enthusiasts" that only one or two people (including myself) are calling it out.
So, with that being said, if space isn't an issue, why compress using a lossy codec? Please don't counter that it's a matter of cost unless you really mean that the costs of a human being pressing a button while a computer does the heavy lifting is exhorbitant.

 

[TonyD]

Why aren't the audio flaws fixed. Video flaws get fixed.