High-quality video in versatile camera?

PTravel wrote:
> "Paul Rubin" <http://phr.cx@NOSPAM.invalid> wrote in message
> news:7xbqo7x4nt.fsf@ruckus.brouhaha.com...
>> "PTravel" <ptravel.RemoveThis@travelersvideo.com> writes:
>>> Ah, you beat me to it. I don't know of any professional use of HDV,
>>
>> According to Wikipedia, some parts of the JAG television series are
>> shot on HDV.
>>
>>> I'm a casual video user, and I care very much about frame-accurate
>>> editing, as well as achieving the best possible video quality.
>
> I'm not sure Wikipedia is a reliable authority, but I won't take
> issue with it. The point, though, is that there are significant
> issues with HDV related to motion artifacts which, to date, has
> slowed its adoption as a prosumer/professional format.
>
>
>>
>> I think I'd take issue with "casual video user" in that case, but
>> it's not exactly a club that anyone is clamoring to get into. A
>> casual user is someone who isn't very fussy, pretty much by
>> definition. The still photography equivalent would be someone who
>> takes snapshots with a disposable camera or cell phone camera.
>
> If that's your definition then, yes, video produced by a still camera
> would be on a par with stills produced by a disposable camera or cell
> phone camera. I don't think that's the typical user, however.

It sounds to me like you haven't seen the kind of video I'm talking about.
There is NO comparison with cell-phone video. It is FAR FAR superior to
that. No contest.
Perhaps I should send a sample to you...? My cell phone shoots video,
and it's no cheapie...but it is absoute C-R-A-P by comparison.

You might be surprised to see what some still cameras can produce.

--
Images (Plus Snaps & Grabs) by Mark² at:
www.pbase.com/markuson
 >> Stay informed about: High-quality video in versatile camera? 

"PTravel" <ptravel.TakeThisOut@travelersvideo.com> writes:

>Sorry, but the measure of video quality is the frames per second. Still
>cameras must use very high compression rates to store video on an SD or CF
>card. All of the ones that I've seen use either mpeg1 or mpeg2.

The Canon digicams I've used produce motion-JPEG encoded video. Each
frame is independently compressed, much like DV encoding. Yes, the
files are large.

Your comments about the undesirability of MPEG1 and MPEG2 don't apply to
motion-JPEG video.

Dave
 >> Stay informed about: High-quality video in versatile camera? 

"Mark²" <mjmorgan(lowest even number here)@cox..net> wrote in message
news:vHWZg.50254$nm1.31750@fed1read04...
> PTravel wrote:
>> "Paul Rubin" <http://phr.cx@NOSPAM.invalid> wrote in message
>> news:7xbqo7x4nt.fsf@ruckus.brouhaha.com...
>>> "PTravel" <ptravel DeleteThis @travelersvideo.com> writes:
>>>> Ah, you beat me to it. I don't know of any professional use of HDV,
>>>
>>> According to Wikipedia, some parts of the JAG television series are
>>> shot on HDV.
>>>
>>>> I'm a casual video user, and I care very much about frame-accurate
>>>> editing, as well as achieving the best possible video quality.
>>
>> I'm not sure Wikipedia is a reliable authority, but I won't take
>> issue with it. The point, though, is that there are significant
>> issues with HDV related to motion artifacts which, to date, has
>> slowed its adoption as a prosumer/professional format.
>>
>>
>>>
>>> I think I'd take issue with "casual video user" in that case, but
>>> it's not exactly a club that anyone is clamoring to get into. A
>>> casual user is someone who isn't very fussy, pretty much by
>>> definition. The still photography equivalent would be someone who
>>> takes snapshots with a disposable camera or cell phone camera.
>>
>> If that's your definition then, yes, video produced by a still camera
>> would be on a par with stills produced by a disposable camera or cell
>> phone camera. I don't think that's the typical user, however.
>
> It sounds to me like you haven't seen the kind of video I'm talking about.
> There is NO comparison with cell-phone video. It is FAR FAR superior to
> that. No contest.
> Perhaps I should send a sample to you...? My cell phone shoots video,
> and it's no cheapie...but it is absoute C-R-A-P by comparison.
>
> You might be surprised to see what some still cameras can produce.

I'm afraid we're going in circles. I know what still camera video looks
like and have explained at length why it's of substantially lower quality
than DV-25 from a decent miniDV camcorder.

>
> --
> Images (Plus Snaps & Grabs) by Mark² at:
> www.pbase.com/markuson
>
>
 >> Stay informed about: High-quality video in versatile camera? 

"Dave Martindale" <davem.TakeThisOut@cs.ubc.ca> wrote in message
news:eh9oe0$sqn$1@swain.cs.ubc.ca...
> "PTravel" <ptravel.TakeThisOut@travelersvideo.com> writes:
>
>>Sorry, but the measure of video quality is the frames per second. Still
>>cameras must use very high compression rates to store video on an SD or CF
>>card. All of the ones that I've seen use either mpeg1 or mpeg2.
>
> The Canon digicams I've used produce motion-JPEG encoded video. Each
> frame is independently compressed, much like DV encoding. Yes, the
> files are large.
>
> Your comments about the undesirability of MPEG1 and MPEG2 don't apply to
> motion-JPEG video.

Only with respect to the need for multi-pass analysis and lack of ease of
editing. Because mjpeg doesn't use intraframe compression, like mpeg, it
cannot compress as efficiently. That results in far higher compression
ratios being required to accomodate the limited amount of storage media
available, far lower data bandwidth and a corresponding decrease in quality.

Also, unless the camera comes with software that will transcode captured
mjpeg to DV-codec-encoded AVI (which will result in still more quality
loss), the mjpeg-encoded AVI (or whatever it produces) will not be readily
editable in most consumer and prosumer editing programs. At best, every
single frame will have to be rendered before it can be worked with. At
worst, it won't be imported at all without transcoding and subsequent
quality loss.

>
> Dave
 >> Stay informed about: High-quality video in versatile camera? 

"Dave Martindale" <davem DeleteThis @cs.ubc.ca> wrote in message
news:eh9oe0$sqn$1@swain.cs.ubc.ca...
> "PTravel" <ptravel DeleteThis @travelersvideo.com> writes:
>
>>Sorry, but the measure of video quality is the frames per second. Still
>>cameras must use very high compression rates to store video on an SD or CF
>>card. All of the ones that I've seen use either mpeg1 or mpeg2.

I just noticed a significant typo in what you've quoted. I should have
written, "Sorry, but the measure of video quality is NOT the frames per
second."


>
> The Canon digicams I've used produce motion-JPEG encoded video. Each
> frame is independently compressed, much like DV encoding. Yes, the
> files are large.
>
> Your comments about the undesirability of MPEG1 and MPEG2 don't apply to
> motion-JPEG video.
>
> Dave
 >> Stay informed about: High-quality video in versatile camera? 

"PTravel" <ptravel DeleteThis @travelersvideo.com> writes:
> Only with respect to the need for multi-pass analysis and lack of ease of
> editing.

It occurs to me that nothing stops a camera from doing multi-pass
compression if it has enough ram to buffer some extra frames, where a
few dozen frames ought to be enough, not that big a deal in terms of
hardware cost. I don't know if any actually do that.

> Because mjpeg doesn't use intraframe compression, like mpeg, it
> cannot compress as efficiently. That results in far higher compression
> ratios being required to accomodate the limited amount of storage media
> available, far lower data bandwidth and a corresponding decrease in quality.

Now you're going around in circles--the DV25 that you like has the
exact same issue.

> Also, unless the camera comes with software that will transcode captured
> mjpeg to DV-codec-encoded AVI (which will result in still more quality
> loss), the mjpeg-encoded AVI (or whatever it produces) will not be readily
> editable in most consumer and prosumer editing programs. At best, every
> single frame will have to be rendered before it can be worked with.

???? Same is true for DV25. I don't know whether typical video
editing programs edit mjpeg directly (same way the edit dv25) but
there's certainly no technical reason why they shouldn't, if enough
users care about it.
 >> Stay informed about: High-quality video in versatile camera? 

"Paul Rubin" <http://phr.cx@NOSPAM.invalid> wrote in message
news:7x3b9jw2ig.fsf@ruckus.brouhaha.com...
> "PTravel" <ptravel RemoveThis @travelersvideo.com> writes:
>> Only with respect to the need for multi-pass analysis and lack of ease of
>> editing.
>
> It occurs to me that nothing stops a camera from doing multi-pass
> compression if it has enough ram to buffer some extra frames, where a
> few dozen frames ought to be enough, not that big a deal in terms of
> hardware cost. I don't know if any actually do that.

Mutli-pass analysis requires more than just a few buffered frames. It
requires analysis of the entire scene, as well as scene change detection. I
suppose with a big enough buffer it could be done, but I've never heard of
any hardware-based multi-pass solution, at least not at the consumer or
prosumer level.

>
>> Because mjpeg doesn't use intraframe compression, like mpeg, it
>> cannot compress as efficiently. That results in far higher compression
>> ratios being required to accomodate the limited amount of storage media
>> available, far lower data bandwidth and a corresponding decrease in
>> quality.
>
> Now you're going around in circles--the DV25 that you like has the
> exact same issue.

DV-25 compresses at roughly a 5-to-1 ratio. Mpeg2, which uses temporal
compression, compresses 10-to-1 or more with a noticeable quality hit.
Mjpeg, which does not use temporal compression, must fit the same amount of
data in far, far less space. It's compression ratio is going to be much
higher, resulting in a much lower bandwidth, i.e. less detail is stored and
the video is of lower quality.

>
>> Also, unless the camera comes with software that will transcode captured
>> mjpeg to DV-codec-encoded AVI (which will result in still more quality
>> loss), the mjpeg-encoded AVI (or whatever it produces) will not be
>> readily
>> editable in most consumer and prosumer editing programs. At best, every
>> single frame will have to be rendered before it can be worked with.
>
> ???? Same is true for DV25.

Not true. Virtually all consumer and prosumer edits use DV-codec-encoded
AVI (or equivalent) as their native format. No transcdoing is necessary.
In order to work with the mjpeg, editors that use DV-codec-encodec AVI as
their native format will have to transcode the mjpeg.

> I don't know whether typical video
> editing programs edit mjpeg directly (same way the edit dv25) but
> there's certainly no technical reason why they shouldn't, if enough
> users care about it.

That's correct, there's no technical reason that they couldn't. However,
they don't because the vast majority of editing is done working with DV-25
source, not mjpeg source.
 >> Stay informed about: High-quality video in versatile camera? 

"PTravel" <ptravel RemoveThis @travelersvideo.com> writes:
> > It occurs to me that nothing stops a camera from doing multi-pass
> > compression if it has enough ram to buffer some extra frames, where a
> > few dozen frames ought to be enough, not that big a deal in terms of
> > hardware cost. I don't know if any actually do that.
>
> Mutli-pass analysis requires more than just a few buffered frames. It
> requires analysis of the entire scene, as well as scene change detection.
> I suppose with a big enough buffer it could be done, but I've never
> heard of any hardware-based multi-pass solution,

According to Wikipedia, which is pretty reliable about stuff like
this, there's commonly 12 or 15 frames per group of pictures where the
group is delimited by I frames (those are the frames with no temporal
compression). A few dozen frames should be enough for analysis of a
group of that size.

> DV-25 compresses at roughly a 5-to-1 ratio. Mpeg2, which uses temporal
> compression, compresses 10-to-1 or more with a noticeable quality hit.

If you're talking about DVD, that usually runs at 4 mbit/sec, which is
about 1/6th the bit rate of DV25, so

> Mjpeg, which does not use temporal compression, must fit the same amount of
> data in far, far less space.

On the other hand JPEG compression is more space efficient (but more
cpu intensive) than DV25 compression.

Point is though, most users don't care about this. DVD quality is
good enough. VHS quality was good enough and 100's of millions of
people used it. Yes, laserdisc was better, but most users didn't
care. VHS sucked badly when the tape got worn, or got partially
demagnetized from sitting on top of a TV set, or after multiple
generations of analog copying especially in the long-play modes.
Also, the tapes were bulky and relatively expensive, the VCR itself
was large and fragile, etc. But the video quality when you had a tape
that wasn't trashed by the above factors was perfectly fine by most
users' standards. People watch broadcast TV all the time too, after
all.

> > ???? Same is true for DV25.
>
> Not true. Virtually all consumer and prosumer edits use DV-codec-encoded
> AVI (or equivalent) as their native format. No transcdoing is necessary.
> In order to work with the mjpeg, editors that use DV-codec-encodec AVI as
> their native format will have to transcode the mjpeg.

There's in a sense no such thing as "native format" for software, it's
just ones and zeros. Of course for stuff like adding video effects,
the frames have to actually be rendered into pixels, so by that
standard even DV isn't "native". Some editing programs might be silly
enough to transcode mjpeg to dv before editing, but there's not an
especially good reason for them to do that.

> That's correct, there's no technical reason that they couldn't. However,
> they don't because the vast majority of editing is done working with DV-25
> source, not mjpeg source.

That's probably because most users who care about precise editing are
shooting on DV, so mjpeg editing in those programs was slapped on as
an afterthought without much attention to doing it the best way.
That's a problem with the silly software, not with the mjpeg format.
 >> Stay informed about: High-quality video in versatile camera? 

"Paul Rubin" <http://phr.cx@NOSPAM.invalid> wrote in message
news:7xirieed54.fsf@ruckus.brouhaha.com...
> "PTravel" <ptravel.DeleteThis@travelersvideo.com> writes:
>> > It occurs to me that nothing stops a camera from doing multi-pass
>> > compression if it has enough ram to buffer some extra frames, where a
>> > few dozen frames ought to be enough, not that big a deal in terms of
>> > hardware cost. I don't know if any actually do that.
>>
>> Mutli-pass analysis requires more than just a few buffered frames. It
>> requires analysis of the entire scene, as well as scene change detection.
>> I suppose with a big enough buffer it could be done, but I've never
>> heard of any hardware-based multi-pass solution,
>
> According to Wikipedia, which is pretty reliable about stuff like
> this, there's commonly 12 or 15 frames per group of pictures where the
> group is delimited by I frames (those are the frames with no temporal
> compression). A few dozen frames should be enough for analysis of a
> group of that size.

12 to 15 frames per I groups is more-or-less standard, but still doesn't
reflect multiple-pass analysis, since the I frames have to be selected and
that is based on forward and backward comparison.

>
>> DV-25 compresses at roughly a 5-to-1 ratio. Mpeg2, which uses temporal
>> compression, compresses 10-to-1 or more with a noticeable quality hit.
>
> If you're talking about DVD, that usually runs at 4 mbit/sec, which is
> about 1/6th the bit rate of DV25, so

DVDs usually use variable bit rate, but the max is 10 mbs (8 for
compatibility with older set-top players).

>
>> Mjpeg, which does not use temporal compression, must fit the same amount
>> of
>> data in far, far less space.
>
> On the other hand JPEG compression is more space efficient (but more
> cpu intensive) than DV25 compression.

I don't know whether that is true or not but, regardless, it is not so
efficient that it can produce the same quality with less than half the
bandwidth.

>
> Point is though, most users don't care about this.

And that always seems to be the arguement of those who advocate for the
low-end consumer formats. I don't know where you get this "most users"
from. I certainly care. Those of us who who buy higher-end consumer and
prosumer cameras certainly care.

> DVD quality is
> good enough. VHS quality was good enough and 100's of millions of
> people used it.

1/2" black-and-white video was good enough when that was all that was
available. I didn't start shooting video, though, until Hi8 came out with
close-to-broadcast quality. At that point, it made sense to me use that as
a medium for preserving memories that were important to me. Drugstore
printing of still photos was good enough for many people, but there were
still plenty of us of had our own darkrooms.

I edit my videos rather extensively. I correct them. I add effects. I
composite. I title. I do all of this because, 10 or 20 or 30 years from
now, I went the best record of what I shoot. And, believe me, there are
plenty of people like me or Adobe wouldn't sell as many copies of Premiere
as it does.

> Yes, laserdisc was better, but most users didn't
> care.

Most users didn't care because (1) large screen televisions were hideously
expensive then, and (2) laser disk was a pain because of the severe time
limitations.

> VHS sucked badly when the tape got worn, or got partially
> demagnetized from sitting on top of a TV set, or after multiple
> generations of analog copying especially in the long-play modes.
> Also, the tapes were bulky and relatively expensive, the VCR itself
> was large and fragile, etc. But the video quality when you had a tape
> that wasn't trashed by the above factors was perfectly fine by most
> users' standards. People watch broadcast TV all the time too, after
> all.

Sure, but VHS was never even close to broadcast quality. S-VHS was and,
given the development of inexpensive, quality large-screen monitors, would
have caught on more than it did except that DVDs were a far better solution
for pre-recorded material.

>
>> > ???? Same is true for DV25.
>>
>> Not true. Virtually all consumer and prosumer edits use DV-codec-encoded
>> AVI (or equivalent) as their native format. No transcdoing is necessary.
>> In order to work with the mjpeg, editors that use DV-codec-encodec AVI as
>> their native format will have to transcode the mjpeg.
>
> There's in a sense no such thing as "native format" for software, it's
> just ones and zeros.

On the contrary, all editors have their preferred format. Premiere Pro, for
example, must transcode ANYTHING that isn't already DV-codec-encoded AVI.
That means that if you import anything than DV-25 source, Premiere Pro will
need to render it before you can do anything with it. This has been true as
long as I've been using Premiere, which goes back to version 6.0.

> Of course for stuff like adding video effects,
> the frames have to actually be rendered into pixels, so by that
> standard even DV isn't "native". Some editing programs might be silly
> enough to transcode mjpeg to dv before editing, but there's not an
> especially good reason for them to do that.

Virtually all editing programs will have to do that.

>
>> That's correct, there's no technical reason that they couldn't. However,
>> they don't because the vast majority of editing is done working with
>> DV-25
>> source, not mjpeg source.
>
> That's probably because most users who care about precise editing are
> shooting on DV, so mjpeg editing in those programs was slapped on as
> an afterthought without much attention to doing it the best way.
> That's a problem with the silly software, not with the mjpeg format.

As you note, those who care about precise editing, i.e. quality, are
shooting DV-25. It's not silly software, it's software designed for those
who need to use it. There may be nothing wrong with mjpeg, per se, but as
implemented in digital still cameras, i.e. low bandwidth, high compression
(and we haven't even discussed audio, which is another issue altogether --
synced high-sampling stereo audio is part of the DV-25 spec) to accomodate
digital still solid state memory limitations results in video which is
significantly and visibly inferior video.
 >> Stay informed about: High-quality video in versatile camera? 

PTravel wrote:
> "Paul Rubin" <http://phr.cx@NOSPAM.invalid> wrote in message
> news:7xirieed54.fsf@ruckus.brouhaha.com...
>> "PTravel" <ptravel RemoveThis @travelersvideo.com> writes:
>>>> It occurs to me that nothing stops a camera from doing multi-pass
>>>> compression if it has enough ram to buffer some extra frames,
>>>> where a few dozen frames ought to be enough, not that big a deal
>>>> in terms of hardware cost. I don't know if any actually do that.
>>>
>>> Mutli-pass analysis requires more than just a few buffered frames. It
>>> requires analysis of the entire scene, as well as scene change
>>> detection. I suppose with a big enough buffer it could be done, but
>>> I've never heard of any hardware-based multi-pass solution,
>>
>> According to Wikipedia, which is pretty reliable about stuff like
>> this, there's commonly 12 or 15 frames per group of pictures where
>> the group is delimited by I frames (those are the frames with no
>> temporal compression). A few dozen frames should be enough for
>> analysis of a group of that size.
>
> 12 to 15 frames per I groups is more-or-less standard, but still
> doesn't reflect multiple-pass analysis, since the I frames have to be
> selected and that is based on forward and backward comparison.
>
>>
>>> DV-25 compresses at roughly a 5-to-1 ratio. Mpeg2, which uses
>>> temporal compression, compresses 10-to-1 or more with a noticeable
>>> quality hit.
>>
>> If you're talking about DVD, that usually runs at 4 mbit/sec, which
>> is about 1/6th the bit rate of DV25, so
>
> DVDs usually use variable bit rate, but the max is 10 mbs (8 for
> compatibility with older set-top players).
>
>>
>>> Mjpeg, which does not use temporal compression, must fit the same
>>> amount of
>>> data in far, far less space.
>>
>> On the other hand JPEG compression is more space efficient (but more
>> cpu intensive) than DV25 compression.
>
> I don't know whether that is true or not but, regardless, it is not so
> efficient that it can produce the same quality with less than half the
> bandwidth.
>
>>
>> Point is though, most users don't care about this.
>
> And that always seems to be the arguement of those who advocate for
> the low-end consumer formats. I don't know where you get this "most
> users" from.

I got it from the OP... This is...after all about his question.
He's indicated quite clearly that he would be satisfied with exactly the
kind of "quality" (or lack of) which I referred to. When did this thread
become all about your needs? I posted with regard to the OP.
I'm more picky about still-image pictures than anyone I know, but that
doesn't mean most are that way. The overwhelming majority simply aren't.
 >> Stay informed about: High-quality video in versatile camera? 

"PTravel" <ptravel.TakeThisOut@travelersvideo.com> writes:
> 12 to 15 frames per I groups is more-or-less standard, but still doesn't
> reflect multiple-pass analysis, since the I frames have to be selected and
> that is based on forward and backward comparison.

Yes but you should not have to go forward or back more than 12 or 15
frames. Once you've decided to emit an I frame, everything starts
again from there.

> > If you're talking about DVD, that usually runs at 4 mbit/sec, which is
> > about 1/6th the bit rate of DV25, so
>
> DVDs usually use variable bit rate, but the max is 10 mbs (8 for
> compatibility with older set-top players).

4 mbps average is what I've always heard. Maybe there are occasional
bursts to 8 mbps.

> > Point is though, most users don't care about this.
>
> And that always seems to be the arguement of those who advocate for the
> low-end consumer formats. I don't know where you get this "most users"
> from. I certainly care. Those of us who who buy higher-end consumer and
> prosumer cameras certainly care.

Yes, and most users don't buy higher end equipment or else that
equipment wouldn't be called "higher end". You are not "most users".
Remember Betamax? It was better than VHS, but VHS beat it out,
because VHS was good enough and users felt that its advantages
(particularly the even lower quality long-play modes) outweighed
Betamax's slightly better quality. Now we've got HD-DVD and/or
Blu-Ray which are eagerly awaited by computer users with their
bottomless appetite for digital storage, but which appear to be
bombing out with video users, because the video quality improvement,
while real, just isn't important to as many users as they'd hoped.

There was a similar situation for audio in the pre-CD era.
Prerecorded music cassettes consistently beat out vinyl LP records in
sales volume, despite costing about the same and sounding worse,
because people liked the convenience of portable players etc. CD's
came along and the quality improved, but people mainly liked the
convenience. Nowadays, CD's seem almost archaic and most people are
listening to downloaded mp3's on their ipods. There is a definite
quality loss compared with CD but as the story always goes, once it's
good enough, for most users, vastly improved convenience counts far
more than some tiny quality increment.

Getting out of gadgetry altogether, fancy audio and video is not too
different from gourmet dining. Going out to a fancy restaurant or
cooking something elaborate is nice if you're into that. Some
connoisseurs can't stand anything less. Most of us don't care. If we
can warm up something at home that reaches a reasonable standard of
taste and nourishment, it's just not worth doing an elaborate
preparation or going out and spending big $$$, just to get something
better, unless it's a special occasion.

> I edit my videos rather extensively. I correct them. I add effects. I
> composite. I title. I do all of this because, 10 or 20 or 30 years from
> now, I went the best record of what I shoot. And, believe me, there are
> plenty of people like me or Adobe wouldn't sell as many copies of Premiere
> as it does.

Maybe you also alphabetize your socks. Most people don't. If you
divide the number of copies of Premiere sold by the number of
camcorders (of all formats) sold, you'll probably get a percentage in
single digits at most. And even people who bought Premiere are in
many cases probably not using it in more than the simplest ways.
Actually, a lot were probably bundled with video equipment and never
used at all. I have a copy of Photoshop that came with my scanner and
I've never opened the CD. I don't even have a Windows machine to run
it on. And the program I use instead (GIMP) has lots of fancy
features like Photoshop does, but I never use any of them.

> > Yes, laserdisc was better, but most users didn't care.
>
> Most users didn't care because (1) large screen televisions were
> hideously expensive then, and (2) laser disk was a pain because of
> the severe time limitations.

I watched VHS on large screens and it looked fine. Maybe it had some
problems that I could have noticed if I were fussy, but I was always
more interested in watching the movie than scrutinizing the video.

> > VHS sucked badly when the tape got worn, ... People watch
> > broadcast TV all the time too, after all.
> Sure, but VHS was never even close to broadcast quality.

I'll take your word for it. I'm used to broadcast TV looking awful
because of multipath distortion, static and snow from marginal
reception, etc. Even comparing cable TV to VHS, I don't think I ever
really was bothered by the difference, though I suppose it was
noticable to anyone paying attention to such things.

> > There's in a sense no such thing as "native format" for software, it's
> > just ones and zeros.
>
> On the contrary, all editors have their preferred format. Premiere Pro, for
> example, must transcode ANYTHING that isn't already DV-codec-encoded AVI.

What do you mean "must"? That just means the implementers left an
annoying limitation in it that they could have fixed if they thought
it mattered. So I'd say "does" but not "must".

> > Some editing programs might be silly enough to transcode mjpeg to
> > dv before editing, but there's not an especially good reason for
> > them to do that.
> Virtually all editing programs will have to do that.

"Have to"? The programs do what the implementers decide to implement.

> As you note, those who care about precise editing, i.e. quality, are
> shooting DV-25. It's not silly software, it's software designed for those
> who need to use it.

There we have it then, the implementers correctly notice that non-DV25
users don't care enough about precision editing to make it worth their
while to do a better job with mjpeg.

> There may be nothing wrong with mjpeg, per se, but as
> implemented in digital still cameras, i.e. low bandwidth, high compression

I think we're mostly in agreement, that the video coming out of these
still cameras (mjpeg or whatever) is somewhat worse than DV video and
somewhat better than VHS. It may be roughly equivalent to DVD.
Whether that's up to the standards of any particular potential user is
up to that user. I think most people have seen enough VHS and DVD
video to make their own call.

> (and we haven't even discussed audio, which is another issue altogether --

We did discuss audio. Audio is the main limitation of those still
cameras as I see it. They generally only have a mono microphone with
no provision for plugging in an external mic. I'll add that the
internal mic in mine seems to pick up an awful lot of noise despite
the camera having no moving parts.
 >> Stay informed about: High-quality video in versatile camera? 

"Mark²" <mjmorgan(lowest even number here)@cox..net> wrote in message
news:4le_g.51258$nm1.42002@fed1read04...
> PTravel wrote:
>> "Paul Rubin" <http://phr.cx@NOSPAM.invalid> wrote in message
>> news:7xirieed54.fsf@ruckus.brouhaha.com...
>>> "PTravel" <ptravel RemoveThis @travelersvideo.com> writes:
>>>>> It occurs to me that nothing stops a camera from doing multi-pass
>>>>> compression if it has enough ram to buffer some extra frames,
>>>>> where a few dozen frames ought to be enough, not that big a deal
>>>>> in terms of hardware cost. I don't know if any actually do that.
>>>>
>>>> Mutli-pass analysis requires more than just a few buffered frames. It
>>>> requires analysis of the entire scene, as well as scene change
>>>> detection. I suppose with a big enough buffer it could be done, but
>>>> I've never heard of any hardware-based multi-pass solution,
>>>
>>> According to Wikipedia, which is pretty reliable about stuff like
>>> this, there's commonly 12 or 15 frames per group of pictures where
>>> the group is delimited by I frames (those are the frames with no
>>> temporal compression). A few dozen frames should be enough for
>>> analysis of a group of that size.
>>
>> 12 to 15 frames per I groups is more-or-less standard, but still
>> doesn't reflect multiple-pass analysis, since the I frames have to be
>> selected and that is based on forward and backward comparison.
>>
>>>
>>>> DV-25 compresses at roughly a 5-to-1 ratio. Mpeg2, which uses
>>>> temporal compression, compresses 10-to-1 or more with a noticeable
>>>> quality hit.
>>>
>>> If you're talking about DVD, that usually runs at 4 mbit/sec, which
>>> is about 1/6th the bit rate of DV25, so
>>
>> DVDs usually use variable bit rate, but the max is 10 mbs (8 for
>> compatibility with older set-top players).
>>
>>>
>>>> Mjpeg, which does not use temporal compression, must fit the same
>>>> amount of
>>>> data in far, far less space.
>>>
>>> On the other hand JPEG compression is more space efficient (but more
>>> cpu intensive) than DV25 compression.
>>
>> I don't know whether that is true or not but, regardless, it is not so
>> efficient that it can produce the same quality with less than half the
>> bandwidth.
>>
>>>
>>> Point is though, most users don't care about this.
>>
>> And that always seems to be the arguement of those who advocate for
>> the low-end consumer formats. I don't know where you get this "most
>> users" from.
>
> I got it from the OP... This is...after all about his question.
> He's indicated quite clearly that he would be satisfied with exactly the
> kind of "quality" (or lack of) which I referred to. When did this thread
> become all about your needs? I posted with regard to the OP.
> I'm more picky about still-image pictures than anyone I know, but that
> doesn't mean most are that way. The overwhelming majority simply aren't.

Please read the thread title, which is what this thread is about:
"High-quality video in versatile camera?"

The answer to the question remains what it was when I first posted: Digital
still cameras do not produce high-quality video.

>
>
 >> Stay informed about: High-quality video in versatile camera? 

PTravel wrote:
> "Mark²" <mjmorgan(lowest even number here)@cox..net> wrote in message
> news:4le_g.51258$nm1.42002@fed1read04...
>> PTravel wrote:
>>> "Paul Rubin" <http://phr.cx@NOSPAM.invalid> wrote in message
>>> news:7xirieed54.fsf@ruckus.brouhaha.com...
>>>> "PTravel" <ptravel.RemoveThis@travelersvideo.com> writes:
>>>>>> It occurs to me that nothing stops a camera from doing multi-pass
>>>>>> compression if it has enough ram to buffer some extra frames,
>>>>>> where a few dozen frames ought to be enough, not that big a deal
>>>>>> in terms of hardware cost. I don't know if any actually do that.
>>>>>
>>>>> Mutli-pass analysis requires more than just a few buffered
>>>>> frames. It requires analysis of the entire scene, as well as
>>>>> scene change detection. I suppose with a big enough buffer it
>>>>> could be done, but I've never heard of any hardware-based
>>>>> multi-pass solution,
>>>>
>>>> According to Wikipedia, which is pretty reliable about stuff like
>>>> this, there's commonly 12 or 15 frames per group of pictures where
>>>> the group is delimited by I frames (those are the frames with no
>>>> temporal compression). A few dozen frames should be enough for
>>>> analysis of a group of that size.
>>>
>>> 12 to 15 frames per I groups is more-or-less standard, but still
>>> doesn't reflect multiple-pass analysis, since the I frames have to
>>> be selected and that is based on forward and backward comparison.
>>>
>>>>
>>>>> DV-25 compresses at roughly a 5-to-1 ratio. Mpeg2, which uses
>>>>> temporal compression, compresses 10-to-1 or more with a noticeable
>>>>> quality hit.
>>>>
>>>> If you're talking about DVD, that usually runs at 4 mbit/sec, which
>>>> is about 1/6th the bit rate of DV25, so
>>>
>>> DVDs usually use variable bit rate, but the max is 10 mbs (8 for
>>> compatibility with older set-top players).
>>>
>>>>
>>>>> Mjpeg, which does not use temporal compression, must fit the same
>>>>> amount of
>>>>> data in far, far less space.
>>>>
>>>> On the other hand JPEG compression is more space efficient (but
>>>> more cpu intensive) than DV25 compression.
>>>
>>> I don't know whether that is true or not but, regardless, it is not
>>> so efficient that it can produce the same quality with less than
>>> half the bandwidth.
>>>
>>>>
>>>> Point is though, most users don't care about this.
>>>
>>> And that always seems to be the arguement of those who advocate for
>>> the low-end consumer formats. I don't know where you get this "most
>>> users" from.
>>
>> I got it from the OP... This is...after all about his question.
>> He's indicated quite clearly that he would be satisfied with exactly
>> the kind of "quality" (or lack of) which I referred to. When did
>> this thread become all about your needs? I posted with regard to
>> the OP. I'm more picky about still-image pictures than anyone I know, but
>> that doesn't mean most are that way. The overwhelming majority
>> simply aren't.
>
> Please read the thread title, which is what this thread is about:
> "High-quality video in versatile camera?"
>
> The answer to the question remains what it was when I first posted:
> Digital still cameras do not produce high-quality video.

You're assuming an absolute value where none has been assigned...except by
you.
"High quality" is absolutely relative in its meaning.
Compared to Cell phone video, it is extremely high quality.
Compared to VHS, is is better.
Compared to Better DV camcorders, it is worse.
Go back and read the response of the OP and you'll see what I'm referring
to.

In any case, it matters very little, because it is what it is.
It suits the purposes of some very well... -Just not yours.

--
Images (Plus Snaps & Grabs) by Mark² at:
www.pbase.com/markuson
 >> Stay informed about: High-quality video in versatile camera?