Sign in with
Sign up | Sign in
Your question

Where are the cheap MPEG2 hardware accelerators?

Last response: in Graphics & Displays
Share
Anonymous
August 31, 2004 6:52:04 PM

Archived from groups: rec.video.desktop (More info?)

Given that DVD recorders are now around $300 and include DVD burners and a
lot of other logic and components, why haven't we seen MPEG2 hardware
acceleration for PCs that's not really expensive?

Is it realistic or even possible to expect a sub-$100 board to provide
one-pass MPEG2 acceleration at 2x real time speeds?

Even on my P4-3.2c system, I'm still looking at 120% of real time for
low-bitrate one-pass MPEG2 encodes with TMPGEnc. Two pass is well over 200%
of real time, and it's all CPU, not a disk or other bottleneck that I can
tell.

I've read that "in the future" encoding processing will be a feature of
programmable GPUs (apparently already is a feature of the GeForce series 6
GPUs), but I'm frankly surprised that you can't buy a $100 card that will
provide at least real-time MPEG2 encoding, if not faster than real time.
Anonymous
September 1, 2004 1:50:20 PM

Archived from groups: rec.video.desktop (More info?)

On Tue, 31 Aug 2004 14:52:04 -0500, "Shawn Barnhart"
<usenet@grasslake.net> wrote:

>Given that DVD recorders are now around $300 and include DVD burners and a
>lot of other logic and components, why haven't we seen MPEG2 hardware
>acceleration for PCs that's not really expensive?
>
>Is it realistic or even possible to expect a sub-$100 board to provide
>one-pass MPEG2 acceleration at 2x real time speeds?
>
>Even on my P4-3.2c system, I'm still looking at 120% of real time for
>low-bitrate one-pass MPEG2 encodes with TMPGEnc. Two pass is well over 200%
>of real time, and it's all CPU, not a disk or other bottleneck that I can
>tell.
>
>I've read that "in the future" encoding processing will be a feature of
>programmable GPUs (apparently already is a feature of the GeForce series 6
>GPUs), but I'm frankly surprised that you can't buy a $100 card that will
>provide at least real-time MPEG2 encoding, if not faster than real time.


The reason for this, in my opinion, is the fact that the technology
for this (mpeg-2 encoding) isn't freeware. The developers ask there
more than fair share (licensing) of anyone who implements mpeg-2. As
the demand for standalone dvd-recorders is large and number of units
sold are high the costs of using mpeg-2 technology can be held
relatively low per unit sold. In case of mpeg-2 boards there is much
less demand and much lower sales are expected, so the very few
producers of this type of equipment are forced to keep there prices
high to compensate the relatively high cost of being able to use the
mpeg-2 technology.
Anonymous
September 1, 2004 1:50:21 PM

Archived from groups: rec.video.desktop (More info?)

"Funprice" <jhbusscherNOSPAM@freemail.nl> wrote in message
news:12vaj0pnou29vvi0l76gcf4mb36fgrg16j@4ax.com...

> The reason for this, in my opinion, is the fact that the technology
> for this (mpeg-2 encoding) isn't freeware. The developers ask there
> more than fair share (licensing) of anyone who implements mpeg-2. As

Sure, but for $79 or whatever you can buy a half-dozen DVD/MPEG2 authoring
programs. If MPEG2 licensing was the reason, the cost of software-only
solutions ought to be a lot higher, as should the cost of MPEG2 tuner cards,
like the Hauppage PVR 250/350 series.

> the demand for standalone dvd-recorders is large and number of units
> sold are high the costs of using mpeg-2 technology can be held
> relatively low per unit sold. In case of mpeg-2 boards there is much
> less demand and much lower sales are expected, so the very few
> producers of this type of equipment are forced to keep there prices
> high to compensate the relatively high cost of being able to use the
> mpeg-2 technology.

I'd guess I'd buy a demand rationale -- I don't know how many people would
buy an MPEG2 accelerator card, although perhaps the real reason is that the
inexpensive chipsets that implement MPEG2 are probably integrated with
tuners and/or expect an analog signal, making them less suitable for
accelerating digital file conversion to MPEG2.
Related resources
Anonymous
September 1, 2004 7:07:44 PM

Archived from groups: rec.video.desktop (More info?)

On Wed, 1 Sep 2004 07:33:22 -0500, "Shawn Barnhart"
<usenet@grasslake.net> wrote:

>
>"Funprice" <jhbusscherNOSPAM@freemail.nl> wrote in message
>news:12vaj0pnou29vvi0l76gcf4mb36fgrg16j@4ax.com...
>
>> The reason for this, in my opinion, is the fact that the technology
>> for this (mpeg-2 encoding) isn't freeware. The developers ask there
>> more than fair share (licensing) of anyone who implements mpeg-2. As
>
>Sure, but for $79 or whatever you can buy a half-dozen DVD/MPEG2 authoring
>programs. If MPEG2 licensing was the reason, the cost of software-only
>solutions ought to be a lot higher, as should the cost of MPEG2 tuner cards,
>like the Hauppage PVR 250/350 series.
>
>> the demand for standalone dvd-recorders is large and number of units
>> sold are high the costs of using mpeg-2 technology can be held
>> relatively low per unit sold. In case of mpeg-2 boards there is much
>> less demand and much lower sales are expected, so the very few
>> producers of this type of equipment are forced to keep there prices
>> high to compensate the relatively high cost of being able to use the
>> mpeg-2 technology.
>
>I'd guess I'd buy a demand rationale -- I don't know how many people would
>buy an MPEG2 accelerator card, although perhaps the real reason is that the
>inexpensive chipsets that implement MPEG2 are probably integrated with
>tuners and/or expect an analog signal, making them less suitable for
>accelerating digital file conversion to MPEG2.

It could be the fact that implementing mpeg-2 technology in software
isn't patented and in hardware it is. The Hauppage 250+350 cards are
noticable higher in price than tv-cards without mpeg-2 hardware
encoding! This maybe backups my assumption of a price to pay for using
mpeg-2 technology.

Sure. I stick to my idea of to little demand for cards which do
encoding to mpeg-2 from a digital source. Chips are available (c-cube)
that can do this though, so check out there prices, I expect them to
be high priced. (Due to licensing costs of the MPEG organisation.)
Anonymous
September 1, 2004 7:07:45 PM

Archived from groups: rec.video.desktop (More info?)

"Funprice" <jhbusscherNOSPAM@freemail.nl> wrote in message
news:h6hbj09ccolm1096fk5n6ohov1lnahuc0m@4ax.com...

> It could be the fact that implementing mpeg-2 technology in software
> isn't patented and in hardware it is. The Hauppage 250+350 cards are
> noticable higher in price than tv-cards without mpeg-2 hardware
> encoding! This maybe backups my assumption of a price to pay for using
> mpeg-2 technology.

OK, so even if this was the case, the Hauppage 350 is a whopping $184. I'd
pay that much for a card that could do MPEG2 from DV sources at 2x real
time. I kind of doubt that MPEG2 licensing fees of any kind are major
(beyond $50) price contributors in any (sub-$300) MPEG2 hardware encoders.

> Sure. I stick to my idea of to little demand for cards which do
> encoding to mpeg-2 from a digital source. Chips are available (c-cube)
> that can do this though, so check out there prices, I expect them to
> be high priced. (Due to licensing costs of the MPEG organisation.)

I'd wager that most of the implementations are A->D and not D->D. There are
a fair number of boards that do that (albeit at real time as part of a
capture process; not as part of a post-capture encoding acceleration).

Given that the capture boards are sub $200, I still wonder why I can't do at
least real-time from digital source for that kind of money.
Anonymous
September 2, 2004 12:14:09 AM

Archived from groups: rec.video.desktop (More info?)

"Funprice" <jhbusscherNOSPAM@freemail.nl> wrote in message
news:h6hbj09ccolm1096fk5n6ohov1lnahuc0m@4ax.com...
> It could be the fact that implementing mpeg-2 technology in software
> isn't patented and in hardware it is. The Hauppage 250+350 cards are
> noticable higher in price than tv-cards without mpeg-2 hardware
> encoding! This maybe backups my assumption of a price to pay for using
> mpeg-2 technology.

Both software and hardware implementations of MPEG encoding are subject to
MPEG licensing. The license fee is small and has little to do with the
price of hardware encoders. The pricing is entirely a supply and demand
issue. There is little demand for realtime h/w encoders in PC form factor,
and even less demand for h/w based transcoders. You can get realtime
software encoding from a cheap tv capture card but quality is a significant
issue. I believe that the CinemaCraft Basic encoder for less than $60
provides good quality encoding at much higher speed than Tmpgenc. See
www.visiblelight.com for CCBasic.

> Sure. I stick to my idea of to little demand for cards which do
> encoding to mpeg-2 from a digital source. Chips are available (c-cube)
> that can do this though, so check out there prices, I expect them to
> be high priced. (Due to licensing costs of the MPEG organisation.)
Anonymous
September 2, 2004 12:14:10 AM

Archived from groups: rec.video.desktop (More info?)

"FLY135" <fly_135(@ hot not not)notmail.com> wrote in message
news:l4qZc.521$Wv5.123@newsread3.news.atl.earthlink.net...
>
> "Funprice" <jhbusscherNOSPAM@freemail.nl> wrote in message
> news:h6hbj09ccolm1096fk5n6ohov1lnahuc0m@4ax.com...
> > It could be the fact that implementing mpeg-2 technology in software
> > isn't patented and in hardware it is. The Hauppage 250+350 cards are
> > noticable higher in price than tv-cards without mpeg-2 hardware
> > encoding! This maybe backups my assumption of a price to pay for using
> > mpeg-2 technology.
>
> Both software and hardware implementations of MPEG encoding are subject to
> MPEG licensing. The license fee is small and has little to do with the
> price of hardware encoders. The pricing is entirely a supply and demand
> issue. There is little demand for realtime h/w encoders in PC form
factor,
> and even less demand for h/w based transcoders. You can get realtime
> software encoding from a cheap tv capture card but quality is a
significant
> issue. I believe that the CinemaCraft Basic encoder for less than $60
> provides good quality encoding at much higher speed than Tmpgenc. See
> www.visiblelight.com for CCBasic.

You'd think that a hardware based transcoder would be a big selling point
for home DVD. Oh well. For $499 or so you can get one from Canopus that
promises "near-realtime" DV->MPEG2 conversion.

I'm holding out for Sonic SD-2000.
Anonymous
September 2, 2004 9:19:00 PM

Archived from groups: rec.video.desktop (More info?)

On Wed, 1 Sep 2004 16:50:05 -0500, "Shawn Barnhart"
<usenet@grasslake.net> wrote:

>
>"FLY135" <fly_135(@ hot not not)notmail.com> wrote in message
>news:l4qZc.521$Wv5.123@newsread3.news.atl.earthlink.net...
>>
>> "Funprice" <jhbusscherNOSPAM@freemail.nl> wrote in message
>> news:h6hbj09ccolm1096fk5n6ohov1lnahuc0m@4ax.com...
>> > It could be the fact that implementing mpeg-2 technology in software
>> > isn't patented and in hardware it is. The Hauppage 250+350 cards are
>> > noticable higher in price than tv-cards without mpeg-2 hardware
>> > encoding! This maybe backups my assumption of a price to pay for using
>> > mpeg-2 technology.
>>
>> Both software and hardware implementations of MPEG encoding are subject to
>> MPEG licensing. The license fee is small and has little to do with the
>> price of hardware encoders. The pricing is entirely a supply and demand
>> issue. There is little demand for realtime h/w encoders in PC form
>factor,
>> and even less demand for h/w based transcoders. You can get realtime
>> software encoding from a cheap tv capture card but quality is a
>significant
>> issue. I believe that the CinemaCraft Basic encoder for less than $60
>> provides good quality encoding at much higher speed than Tmpgenc. See
>> www.visiblelight.com for CCBasic.
>
>You'd think that a hardware based transcoder would be a big selling point
>for home DVD. Oh well. For $499 or so you can get one from Canopus that
>promises "near-realtime" DV->MPEG2 conversion.
>
>I'm holding out for Sonic SD-2000.

Okay, but is your question answered by now?
Anonymous
September 3, 2004 2:14:11 PM

Archived from groups: rec.video.desktop (More info?)

"Funprice" <jhbusscherNOSPAM@freemail.nl> wrote in message

>
> Okay, but is your question answered by now?

I guess. It still strikes me as reasonable that an inexpensive transcoding
board capable of at least 2x real time encoding should be possible, given
the proliferation of inexpensive devices relying on hardware MPEG2 encoding.

Canopus apparently does make a direct-to-MPEG2 capture board that can
transcode AVI to MPEG2 "at near real time" for $600. "At near real time"
being something less than interesting.
Anonymous
September 4, 2004 3:43:45 AM

Archived from groups: rec.video.desktop (More info?)

On Fri, 3 Sep 2004 10:14:11 -0500, "Shawn Barnhart"
<usenet@grasslake.net> wrote:

>
>"Funprice" <jhbusscherNOSPAM@freemail.nl> wrote in message
>
>>
>> Okay, but is your question answered by now?
>
>I guess. It still strikes me as reasonable that an inexpensive transcoding
>board capable of at least 2x real time encoding should be possible, given
>the proliferation of inexpensive devices relying on hardware MPEG2 encoding.
>
>Canopus apparently does make a direct-to-MPEG2 capture board that can
>transcode AVI to MPEG2 "at near real time" for $600. "At near real time"
>being something less than interesting.
>

Probably only equipment that can be mass-produced and sold to the
masses is relatively cheap. Special equipment will always be built,
and why not try to get a high price for it. You're just unlucky to be
interested in something not many others are, it seeems to me.
!