Digital video on a 4200 RPM laptop HD???

[Chuck Antinoro]

OK, I'm not trying to be a pest, intentionally anyway. BUT, is capturing digital video on a 4200 RPM laptop HD doable, or will I drop a ton of frames? I would only be doing digital transferring (firewire) and no analog stuff. I guess I'm hoping that firewire transfer to HD is an exact bit for bit transfer and the low RPMs would just affect access time. Most of the laptops I see out there are equipped with HDs plenty big enough for my needs (100GB) but most are 4200 RPM. Or should I sacrafice inital HD size and get an external firewire or USB 2.0 HD that is 7200 or 10000 RPM? I'm not looking to make a living as a professional editor, just the occasional vaction video, but don't want a herky jerky video due to tons of dropped frames.

Or should I just forget the whole idea of digital video on a laptop.

Any suggestions or advice is greatly appreciated.

[BJ_M]

if capturing dv -- you will be ok .. though faster is better .

ive never had a problem w/ laptop 4200rpm drives and dv

[edDV]

I hope this isn't too detailed but some basic concepts are being missed concerning DV streams ...


"I would only be doing digital transferring (firewire) and no analog stuff."

IEEE-1394 DV streams are always 25Mb/s fixed. The stream comes from the camera section, the MiniDV tape or as pass-through DV streams from digitized analog inputs. It doesn't matter which. The stream is recorded to a DV-AVI file on the notebook HDD. The HDD must be capable of keeping up with the fixed stream under all conditions or you will get a glitch.

In contrast, if you were transferring a DV-AVI file between drives through a network, even one including a IEEE-1394 link hooked up for MS networking, the DV-AVI would be handled as a file and it would be transferred normally at network speeds.


"Or should I sacrafice inital HD size and get an external firewire or USB 2.0 HD that is 7200 or 10000 RPM?"

To record a DV stream to an external HDD, the stream would still run 25Mb/s constant regardless of HDD RPM. The external drive would just need to keep up or there will be a glitch.

If you were using USB2 to transfer a file from the external HDD to the laptop HDD, the transfer would be limited to the speed of the slower drive or the USB2 connection but the file would make it without errors. This is not necessarily the case for fixed stream transfers.

I've had many tell me that recording a 25Mb/s DV stream to an external notebook drive (connected IEEE-1394 or USB2) can be problematic and their recomentdation has been to record the stream first to a DV-AVI file on the local drive and then copy the file to the external drive. This may be old news but the warning stands.

The important issue is whether the internal drive can robustly handle a fixed 25Mb/s DV stream under all conditions including shared Windows and background process interupts.


"I'm not looking to make a living as a professional editor, just the occasional vaction video, but don't want a herky jerky video due to tons of dropped frames."

I have no direct DV editing experience with 4200 RPM notebook drives. 5400 RPM drives have worked for me. You should research that the 4200 RPM drive will work before you commit. That is your weakest link to smooth DV video.

[Richard_G]

IEEE-1394 DV streams are always 25Mb/s fixed.

I don't know where you got this idea from, DV is 13Gb per hour which works out to around 3.8Mb per second. Any hard drive that can't achieve this as a sustained data rate is either very old, very slow or very fragmented. You should be able to transfer DV over Firewire without any problems.

[edDV]

Bytes and bits

25Mb/s x60s/min x60min/hr = 90Gbits / 8 bits/byte = 11.25 GB per hour.

While the video stream is 25 megabits per second, the actual IEEE-1394 data stream is 36Mb/sec after you add in audio, subcode (including timecode), insert and track Information (ITI), and error correction.

The way DV actually is recorded to tape works out to around 30Mb/sec or 13.5GB per hour.

True, it is a sustained rate well within modern hard disk drive and controller capability, but is a much higher rate than other external data communication I/O for a typical PC. Normal HDD to HDD or Ethernet data transfer includes a mechanism to replace corrupt packets by requesting replacement data from the source. This doesn't happen with a continous data stream from a DV camcorder. A frame lost is a frame lost.

There is minimal buffering on a fixed rate external data stream and any interruption to sustained HDD write will result in a glitch or lost frames. A notebook has a single hard drive that is shared with the OS and other running processes. If any of those take interrupt priority on the drive the internal buffering can be used up fast and a frame dropout will occur. When you consider that the data must arrive with no frames lost for an hour or more of transfer, you begin to see the problem.

For a desktop PC the problem can solved by using redundant drives and hard disk controllers, one for the OS and background processes and the other dedicated to sustained video transfer. A second external drive on a notebook is a different animal. A notebook external HDD sits on a USB or IEEE-1394 connection that adds additional probability for data loss regardless of the HDD's sustained data capability.

[BJ_M]

an external drive on USB also usually causes a lot of cpu activity and would not be able to sustain long periods of transfer ...

[tmh]

I have a Prostar (Clevo makes - like Alienware,Sager) "desktop replacement" type laptop with a P4 2.4 processor (desktop processor) and so far I've had no problems with DV (Sony Camcorder) capture through a firewire card to an external USB 2.0 Maxtor 160gb 8mb buffer HD conected through the same card. I close all other apps I can see running. The laptop does get a bit hot, but it's made for gaming so it's within "normal" temps, I think.

So far, I've done most of my editing on the laptop, but have encoded on my desktop because I normally do this while I sleep and I'm nervous about frying my laptop during encoding. Question: If it transfers fine from camcorder (more than 1 hr so far), am I likely OK with encoding on the laptop? Does encoding normally take more or less CPU power as compared to transfer via firewire? Will the usb 2.0 external harddrive be an issue when encoding?

Thanks. I want to be as portable as possible with my setup so I look forward to hearing others' experience with this.

[edDV]

quote tmh

"I have a Prostar (Clevo makes - like Alienware,Sager) "desktop replacement" type laptop with a P4 2.4 processor (desktop processor) and so far I've had no problems with DV (Sony Camcorder) capture through a firewire card to an external USB 2.0 Maxtor 160gb 8mb buffer HD conected through the same card. I close all other apps I can see running."

I'm not saying it can't be done, I'm just showing how easy it would be to loose frames as the margin of error is dropped with a slower drive and the vulnerability of the process to OS and background process interrupts.

Those of us who have been capturing and transferring video since the 386/486 CPU days using early ATA and SCSI drives have experienced the delicate balance of the process first hand. In those days RAID was a requirement both to achieve sustained transfer rates and to isolate CPU interrupt activity. It wasn’t until the Pentium II machines added bus mastering to the PCI bus that true capture isolation was achieved. PCI Bus mastering allowed capture cards and disk controllers to operate in isolation from the CPU allowing much more reliable video capture.

The DV standard has cut the required sustained data transfer rate almost in half compared to MJPEG for similar quality video (i.e. 4 vs. 8 MB/s capture). Likewise modern ATA-100 disk controller and drive technology has increased single disk performance for sustained transfer. However, even fast single drive desktop machines can experience frame loss from OS and background process activity.

See this recent forum thread “Dropped frames no matter what” https://www.videohelp.com/forum/viewtopic.php?t=249078&highlight=

A single drive notebook operates much closer to the safety margin edge. It can work well if the process is properly monitored and managed.


“The laptop does get a bit hot, but it's made for gaming so it's within "normal" temps, I think.

So far, I've done most of my editing on the laptop, but have encoded on my desktop because I normally do this while I sleep and I'm nervous about frying my laptop during encoding. Question: If it transfers fine from camcorder (more than 1 hr so far), am I likely OK with encoding on the laptop? Does encoding normally take more or less CPU power as compared to transfer via firewire? "

Transfer is a separate process from CPU intensive encoding. While transferring (“Capturing”) the DV stream it is good practice to minimize all other CPU tasks.

Encoding a Mpeg is an intense CPU process that will cause your laptop to run at 100% CPU speed for extensive periods. Heat dissipation will be a problem depending on notebook design. Most laptops get extremely hot during encoding. When my PIII 750MHz VAIO notebook is rendering for long periods I set a small desk fan to blow across the keyboard to aide cooling.


"Will the usb 2.0 external harddrive be an issue when encoding?”

No because the data transfer is done by normal disk control. There will be no "lost frames" because encoding will be paused until the needed data transfer is complete with no errors.

[tmh]

Thanks for the info, ed. I'll just get a fire extinguisher ready before I encode from the lappy.

[Richard_G]

Bytes and bits

25Mb/s x60s/min x60min/hr = 90Gbits / 8 bits/byte = 11.25 GB per hour.

While the video stream is 25 megabits per second, the actual IEEE-1394 data stream is 36Mb/sec after you add in audio, subcode (including timecode), insert and track Information (ITI), and error correction.

Sorry for that, a bit of confusion between bits and bytes. I totally agree that the DV stream is in bits but the hard disk sustained transfer rate is normally quoted in bytes (or Mbytes) per second. Hence it is normally quoted as 3.8 Mbytes per second.

When I first started working with video, everyone's advice was that you should really be using SCSII hard drives in a RAID array. Being a cheapskate, I was using a 350 MHz AMD K6 on a motherboard with ATA66 IDE controller. The maximum sustained data rate on the hard drives in those days was 66 Mbytes per second burst (if you were lucky) and anything from 10 to 18 sustained rate. As DV hadn't been invented, everything was analogue captures and it was a carefull juggling act between decent quality and not dropping frames (because the data was arriving faster than you could get it on the hard disk). I used to capture in 640 x 480 and use the PicVideo MJPEG codec which required a rate of around 12 Mbytes per second! The slightest hiccup or disk fragmentation and the whole lot was ruined.

Over the years, hard drives and controllers have got much faster and, with DV, the requirement is much lower. So now we have the situation where anything remotely modern is capable of doing the job, including 4200 rpm laptop drives!