DA Lite HDO

Troy, we don't yet have a planto sell G2 yet for 2 reasons:
  • I have tried the demo unit from factory and I was not impressed. The speed was still very slow. I have no idea why Jobo claims 16M/s. That sounds a typo to me
  • The price is not reasonable. I don't think it is anything better than Tripper but the price is way higher than Tripper. I just can't justify that.
On the other hand, we may sell Image Tank III, which is a USB 2.0 device with multiple card slots. I will get the evaluation unit in 2 weeks.
This is not impressive by any standard.
  • The transfer speed is is about half of Tripper, or 40% of Super
Digibin, or same as Image Tank, a 3 year old product
  • It is simply impossible to do bitwise transfer verification with
20 seconds. By common sense, that process should be longer than the
transfer time, or 5 minutes
Charles:

Do you have the new G2 Image Tank yet? R those transfer rates I
mentioned (Jobo web site source) accurate? 16 MB/s seems faster
than the card itself--there must be a bottleneck somewhere in the
system.

Best,

Troy
--
Charles Cheng
[email protected]
 
  • I have tried the demo unit from factory and I was not impressed.
The speed was still very slow. I have no idea why Jobo claims
16M/s. That sounds a typo to me
  • The price is not reasonable. I don't think it is anything better
than Tripper but the price is way higher than Tripper. I just can't
justify that.

On the other hand, we may sell Image Tank III, which is a USB 2.0
device with multiple card slots. I will get the evaluation unit in
2 weeks.
Charles Cheng wrote:
Thanks Charles! Good to know. I'm happy with the tripper with the exception of poor battery life and the lack of a backlit LCD. It would be nice to be able to delete files too, but it's not that big of a deal as I can do that through my PDA and host card.

Thx,

Troy
 
I agree: not impressive at all. I would have expected the DA Lite to be on par or faster than the other devices already out there, but at least it's better than the 0.64 MB/s they told me (I misquoted by .03 MB/s in my earlier post).

CF to HD for MicroDrive:
Super DigiBin: 2.64 MB/s
Tripper: 1.40 MB/s
DA Lite: 1.2 MB/s

It's definitely MUCH slower than the Super DigiBin, but actually only 0.2 MB/s slower than the Tripper when transferring from a MicroDrive. I’ll be interested to see what the transfer rates are like for a high-speed CF card. Might be faster, might not.
This is not impressive by any standard.
  • The transfer speed is is about half of Tripper, or 40% of Super
Digibin, or same as Image Tank, a 3 year old product
  • It is simply impossible to do bitwise transfer verification with
20 seconds. By common sense, that process should be longer than the
transfer time, or 5 minutes
 
  • It is simply impossible to do bitwise transfer verification with
20 seconds. By common sense, that process should be longer than the
transfer time, or 5 minutes
Yes, but a bitwise comparison is not the only way to accurately verify the fidelity of the transferred files. A very accurate verification can be quickly performed using cyclic redundancy checking (checksum for short), which exploits some binary arithmetic tricks to make the required calculations very fast. This is the de facto method for maintaing the integrity of computer files systems and also all sorts of other data transfers, such as modem and Internet protocols. It would not be very difficult to implement in products such as Nixvue, ImageBank, Tripper, etc. In fact, I am rather surprised it's not a standard feature.

That said, Nixvue's Kerry Reuer has previously stated in this forum that the DA products perform a bitwise verification, and that takes as long as the original file transfer. So I am curious now as to what is really going on; if they are doing a checksum comparison I think that is an excellent answer, since it is accurate and fast (saves a lot of battery power). And it definitely would place the DA series of products in the top tier in terms of maintaining data integrity. I know Kerry has been reading and posting to this thread and I hope he comments soon.

Andrew
 
Andrew, nice comment. I understand your point. It is fair to claim the actual comparsion can be very short. However, I don't think the bottle neck is at the comparsion. It is the data access time that causes the slowness. It has to get the data from the CF card first, then does the comparision. Unless it has very advanced way of fetching the right bits from memory card, this process will take the same amount of time as reading the whole card.
  • It is simply impossible to do bitwise transfer verification with
20 seconds. By common sense, that process should be longer than the
transfer time, or 5 minutes
Yes, but a bitwise comparison is not the only way to accurately
verify the fidelity of the transferred files. A very accurate
verification can be quickly performed using cyclic redundancy
checking (checksum for short), which exploits some binary
arithmetic tricks to make the required calculations very fast. This
is the de facto method for maintaing the integrity of computer
files systems and also all sorts of other data transfers, such as
modem and Internet protocols. It would not be very difficult to
implement in products such as Nixvue, ImageBank, Tripper, etc. In
fact, I am rather surprised it's not a standard feature.

That said, Nixvue's Kerry Reuer has previously stated in this forum
that the DA products perform a bitwise verification, and that takes
as long as the original file transfer. So I am curious now as to
what is really going on; if they are doing a checksum comparison I
think that is an excellent answer, since it is accurate and fast
(saves a lot of battery power). And it definitely would place the
DA series of products in the top tier in terms of maintaining data
integrity. I know Kerry has been reading and posting to this thread
and I hope he comments soon.

Andrew
--
Charles Cheng
[email protected]
 
Andrew, nice comment. I understand your point. It is fair to claim
the actual comparsion can be very short. However, I don't think the
bottle neck is at the comparsion. It is the data access time that
causes the slowness. It has to get the data from the CF card first,
then does the comparision. Unless it has very advanced way of
fetching the right bits from memory card, this process will take
the same amount of time as reading the whole card.
Hi Charles, actually I was just doing some calculations on this!

The VxWorks operating system used by Nixvue is very efficient, and likely gets to a file's checksum with a small number of 32 bit reads from the CF media per file (possibly only one read). On a microdrive the average latency plus seek time is about 20.33 mSec, and the data transfer time for 32 bits is way less than that, so for a rough estimate I ignored data transfer time. If more than one read needs to be performed, the access time is going to be much smaller, since the disk heads are already on the right data track. Latency - who knows, it is the luck of the draw, but probably less than the average. Since this was a very quick and dirty calculation, I just ignored the time to make any additional reads.

Access to the checksum of the copied file is actually not much faster, on the Fujitsu drive in my DA Lite (it lets me look up the model number), I see the average seek plus latency time is 19.14 mSec, and the time to read the data even more inconsequential than on the microdrive. (It may sound strange that the Fujitsu is not much faster than the microdive, but remember it is 40GB vs. 1GB for the microdrive, so it has a lot more magnetic real estate to negotiate).

Therefore a very rough guesstimate is that it takes roughly 20.33 + 19.14 = approx. 40 mSec to get the data to do a checksum comparison, which should only take a few hundred microseconds to compute. Maybe a little more since the processor in the DA is very tiny.

Now - it happens I have just done a fresh transfer from my microdrive to the DA Lite, and this time I used a stopwatch. The verification for 139 files took 8 seconds, which is 57 mSec per file! I say that's in the ballpark!

Andrew
 
Miles -

If I may respond - Nixvue has been very hesitant to post transfer speeds since in real life situations there can be large differences - If a product has a stated speed of transfer and maker claims it for all types of media then they are being less than honest in publishing the figures unless they claim it is the minimum transfer speed - Also to make comparisions between different units one needs to use the exact same media with the exact same files otherwise you are comparing apples to oranges.

Let me explain - the speed of the controller on the card (or in the adapter if one used) will have a major effect on the speed of the transfer - A recent test I did on the DA Lite showed how much - I took a full 256MB generic CF card (PQI controller) and got a transfer rate of .62 MB sec. I then took a Delkin CF card using a 3S controller and got a transfer rate of 1.2MB sec. Almost double the speed. Another factor is the type of media - SmartMedia can be very slow, SD is fast, Memory stick is so-so.

For the same reason we are hesitant to post battery life since it can vary depending on the speed of the card - the longer it takes to copy the longer the HD spins and uses power. The number of transfers will also the battery life since there is a large current draw each time the HD is spun up. The copying of small cards many times can actually mean less data is transfered than if the user has a large 512MB card and copies it only a few times.

Hope this helps -

Kerry Reuer, Dir Bus Dev., Nixvue Systems
There was one important (to me) spec for the DA Lite that I
couldn't find on the Nixvue website: CF to HD speed. I sent an
email to their sales department and they responded that it was 0.61
MB/s. Can anyone confirm this? Seems pretty slow to me,
especially when Tripper and Super DigiBin are getting around 2 MB/s.
 
Kerry, thanks for the good explanation.

Since a lot of people often refer the test I did, I just want to make some comments here: I did use the exactly same testing conditions. In fact, I used the most common setup found in digital photographers:

Nikon D100
Ridata Professional 20X CF card (the fastest back then)
Large image files (at 17MB per picture)

You are right, it is not a good idea for manufacturers to publish the transfer rate since it can be misleading. In fact, the manufacturers of Super Digibin and Trippers don't do that either. However, most customers (at least those from this forum) reported very similar results, which makes think the bottle neck is NOT the card, even though I have only tested the high speed card.
If I may respond - Nixvue has been very hesitant to post transfer
speeds since in real life situations there can be large differences
  • If a product has a stated speed of transfer and maker claims it
for all types of media then they are being less than honest in
publishing the figures unless they claim it is the minimum transfer
speed - Also to make comparisions between different units one needs
to use the exact same media with the exact same files otherwise you
are comparing apples to oranges.

Let me explain - the speed of the controller on the card (or in the
adapter if one used) will have a major effect on the speed of the
transfer - A recent test I did on the DA Lite showed how much - I
took a full 256MB generic CF card (PQI controller) and got a
transfer rate of .62 MB sec. I then took a Delkin CF card using a
3S controller and got a transfer rate of 1.2MB sec. Almost double
the speed. Another factor is the type of media - SmartMedia can be
very slow, SD is fast, Memory stick is so-so.


For the same reason we are hesitant to post battery life since it
can vary depending on the speed of the card - the longer it takes
to copy the longer the HD spins and uses power. The number of
transfers will also the battery life since there is a large current
draw each time the HD is spun up. The copying of small cards many
times can actually mean less data is transfered than if the user
has a large 512MB card and copies it only a few times.

Hope this helps -

Kerry Reuer, Dir Bus Dev., Nixvue Systems
There was one important (to me) spec for the DA Lite that I
couldn't find on the Nixvue website: CF to HD speed. I sent an
email to their sales department and they responded that it was 0.61
MB/s. Can anyone confirm this? Seems pretty slow to me,
especially when Tripper and Super DigiBin are getting around 2 MB/s.
--
Charles Cheng
[email protected]
 
Andrew, Thanks for the very detalied analysis. I believe this is very valuable to folks on this forum.
Andrew, nice comment. I understand your point. It is fair to claim
the actual comparsion can be very short. However, I don't think the
bottle neck is at the comparsion. It is the data access time that
causes the slowness. It has to get the data from the CF card first,
then does the comparision. Unless it has very advanced way of
fetching the right bits from memory card, this process will take
the same amount of time as reading the whole card.
Hi Charles, actually I was just doing some calculations on this!

The VxWorks operating system used by Nixvue is very efficient, and
likely gets to a file's checksum with a small number of 32 bit
reads from the CF media per file (possibly only one read). On a
microdrive the average latency plus seek time is about 20.33 mSec,
and the data transfer time for 32 bits is way less than that, so
for a rough estimate I ignored data transfer time. If more than one
read needs to be performed, the access time is going to be much
smaller, since the disk heads are already on the right data track.
Latency - who knows, it is the luck of the draw, but probably less
than the average. Since this was a very quick and dirty
calculation, I just ignored the time to make any additional reads.

Access to the checksum of the copied file is actually not much
faster, on the Fujitsu drive in my DA Lite (it lets me look up the
model number), I see the average seek plus latency time is 19.14
mSec, and the time to read the data even more inconsequential than
on the microdrive. (It may sound strange that the Fujitsu is not
much faster than the microdive, but remember it is 40GB vs. 1GB for
the microdrive, so it has a lot more magnetic real estate to
negotiate).

Therefore a very rough guesstimate is that it takes roughly 20.33 +
19.14 = approx. 40 mSec to get the data to do a checksum
comparison, which should only take a few hundred microseconds to
compute. Maybe a little more since the processor in the DA is very
tiny.

Now - it happens I have just done a fresh transfer from my
microdrive to the DA Lite, and this time I used a stopwatch. The
verification for 139 files took 8 seconds, which is 57 mSec per
file! I say that's in the ballpark!

Andrew
--
Charles Cheng
[email protected]
 
Kerry,

Thanks for helping me understand the manufacturer’s side of things. True, the media used will have a drastic effect on the speed of the transfer, and maybe there are many consumers who would not understand that.

However, once you start using high-speed CF cards, the card doesn’t seem to be the bottleneck. I can expect 5 MB/s read speed from my 24x Lexar CF in a firewire card reader, but there are no portable storage devices that come anywhere near that. It just seems to me that there comes a point where the device is limiting the transfer speed, and that is what I would like to see posted. I don’t expect it to be a guaranteed transfer speed.

Additionally, it seems to me that a 5 MB/s CF to HD transfer speed would greatly increase the amount of data that could be stored on a single battery charge. So faster card reading both takes less time and increases the amount of data that can be stored on a single charge, making it doubly important. That’s why I want to know about it before I buy.

Just out of curiosity, is there any reason portable storage devices aren’t able to match the speed of a firewire/USB2 card reader?

Thanks,
Miles
If I may respond - Nixvue has been very hesitant to post transfer
speeds since in real life situations there can be large differences
  • If a product has a stated speed of transfer and maker claims it
for all types of media then they are being less than honest in
publishing the figures unless they claim it is the minimum transfer
speed - Also to make comparisions between different units one needs
to use the exact same media with the exact same files otherwise you
are comparing apples to oranges.

Let me explain - the speed of the controller on the card (or in the
adapter if one used) will have a major effect on the speed of the
transfer - A recent test I did on the DA Lite showed how much - I
took a full 256MB generic CF card (PQI controller) and got a
transfer rate of .62 MB sec. I then took a Delkin CF card using a
3S controller and got a transfer rate of 1.2MB sec. Almost double
the speed. Another factor is the type of media - SmartMedia can be
very slow, SD is fast, Memory stick is so-so.


For the same reason we are hesitant to post battery life since it
can vary depending on the speed of the card - the longer it takes
to copy the longer the HD spins and uses power. The number of
transfers will also the battery life since there is a large current
draw each time the HD is spun up. The copying of small cards many
times can actually mean less data is transfered than if the user
has a large 512MB card and copies it only a few times.

Hope this helps -

Kerry Reuer, Dir Bus Dev., Nixvue Systems
There was one important (to me) spec for the DA Lite that I
couldn't find on the Nixvue website: CF to HD speed. I sent an
email to their sales department and they responded that it was 0.61
MB/s. Can anyone confirm this? Seems pretty slow to me,
especially when Tripper and Super DigiBin are getting around 2 MB/s.
 

Keyboard shortcuts

Back
Top