HD-P2 arrived, one problem
Stonewall Ballard
work out very nicely for me. It's an impressive device. Very usable.
The only problem I've found is that with pre-record turned on, and with
no existing takes, I always get a "buffer overrun" error when I hit
"record" if there's even a tiny amount of pre-recording. If there are
existing takes, the probability of this overrun error goes down with
the number of takes.
I've tried this with two different 2GB CF carts, but the same thing
happens on both. The media tester in the HD-P2 says that both CF carts
are fine at all speeds. I'm running the latest 1.03 firmware.
Anyone else seen this?
- Stoney
Leo Sreto
Could you please elaborate on that? I simply don't understand what you
mean... :-( Is this happening only if your CF is empty or...?
Do you have any other comments on the device? What about its sturdyness? Is
it good "qualityfeeling"?
And most importently: how are the pre-amps? And noise?
I'm very interested in this machine but I'm in Denmark and don't know when
it will ship here...
Leo
"Stonewall Ballard" <sb.n...@sb.org> wrote in message
news:2005120814473116807-sbnospam@sborg...
David Satz
HD-P2. It was shown at the recent AES Convention in New York, which I
attended, and one of the many obnoxious things I do at such shows is to
carry with me a pair of miniature devices which test the phantom
powering capabilities of a microphone input. I asked permission of the
nice people at the Tascam booth and tried these testers on the HD-P2.
It passed the test very well.
What I should explain, however, is that these testers only draw about
4.5 - 5 mA per input, not the full 10 mA which the IEC standard permits
for 48 Volt phantom powering. So this doesn't prove that, for example,
Earthworks condenser microphones would work with the HD-P2 (at least
some of their models really require the full 10 mA). But the available
current is adequate for any 48-Volt phantom powered Neumann or Schoeps
microphone, as well as many other types. That's welcome news, since
many items of portable equipment have inadequate phantom supplies.
--best regards
Stonewall Ballard
the inputs and keeping up to the last 15 seconds of audio in a circular
buffer. When you hit Record, it dumps what it has in its pre-record
buffer onto the CF card, and continues from there. This is intended to
keep you from hitting record too late and missing the beginning of
something.
I haven't determined what state is necessary to cause the problem.
Clearly an empty CF card will do it. After recording about 30 seconds,
pre-record works most of the time. I think it's a firmware bug. I've
asked Tascam via email. We'll see what they say.
The HD-P2 has excellent fit-and-finish. It feels like a lightweight but
sturdy pro device. Rubber edge on the record gain control. Very clear
LCD, although small font - I need my reading glasses to use it. The
menus and switches are very well-designed.
I have not yet had a chance to see just how good the mic pre's are.
I've only had time to go over it operationally, and still haven't
looked at the timecode or video sync functions.
Can't you buy one here in the US? Or is that a reverse Gray Market?
- Stoney
Tom Jancauskas
There is a software update available at the Tascam site. Maybe this is
fixed with the update?
---
Tom Jancauskas
Imedia
Stonewall Ballard
I have the V1.03 update installed. The HD-P2 came with V1.02.
- Stoney
Leo Sreto
Thanks for you reply. Now I understand the problem... I can see why this is
happening with an empty CF, but you say it is happening even with 30 sec. of
recording on it...? Sounds like a software issue to me too. Does it happens
when you have, lets say, 10 minutes og recording on the CF?
I downloaded the Owners Manual last night form tascam.com and read every
word. Couldn't sleep... and I must say it looks like exactly what I'm
looking for. I have this 10 years old HHB Pro DAT and it has been my
faithfull compagnion for many field recordings through the years. But it's a
bit tired now, and I think the HD-P2 is just the replacement I'm looking for
:-)
I would appreciate any observations/issues from you in the future. I'm very
tempted to order one right away if only I knew it is as good as it sounds.
BTW. I cannot order it from US. I would have to pay shipping, toll, tax and
moms [a special tax we have in Denmark]
regards,
Leo
"Stonewall Ballard" <sb.n...@sb.org> wrote in message
news:2005120814473116807-sbnospam@sborg...
Leo Sreto
Thank you for you info!
I have 2 AKG C 460B mics. that I use for recording acoustic/classical
guitar. And I've used them with my HHB Pro portable DAT with exellent sound.
Do you have a clue if those AKG's is going to work with the HD-P2?
I would appreciate any comment from you.
regards,
Leo
"David Satz" <DS...@msn.com> wrote in message
news:1134077835....@g43g2000cwa.googlegroups.com...
Mike Rivers
Leo Sreto wrote:
> Thanks for you reply. Now I understand the problem... I can see why this is
> happening with an empty CF, but you say it is happening even with 30 sec. of
> recording on it...? Sounds like a software issue to me too.
I understand the situation. Why do you think it's happening with a
blank flash card? Are they harder to write when they're empty? Is there
a "formatting" issue where it's trying to do some sort of
initialization when it's empty, and that takes too long for the buffer
to dump?
Doesn't anybody get these things right before they're released?
Peter Larsen
> I would appreciate any observations/issues from you in the
> future. I'm very tempted to order one right away if only
> I knew it is as good as it sounds BTW. I cannot order it
> from US.
Sure you can, whether it is wise is a different question. Sometimes
shopping in the US via the web can be done with advantage, sometimes
not. The largest drawback is servicing if any is required.
I would have to pay shipping, toll, tax and
> moms [a special tax we have in Denmark]
They call it [v]alue [a)dded (t)ax, ie. VAT abroad and overseas. For a
rule of thumb: USD price times 10 equals DKK price. It seems to be in
the same general price range as the Fostex FR2 that SC Sound imports.
And that one is a hard act to follow .... perhaps someone out there has
a chance to compare?
> Leo
Kind regards
Peter Larsen
--
*******************************************
* My site is at: http://www.muyiovatki.dk *
*******************************************
Leo Sreto
Well, what I meant was that it couldn't pay off ordering it from US. I know
i *can*, but I suspect I end up paying about the same as if I bought it in
DK. And I would like to be able to have it serviced in the future. I just
contacted KinoVox.dk to hear about delivering.
Best regards,
Leo
"Peter Larsen" <SPAMSHIEL...@mail.tele.dk> wrote in message
news:4399983D...@mail.tele.dk...
David Satz
required at most 1 mA from a phantom power supply. That's rather
typical for older, simpler FET amplifier circuits that have output
transformers. The powering really shouldn't be any problem at all.
The sensitivity of the C 460 series with the most common capsule types
is about 8 mV/Pa--not unusually high. I didn't measure the input
overload point of the HD-P2's preamps, but the owner's manual on
Tascam's Web site says that their maximum input level (with the trim
control set to minimum sensitivity) is +13.8 dBu, which is nearly 3.8
Volts. That's considerably more signal than those microphones can put
out, and would require a sound pressure level of around 147 dB, which
is more than they can handle without attenuation. So I don't think that
preamp overload will be a problem, either, as long as you set the trim
control properly.
--best regards
Stonewall Ballard
almost certainly a firmware bug.
The problem in a nutshell is that when you hit the record button, the
recording's start time is the end of the previous recording if
auto-append is on, or is the current time if auto-append is off.
Anything in the pre-record buffer is set to start earlier than the
start time of the recording, so if that would result in the audio
clip's start time being negative, the recording is aborted with the
buffer overrun error.
The tech tried one with V1.02 firmware, and it failed too, but with a
different error message. One with V1.03 firmware gave the same error as
mine.
The net effect of this is that you can't use pre-record and auto-append
on an empty project. If you use pre-record at all, you must ensure that
the recording starts at a high enough time to accomodate the pre-record.
I was surprised to see this elaborate time system in the HD-P2, but I
guess that goes along with the timecode and sync capabilities. Once you
understand how it sets the time of the clips, and how it treats all the
clips in a project as positioned along the timeline, comprising one big
virtual recording, it makes sense. Unfortunately, treating pre-record
audio as happening before the record button is pushed (which it is, of
course) and setting its time as starting before the record button is
pushed, leads to this problem.
- Stoney
Mike Rivers
Stonewall Ballard wrote:
> The problem in a nutshell is that when you hit the record button, the
> recording's start time is the end of the previous recording if
> auto-append is on, or is the current time if auto-append is off.
> Anything in the pre-record buffer is set to start earlier than the
> start time of the recording, so if that would result in the audio
> clip's start time being negative, the recording is aborted with the
> buffer overrun error.
Good diagnosis. Thanks. Sounds like they need to adjust the recording
start time (by whichever definition is appropriate) by the amount of
time in the buffer. But best to leave this to the software engineers.
If they know what's happening, it can be fixed.
> I was surprised to see this elaborate time system in the HD-P2, but I
> guess that goes along with the timecode and sync capabilities.
Either that or they needed to know how to time-tag the first recorded
byte so they took it from the current time. Maybe time-tag the buffer
from the current time, and then start the clock again after it's
loaded?
Leo Sreto
Thanks for all your info! Much appreciated!
I'm a musician who likes to record my own, and others performances, so I'm
really glad you took you time to check this out for me and explaining it so
well.
Well, I think I'm one step closer to order a HD-P2. I've downloaded the
Owners manual and I am very impressed with what this device can do...
I would love though to see a full hands-on-review of it. A job for Stonewall
Ballard...? :-)
Kind regards,
Leo
"David Satz" <DS...@msn.com> wrote in message
news:1134146435....@g44g2000cwa.googlegroups.com...
Stonewall Ballard
> It is not so uncommon that firmware needs updates, The first Tascam
> update was there before the product was rleased. A lot of bugs have been
> fixed, a lot more will be fixed later I think.
Right. Interestingly, an engineer on the HD-P2 team called me today to
ask my opinion on how they should resolve this firmware bug. It turns
out to be a real design flaw rather than a simple bug. He said they'd
been arguing about how to fix it, and he wanted an outside opinion. We
ended up talking for over an hour about various issues with the device.
These were mostly annoyances, like that it's hard to clear all the
audio out of a project.
> Before jumping into details, have you connected microphones and tried to
> make recordings? What is your impression, how does it perform?
I have not had a chance to do more than smoke-test it with mics. My
initial impression is that it sounds very good.
I'm going to record a live concert with it on Dec. 18, so I'll have a
much better idea of how it performs then.
- Stoney
Stonewall Ballard
The Tascam engineer that called me today offered several possible
solutions. I liked the one where when you start the pre-record, the
current time advances until the pre-record buffer fills, then it waits
there until you hit record. So the clip will start with the time
corresponding to the pre-recorded audio's beginning, not the time the
record button is pushed. This seems foolproof to me. I don't care what
the time is when the recording starts, and pre-record is disabled in
timecode mode anyway.
- Stoney
--
------------------------------------------------------
sb.n...@sb.org http://stoney.sb.org
Mike Rivers
Chel van Gennip wrote:
> I don't care too much about these small annoyances. I don't see much use
> for e.g. the pre-roll feature.
Again, it's about "you."
Like a number of things we discussed with the Micro Track, if it's a
feature, it should work. Since you start the recorder and then start
playing the piano, it's of no value to you. But if you're hanging out
and doing some casual recording with a bunch of musicicians who are
talking with their instruments in their lap and one just starts picking
something really cool, you'd be thankful for that feature. Or if
someone is doing a long and boring introduction on stage that you don't
want to record and all of a sudden the band starts playing before he's
finished talking, you'd want that feature. There are many users to whom
this would be valuable. It's not a first with this recorder, and it's
saved many a reporter.
I'm happy to hear that TASCAM understands the problem and I'm sure
they'll eventualy get it working. Still, it would have been nice if
there were release notes that said "The pre-record doesn't work
properly." At least it wouldn't come as a surprise to someone who's
checking it out.
Markus Mietling
>Well, I think I'm one step closer to order a HD-P2. I've downloaded the
>Owners manual and I am very impressed with what this device can do...
One word of caution though: Some time ago, the blurb on Tascam's
European web site [1] reported a
Noise Level < –55 dBu (MIC to LINE OUT)
which is apparently even worse than Marantz's PMD671 (which got much
maligned for this very reason).
I would love to be wrong, but I don't expect the HD-P2's mic preamps to
be usable for music recordings, and I certainly wouldn't buy the machine
without having the option to return it for a full refund.
If you do decide to buy it nonetheless, I would be very interested in
your opinion about the preamp quality.
Best,
MM
[1] They have deleted this particular info since then, but I've seen it myself.
Apart from that, it got quoted on RAMPS by Oleg Kaizerman in article
news:43478914$0$50250$892e...@authen.white.readfreenews.net
Leo Sreto
Don't you think it's a typo?
I would be surprised if TASCAM produce a device like HD-P2 and does a
blunder like that. But you really have a piont there and I would be glad to
hear others comment on this. If it's even worse than PDM671, I will put my
money elsewhere...
Leo
"Markus Mietling" <mietlin...@bigfoot.com> wrote in message
news:uiflp1pdtah11oj4f...@4ax.com...
> Leo Sreto in <4399f9e3$0$8771$edfa...@dread14.news.tele.dk>:
>
>>Well, I think I'm one step closer to order a HD-P2. I've downloaded the
>>Owners manual and I am very impressed with what this device can do...
>
> One word of caution though: Some time ago, the blurb on Tascam's
> European web site [1] reported a
>
> Noise Level < -55 dBu (MIC to LINE OUT)
Markus Mietling
>MM,
>
>Don't you think it's a typo?
No, I don't think it's a typo.
If you look at the specs posted by Oleg [1], it seems that he copied and
pasted these from here: http://www.tascam.de/en/hd-p2.html
The figures are identical, text likewise -- the only difference being
that meanwhile, the previously quoted line with that noise figure is
missing. Tascam simply deleted this bit of information.
This would be a very unusual way to correct a typo: Instead, most other
manufacturers would substitute the incorrect figure with the correct
one, at least that's what I'm thinking.
Be aware that all of this is mere speculation on my side, since I don't
know the thing: I "know" the HD-P2 only from what I read on the net
until know. AFAIC I'd just be wary of buying a cat in the bag for some
1000 Euro, unless I can return it without problems.
Cheers,
MM
[1] news:43478914$0$50250$892e...@authen.white.readfreenews.net
Mike Rivers
Markus Mietling wrote:
> No, I don't think it's a typo.
I don't put a lot of faith in TASCAM's ad copy (or anybody''s, really)
these days. They were running ads for the DVD recorder for months that
had a table of recording time at various sample rates that had a couple
of very wrong numbers. It was eventually corrected.
It could be a copy mistake, a real measurement without knowing what it
really means, or a real measurement mistake. Or it could be real. The
Marantz portable CD recorder has numbers in that ballpark for the mic
inputs. Everybody around here says they have too much noise (based on
the spec sheet) but I don't know anyone who has actually tried one with
real mics and said it had too much noise.
I'll wait for a real review with a real lab test, or at least a
reliable listening test. Right now we're all just guessing and
expecting the worst. It might be better.
Rory
Stonewall Ballard wrote:
>pre-record is disabled >in timecode mode >anyway.
Is it possible to make a recorder that has a pre-record buffer that
works in time code mode?
Rory
FR-2?
Markus Mietling
>Marantz portable CD recorder has numbers in that ballpark for the mic
>inputs. Everybody around here says they have too much noise (based on
>the spec sheet) but I don't know anyone who has actually tried one with
>real mics and said it had too much noise.
There's always Google. I mean, you won't get to know "anyone who has
actually tried" _in person_, but still:
| Think of it like this: you could cook your morning bacon with that
| noise. Remember the days of cassette recorders? You know that constant
| hisssssssssssssssss in the background? That's what it sounds like. So,
| if you're looking for quality pro results don't get the PMD671 if you're
| using external (or internal) microphone inputs. [1]
Doesn't look like "judgement based on the spec sheet" to me.
>I'll wait for a real review with a real lab test, or at least a
>reliable listening test. Right now we're all just guessing and
>expecting the worst. It might be better.
I agree here.
MM
[1] http://4webresults.com/blog/04-05/marantz-pmd660-portable-recording-device
Stonewall Ballard
is not ideal, but it's simple to do. I recorded 30-sec clips with the
input gain set to 0, 5, and 10. 24-bit, 44.1KHz.
I opened the resulting files in Peak to see what the noise looked like,
and into SoundTrack Pro to see the noise in the frequency domain.
The average noise level with the gain all the way up was about 95 db
RMS below 0, digitally. With the gain at 0, the average noise level was
about -100db RMS.
It was clear from the spectrum that it was gettiing interference from
somewhere. There is a radio station close enough that some of my phone
lines pick it up, and at one point, the interference sounded like a
dance beat (listening with 60db of gain). Turning on the 20db pads
and/or setting the input gain to 0 eliminated most of this interference
and produced a very white-noise-like spectrum.
None of this interference exceeded about -90db RMS. Switching between
battery and power supply didn't make any noticable difference.
In addition to the interference, there were noise bursts, typically
lasting about 1/8 sec. in the right channel only. These happend every
few seconds, and were always in the positive direction. The input gain
did not affect these, so I suspect a faulty preamp. Still, this added
only about 6db to the noise while it was happening.
I got similar results, but without any interference pick-up, from my
DigiDesign M-Box. A MicroTrack which I had for a week (too junky for
me) had noise about -60 to -70db down, also tested with shorted inputs.
My conclusion is that the HD-P2 could use better shielding, but it
seems more than adequate to handle live 24-bit recordings.
- Stoney
Mike Rivers
Stonewall Ballard wrote:
> I just tested the HD-P2 with shorted XLR plugs in both mic inputs.
> The average noise level with the gain all the way up was about 95 db
> RMS below 0, digitally. With the gain at 0, the average noise level was
> about -100db RMS.
That's nothing to complain about. "CD quality," fer sure, fer sure.
Markus Mietling
>I just tested the HD-P2 with shorted XLR plugs in both mic inputs.
Thank you for this experiment. Did you simply short the plugs, without
any kind of resistor involved?
>The average noise level with the gain all the way up was about 95 db
>RMS below 0, digitally. With the gain at 0, the average noise level was
>about -100db RMS.
Doesn't sound too bad, at least to my layman's ears. Makes one wonder
how in the world they arrived at the previously reported -55 dBu ...
Regards,
MM
Leo Sreto
Thanks for you testing!
But, in this field I'm an ignoramous :-). I'm a musician, not an
soundingeneer. So, what I would like to know is: can one make high-quality
live recordings with HD-P2? With my HHB DAT I'm used to 100% noisefree
recordings of classical guitar. Do you think I expect that if I buy a
HD-P2...?
Best regards,
Leo
"Stonewall Ballard" <sb.n...@sb.org> wrote in message
news:2005121017295716807-sbnospam@sborg...
Mike Rivers
> what I would like to know is: can one make high-quality
> live recordings with HD-P2? With my HHB DAT I'm used to 100% noisefree
> recordings of classical guitar.
I seriously doubt that youre recordings were 100% noise free (maybe
from a musician's perspective) but if Stoney's measurement technique
was reasonably valid, the HD-P2's noise performance should be on par
with, or better than your HHB DAT.
Scott Dorsey
>Thanks for you testing!
>But, in this field I'm an ignoramous :-). I'm a musician, not an
>soundingeneer. So, what I would like to know is: can one make high-quality
>live recordings with HD-P2? With my HHB DAT I'm used to 100% noisefree
>recordings of classical guitar. Do you think I expect that if I buy a
>HD-P2...?
You are comparing two pieces of equipment that differ in original price
by almost a factor of ten. Why should you be surprised that the HHB
would sound better? For ten times the money it _better_ have higher quality
sound.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."
Mike Rivers
Chel van Gennip wrote:
> That is amazingly good. 95 dB below an input level of -60dBu (at maximum
> gain according to specification) is an EIN of -155 dBu. Normally you
> don't see values above -129 dBu.
You can't really figure EIN like that when you're comparing an analog
input with a digital output since you don't really know what the total
gain from intput to output is. The digital full scale output for a
nominal +4 dBu device is an analog output of somewhere between +20 and
+24 dBu (depending on what digital level they reference the nominal
output to). You can make a better guess at EIN by subtracting another
24 dB from your 155, which is still a bit on the high side. If the
preamp gain is actually 65 dB, that gets you closer to the magic 129
figure for EIN that everybody else comes up with by hook or crook.
As a useful benchmark, the amount of noise present at the output with
no input (shorted) is good to know. Have you tried this yet with your
Micro Track? It woudl be in interesting comparison. Be sure you use
plugs with shielded shells, and tie the "short" to the sleeve. That
will be as good as you can do without a screen room as far as keeping
EMI from interfering with your test.
While we "know" that microphones have a source impedance of a couple of
hundred ohms, a more valid test of the preamp itself is with the input
shorted rather than sourced from a resistor. It takes the noise of the
resistor out of the equation and makes a lower impedance termination
for any common mode noise.
Leo Sreto
| You are comparing two pieces of equipment that differ in original price
| by almost a factor of ten. Why should you be surprised that the HHB
| would sound better? For ten times the money it _better_ have higher
quality
| sound.
| --scott
Scott,
I think you are misstaking what HHB DAT I have. It's the portable HHB Pro.
[long ago discontinued]
In 1996 [10 years ago] the price was 1.600 Euro and the HD-P2 is today 1.300
Euro so I think the comprison is fair...
And I think that the development of soundrecordingdevices contra prices
would give me a better product for the same money today. So I think i's fair
to expect a quality that is the same os better than back in 1996.
regards
Leo
Leo Sreto
You are probably right... But I'm very pleased with my setup and would love
to keep it as it is. But my DAT is old [10 years] and I must have a new
recorder ASAP. Of course it isn't 100% noisefree :-) But I've made
recordings with it that HI-FI enthusiasts/shops use for demonstrations of
HI-FI equiptment...
But as you noticed; I'm *only* a musician and I think our ears are somewhat
subjectiv :-)
Regards,
Leo
"Mike Rivers" <mri...@d-and-d.com> wrote in message
news:1134306644.4...@g14g2000cwa.googlegroups.com...
Mike Rivers
Leo Sreto wrote:
> I'm very pleased with my setup and would love
> to keep it as it is. But my DAT is old [10 years] and I must have a new
> recorder ASAP.
If my porrtable DAT was still working, I would just toss off these
newfangled flash card recorders as a passing fad. As you may have read
in other discussions, I've replaced it functionally with a Nomad
Jukebox 3 which, with the exception of mic inputs (which it essentially
doesn't have, and my DAT did) it makes recordings that sound better and
are quieter than the DAT. But half of the Jukebox feels like a toy and,
while it's never failed me yet, I keep waiting for the day when it will
- either the input jack will fail, something inside will fail, or I'll
not realize that I hadn't held the Record button long enough and it
never started.
If you don't mind the process of tending to the media and you don't
need to make day-long recordings, and you don't mind a few bugs and
flaws that may or may not get fixed, then the HD-P2 would probably be a
suitable replacement for your DAT.
Mike Rivers
> You are comparing two pieces of equipment that differ in original price
> by almost a factor of ten. Why should you be surprised that the HHB
> would sound better? For ten times the money it _better_ have higher quality
> sound.
On the other hand, he's comparing two pieces of equipment that differ
in age by ten years. While there are some asymptopes, at least for a
while, things in this product category get better sounding and cheaper
before the cross a line where they continue to get cheaper at the cost
of sound quality.
But when dealing with a box that's really a system, sometimes parts of
the system improve while others degrade. If the preamps in the PortaDAT
(I think that's what Leo has) are pretty good, I wouldn't expect those
in the HD-P2 to be better, and in fact might not be quite as good. On
the other hand, the A/D converters are probably better if for no other
reason than that they offer 24-bit resolution, which the DAT doesn't.
I think that (unless the mic inputs really suck) what he'll end up with
is something that doesn't sound any worse than what he has now, will
require him to learn a new process of media management (a PITA, IMHO),
will slow him down with new buttons to learn for a while, but will be
new and will probably have a greater no-cost service life than his DAT
currently has.
Hey, he's "only" a musician. ;)
Stonewall Ballard
...
As a useful benchmark, the amount of noise present at the output with
> no input (shorted) is good to know. Have you tried this yet with your
> Micro Track? It woudl be in interesting comparison. Be sure you use
> plugs with shielded shells, and tie the "short" to the sleeve. That
> will be as good as you can do without a screen room as far as keeping
> EMI from interfering with your test.
I no longer have the MicroTrack. In any case, I have no way to measure
noise in the analog domain. It's only by capturing the waveform and
looking at it in an audio editor that I can do this kind of test.
> While we "know" that microphones have a source impedance of a couple of
> hundred ohms, a more valid test of the preamp itself is with the input
> shorted rather than sourced from a resistor. It takes the noise of the
> resistor out of the equation and makes a lower impedance termination
> for any common mode noise.
I believe that's true too. I also swapped the shorting XLR plugs to
ensure that the noise bursts and interference weren't due to poor
solder joints or the like, but swapping made no difference.
I just ran noise tests with the line inputs. Same noise bursts in one
channel, so I think I've got a defective unit. Outside the noise
bursts, it read about -102db RMS below peak.
BTW, this is a kind of strange measurement because it's relative to a
maximum digital peak value, not 0 db RMS. The normalized (peak digital)
residual noise in the HD-P2 has an RMS reading of about -12db. So, you
can probably be justified in adding 12db to these numbers to make them
more realistic as a signal-to-noise ratio.
Mike Rivers
Stonewall Ballard wrote:
> I no longer have the MicroTrack.
I was actaually addressing Chel when I wrote that. I trust he still has
his, and can make a similar measurement.
> I just ran noise tests with the line inputs. Same noise bursts in one
> channel, so I think I've got a defective unit.
Either that, or it's another "design error" like the pre-record buffer.
It's possible that they all do that. Is there a place you could post a
few seconds of a WAV file with the noise burst and without? A listen is
worth a thousand words.
Pooh Bear
Chel van Gennip wrote:
> On Sun, 11 Dec 2005 14:22:34 +0100, Mike Rivers wrote:
>
> > Chel van Gennip wrote:
> >> That is amazingly good. 95 dB below an input level of -60dBu (at
> >> maximum gain according to specification) is an EIN of -155 dBu.
> >> Normally you don't see values above -129 dBu.
> >
> > You can't really figure EIN like that when you're comparing an analog
> > input with a digital output since you don't really know what the total
> > gain from intput to output is.
>
> The test done at maximum gain gave, as Stoney wrote, -95 dB noise relative
> to FS. According to HD-P2 specification, FS corresponds to -60dBu input
> under these (maximun gain) conditions. So Equivalent Input Noise derived
> from these tests is an unbelievable -60 dBu- 95 dB S/N = -155dBu
>
> I don't think I miss anything to meke this conclusion.
There definitely must be. Amplifying devices simply aren't that quiet.
Graham
Stonewall Ballard
> Is there a place you could post a
> few seconds of a WAV file with the noise burst and without? A listen is
> worth a thousand words.
What a good idea! I've uploaded three 30-second clips to my web site:
http://stoney.sb.org/HDP2Gain0.wav
http://stoney.sb.org/HDP2Gain10.wav
http://stoney.sb.org/HDP2-16bit.wav
The first two are about 7.5MB, and the last is 5.2MB.
The first two were recorded with shorting XLR plugs, 24 bit, 44.1KHz.
Gain 0, and Gain 10. Limiter turned off. The last one was recorded in
16-bit, with Gain at 0.
There are little noise bursts in the 16-bit version as well, but there
are long stretches where the maximum data value is 2. Clearly, it's
worth recording in 24-bit with the HD-P2.
Perhaps someone can tell me if my reading of these waves is correct or
not. I'm reporting the RMS signal value from Peak's "Find Peak"
command, and I'm also boosting the gain by 60-70db and looking at the
result on the Elemental Audio Systems IXL Level Meter. Both ways
produce about the same result.
Mike Rivers
Stonewall Ballard wrote:
> What a good idea! I've uploaded three 30-second clips to my web site:
> Perhaps someone can tell me if my reading of these waves is correct or
> not. I'm reporting the RMS signal value from Peak's "Find Peak"
> command, and I'm also boosting the gain by 60-70db and looking at the
> result on the Elemental Audio Systems IXL Level Meter. Both ways
> produce about the same result.
I don't have any software (that I know how to use, anyway) that's smart
enough to read levels that low, but I just listened to the "gain at 10"
clip (I figured that was as bad as it would get) and it souunded quiet
enough so that I wouldn't worry about it at all. I didn't hear anything
that I'd call a noise burst. I amplified it by 60 dB before I could
hear enough to want to do something about, and it sounded like noise,
kind of digital. Probalby it was noise in the converters.
So, my practical conclusion s that the mic preamps are adequately
quiet. How good they sound with microphones remains to be seen.
Pooh Bear
Chel van Gennip wrote:
> On Sun, 11 Dec 2005 22:32:40 +0100, Pooh Bear wrote:
>
> >> The test done at maximum gain gave, as Stoney wrote, -95 dB noise
> >> relative to FS. According to HD-P2 specification, FS corresponds to
> >> -60dBu input under these (maximun gain) conditions. So Equivalent Input
> >> Noise derived from these tests is an unbelievable -60 dBu- 95 dB S/N =
> >> -155dBu
> >>
> >> I don't think I miss anything to meke this conclusion.
> >
> > There definitely must be. Amplifying devices simply aren't that quiet.
>
> Amplifying devices simply can't be that quiet. I don't think the problem
> is in the conclusion. I think there is something strange in the
> measurements.
I'm sure that is the case. Would be nice to know where though.
Graham
Chris Hornbeck
<rabbitsfriend...@hotmail.com> wrote:
>> Amplifying devices simply can't be that quiet. I don't think the problem
>> is in the conclusion. I think there is something strange in the
>> measurements.
>
>I'm sure that is the case. Would be nice to know where though.
The problem seems to me to be in interpretation, and
in two separate places.
First, the tester conflated RMS noise measurements
with peak reference levels; then compounded the
error by correcting in the wrong direction.
Second, the assumption that a -60 dBu reference
level and a (mistaken) 95 dB S/N ratio could be
*added* to give an absolute input noise voltage
with no knowledge of absolute gain before
conversion.
As Mark Twain said " ,and statistics."
Chris Hornbeck
Eric Toline
Re: HD-P2 arrived, one problem
Group: rec.audio.pro Date: Sat, Dec 10, 2005, 7:48am (EST-3) From:
rory...@yahoo.com (Rory)
Sure, the Sound Devices 744T does it very nicely for around $4k.
Eric
Markus Mietling
>What a good idea! I've uploaded three 30-second clips to my web site:
>
>http://stoney.sb.org/HDP2Gain0.wav
>http://stoney.sb.org/HDP2Gain10.wav
>http://stoney.sb.org/HDP2-16bit.wav
>
>Perhaps someone can tell me if my reading of these waves is correct or
>not.
I loaded these files into Adobe Audition 1.0, and used the menu entry
"Analyze -> Statistics..." with the following "RMS Settings":
"O db = FS Sine Wave", "Account for DC" checked, and "Window Width 50ms"
These are the results for HDP2Gain10.wav:
| Left Right
| Min Sample Value: -14.95 -11.64
| Max Sample Value: 17.19 12.98
| Peak Amplitude: -65.73 dB -68.21 dB
| Possibly Clipped: 0 0
| DC Offset: .001 .001
| Minimum RMS Power: -81.24 dB -81.44 dB
| Maximum RMS Power: -77.74 dB -79.34 dB
| Average RMS Power: -80.29 dB -80.76 dB
| Total RMS Power: -80.28 dB -80.76 dB
| Actual Bit Depth: 24 Bits 24 Bits
|
| Using RMS Window of 50 ms
The same for HDP2Gain0.wav:
| Left Right
| Min Sample Value: -2.31 -2.91
| Max Sample Value: 6.74 3.5
| Peak Amplitude: -74.05 dB -80.16 dB
| Possibly Clipped: 0 0
| DC Offset: .001 .001
| Minimum RMS Power: -95.87 dB -94.71 dB
| Maximum RMS Power: -82.66 dB -86.17 dB
| Average RMS Power: -93.36 dB -92.13 dB
| Total RMS Power: -92.96 dB -91.96 dB
| Actual Bit Depth: 24 Bits 24 Bits
|
| Using RMS Window of 50 ms
Finally HDP2-16bit.wav:
| Left Right
| Min Sample Value: -3 -3
| Max Sample Value: 6 3
| Peak Amplitude: -74.41 dB -80.09 dB
| Possibly Clipped: 0 0
| DC Offset: -.001 -.001
| Minimum RMS Power: -94.15 dB -93.51 dB
| Maximum RMS Power: -82.62 dB -86.81 dB
| Average RMS Power: -92.7 dB -91.71 dB
| Total RMS Power: -92.5 dB -91.63 dB
| Actual Bit Depth: 16 Bits 16 Bits
|
| Using RMS Window of 50 ms
MM
Markus Mietling
>> Amplifying devices simply can't be that quiet. I don't think the problem
>> is in the conclusion. I think there is something strange in the
>> measurements.
>
>I'm sure that is the case. Would be nice to know where though.
I would conjecture that the specs Mr Ballard found by shorting the
inputs are not comparable to those found elsewhere, which usually
involve a resistor at the input:
"The standard source impedance is 150 ohms. As unintuitive as it may be,
a plain resistor, hooked up to nothing, generates noise [...] A trick
which unscrupulous manufacturers may use is to spec their mic stage with
the input shorted -- a big no-no, since it does not represent the real
performance of the preamp." (Source: http://www.rane.com/note145.html)
MM
Mike Rivers
Chel van Gennip wrote:
> Well the -60 dBu input level for FS under the described testing conditions
> is from the device specification. This gives an absolute reference for the
> input level. It could be wrong, but it looks like a normal input level for
> a microphone preamp at maximum gain.
I suspect that this is incorrect. -60 dBu for nomonal (not FS)
recording level would be more typical. Not that Mackie is the gold
standard, but it's pretty typical of a contemporary mic input (and I
just happen to have an Onyx sitting next to me). With the mic gain
(trim) fully clockwise and the channel and master set to their unity
gain positions, -60 dBu going in to the mic input gives an output of 0
dBu (which, curiously, indicates at -10 on the VU meters - so much for
calibration). This would represent something around -20 dBFS when
feeding your typical "pro" sound card. I would expect the gain
structure of an integrated device such as a recorder to be similar.
Mike Rivers
Markus Mietling wrote:
> I would conjecture that the specs Mr Ballard found by shorting the
> inputs are not comparable to those found elsewhere, which usually
> involve a resistor at the input:
I'd agree that Stony's measurements don't agree with someone else's.
They never do. However, I don't concur that the reason is due to using
a short rather than a resistor to simulate a source.
Rane may think this is a no-no, but David Josephson (who makes
microphones) believes that a resistor is a no-no, or at least an
inaccurate-inaccurate. The actual difference in broadband RMS noise
between a 150 ohm resistor and a short is probably less than 1 dB, but
that 1 dB can be very important to marketing departments. The noise
spectrum, however, (as apposed to the absolute noise level) will be of
more interest to a preamp designer.
Mike Rivers
Markus Mietling wrote:
> I loaded these files into Adobe Audition 1.0, and used the menu entry
> "Analyze -> Statistics..." with the following "RMS Settings":
> "O db = FS Sine Wave", "Account for DC" checked, and "Window Width 50ms"
>
> These are the results for HDP2Gain10.wav:
So what is your conclusion? Or is that left as an exercise for the
student? I see peak amplitudes below -65 dB (FS assumed?) but the RMS
measurements make no sense to me. Why is this power? How can they
relate power, which implies both voltage and current, to a number of
bits below 24 (OK, 23 and a sign bit)? Perhaps the Audition manual can
shed some light on this.
You opened the can of worms. Can you determine the species of worm that
you let out?
Mike Rivers
Chel van Gennip wrote:
> That would surprise me. The exact specification in the HD-P2 manual is:
> "Mic Input Level: 60 dBu (Trim Max) to 13.8 dBu (Trim Min)"
To me, this only tells me about the overload characteristics. I think
you're reading too much into this bit of markeint. It's telling you
that with the trim set to maximum, the preamp will clip at -60 dBu in,
with the trim set to minimum, it will clip at -13.8 (or is it +13.8?)
dBu. It says nothing about record level.
Now, in a good design, the clip level should be very close to giving FS
recording, so it may be reasonable to conclude that this is indeed FS.
But we don't know that. I suspect that the HD-P2 is designed pretty
well. So what's the actual level of the recorded noise with no input
signal? I doubt that it's actually -95 dBFS. In fact, didn't someone's
Audition show -65 dBFS or so?
It would be so simple to measure this. All you need is the recorder and
some basic test equipment. But I don't have a recorder, and I guess
Stony doesn't have the appropriate test equipment. We'll have to wait
for the planets to collide. Until then, we're just guessing, so I'm
going to stop.
Mike Rivers
Chel van Gennip wrote:
> Do you really think the preamp to clip before the AD converter does in a
> device where preamp and AD converter are combined in a single unit?
I guess you've never heard of a Walkman DAT? Or a Jukebox 3? It may
sound stupid, but it happens more than we'd like it to.
> Let's just assume the specified input level for a recorder specifies the
> input level to make recordings.
Assume what you like. I'll wait for real measurements.
Mike Rivers
Chel van Gennip wrote:
> On Sat, 10 Dec 2005 23:44:46 +0100, Mike Rivers wrote:
> If I remember well, you were content about these measurements, and I was
> questioning them, as they could not be right.
Correction - I was content with the noise level on Stony's samples, by
ear, not by measurement. I would still like to see some numbers that I
can believe. I'll be happy to make measurements if someone sends me a
recorder to measure. I guess that won't be you.
Mike Rivers
Chel van Gennip wrote:
> Correction, you were content about the measured 95dB S/N of the recorder
> with microphone input at full gain:
Will you please go to hell with your misinterpretations of what I said?
I said that a noise level of -95 dB was nothing to complain about. ("CD
quality" = 96 dB) I never said I thought that figure was actually
correct.
There is absolutely no reason to believe that this $1,000 flash card
recorder with mic inputs is 30 dB quieter than well respected mic
preamps. Don't kid yourself, and particualry, don't try to convince
other readers of this newsgroup of that unless you can back it up. I
don't know why youj're stalking me on this. Please go away until you
have some real facts based on real measurements.
gl0...@yahoo.com
he got shafted with the MT ;o) It's all about cognitive dissonance and
the inevitable hot air that follows.
Markus Mietling
>Markus Mietling wrote:
>
>> These are the results for HDP2Gain10.wav:
>
>So what is your conclusion?
My very own conclusion is that the HD-P2 might be noisier than
I originally expected from Mr Ballard's report: He wrote that
| The average noise level with the gain all the way up was about 95 db
| RMS below 0, digitally. [1]
... while Audition's analysis of the corresponding .WAV file yields
>| Average RMS Power: -80.29 dB -80.76 dB
for the left and the right channel respectively. To me, this looks like
a difference of more than 14 dB.
>Or is that left as an exercise for the
>student?
I don't deem my insignificant opinions to be interesting enough to be
"left as an exercise for the student." I'm not sure why you would impute
such arrogance to me.
>I see peak amplitudes below -65 dB (FS assumed?) but the RMS
>measurements make no sense to me. Why is this power? How can they
>relate power, which implies both voltage and current, to a number of
>bits below 24 (OK, 23 and a sign bit)?
Some people [1] say that "The average (mean) power can be computed using
the 'mean squared voltage'. Specifically, power is found using the
square of the 'root mean squared voltage'".
If a .WAV file can be interpreted as a numerical record of voltage
fluctuations over time, then I don't see why it would be counerintuitive
to deduce a power measurement from these numbers.
MM
Markus Mietling
>The actual difference in broadband RMS noise
>between a 150 ohm resistor and a short is probably less than 1 dB,
In at least one case [1], the actual difference between a 150 ohm
resistor and a 50 ohm resistor is already 3 dB.
It seems highly unlikely (to me) that the difference between a 150 ohm
resistor and a *short* is less than that.
MM
Mike Rivers
Chel van Gennip wrote:
> The ratings of this 722 if -128dBu at 150 Ohm / -133dBu at 50 Ohm are
> quite close to the theoretical maximum. These values you see for the
> better preamps. So if you see measurement that result in an EIN of -155dBu
> you should not trust the measurements.
I certainly wouldn't. Did you see an EIN of -155 dBu for anything we've
been talking about?
> The output noise of good microphones
> normally is higher, e.g. -120dBu for my AKG's C480B's or -110dBu for the
> DPA 4003.
And the Johnson noise of a 150 ohm resistor at 20 degrees C over a 10
kHz bandwidth is about -.15 microvolts, about -155 dB below a
reasonable maximum output level of +22 dBu or so. Could this be the
-155 of which you speak? That's not EIN.
Where's Myth Buster Phil Allison when we need him?
Leo Sreto
I joined this group to find out if the TASCAM HD-P2 is a good replacement
for my DAT and if the recording-quality of the device is similar to it. I'm
really impressed with all your numbers and calculations; way above my
insight in these matters. You see, I'm only interested in *how is records*
in real life, with a microphone in front of a musician.
Suggestion: Why don't Mr. Ballard [who is the only one here, (I think), with
a HD-P2] post a .wav-file with some sound, so we/I can hear the quality.
Just the wind in the woods, coffee-machine brewing coffee, his dog barking,
anything :-) -- no, I'm not kidding, for me it says more about the HD'P2
than all the numbers in the world.
Just my humble opinion :-)
Regards, and peace
Leo
"Mike Rivers" <mri...@d-and-d.com> wrote in message
news:1134418612.4...@g14g2000cwa.googlegroups.com...
Mike Rivers
Leo Sreto wrote:
> I'm only interested in *how is records*
> in real life, with a microphone in front of a musician.
> Suggestion: Why don't Mr. Ballard [who is the only one here, (I think), with
> a HD-P2] post a .wav-file with some sound, so we/I can hear the quality.
> Just the wind in the woods, coffee-machine brewing coffee, his dog barking,
> anything :-) -- no, I'm not kidding, for me it says more about the HD'P2
> than all the numbers in the world.
That's a healthy attitude. In fact, I said that listening to the
"quiet" with the mic input shorted meant that it was quiet enouth for
me.
I think he mentioned that he was recording a concert this coming
weekend. Hopefully he'll figure out how to run it by then so he brings
home a recording <g>, and that his subject will allow him to share a
bit of the recording.
Leo Sreto
> weekend. Hopefully he'll figure out how to run it by then so he brings
> home a recording <g>, and that his subject will allow him to share a
> bit of the recording.
>
Mike,
That's right! I forgot!
Looking forward to it.
Leo
Chris Hornbeck
wrote:
>And the Johnson noise of a 150 ohm resistor at 20 degrees C over a 10
>kHz bandwidth is about -.15 microvolts, about -155 dB below a
>reasonable maximum output level of +22 dBu or so. Could this be the
>-155 of which you speak? That's not EIN.
Yeahbut. -155 dBu is less than the thermal noise of the
shunt-plus-parasitic losses. This is bogus.
my best and oldest friend had to lose his old dog
today. God bless us every one.
Chris Hornbeck
Bob Cain
Chel van Gennip wrote:
> On Mon, 12 Dec 2005 05:53:48 +0100, Chris Hornbeck wrote:
>
>> Second, the assumption that a -60 dBu reference level and a (mistaken)
>> 95 dB S/N ratio could be *added* to give an absolute input noise voltage
>> with no knowledge of absolute gain before conversion.
>
> Well the -60 dBu input level for FS under the described testing conditions
> is from the device specification. This gives an absolute reference for the
> input level. It could be wrong, but it looks like a normal input level for
> a microphone preamp at maximum gain.
Sounds like you have the instrumentation to actually measure that. It
can be done on the cheap with just a VOM, a near FS sinusoidal file
played from a DAW, measurement of the voltage at the DAC output that
results from playing it, wrapping that output to the input after digital
scaling of the file to somewhat below the full gain clip point,
recording that and measuring the recorded signal dB re digital FS. It's
clear that you understand the math to get from all that to the FS dBu at
max gain.
> A S/N ratio is a ratio between two
> signals, the Signal level can't be higher than FS, so I think it is right
> to calculate the EIN by applying the measured S/N ratio to the FS input
> level under testing conditions.
Me too.
>
> So I think there must be something wrong with the measured S/N ratio for
> the preamp at full gain in this test. Either it is not at full gain, or
> the S/N ratio isn't 95 dB or both.
I'm suspicious of the noise measurement. It implies a wiggling of just
the low order bit of 16 at full gain, a conspicuous value. I'd expect
to see noise in several of the low order bits at that gain. With 150
Ohm resistive shorting, the Tascam US-122 at full gain has noise mostly
but not exclusively confined to the low order three of 16. That works
out to about -128 dBu EIN after measuring the full gain, FS reference as
sketched above.
Bob
--
"Things should be described as simply as possible, but no simpler."
A. Einstein
Bob Cain
Mike Rivers wrote:
> And the Johnson noise of a 150 ohm resistor at 20 degrees C over a 10
> kHz bandwidth is about -.15 microvolts, about -155 dB below a
> reasonable maximum output level of +22 dBu or so. Could this be the
> -155 of which you speak? That's not EIN.
If it were all of the EIN, it would be -134 dBu.
What noise voltage would be expected from a 3.4k resistor?
Mike Rivers
Bob Cain wrote:
> What noise voltage would be expected from a 3.4k resistor?
Vsquared = 4*K*T*R*Bw
where:
K = Boltzmann's constant (1.38 x 10^-23)
T = Temperature in Kelvin degrees (293 is a good number)
R = Resistance in ohms
Bw = Bandwidth over which you want to know the noise. (10 kHz is a good
number)
Get out your Curta and crank away.
Bob Cain
Chel van Gennip wrote:
> On Tue, 13 Dec 2005 09:14:48 +0100, Bob Cain wrote:
>
>> With 150 Ohm resistive shorting, the Tascam US-122 at full gain has
>> noise mostly but not exclusively confined to the low order three of 16.
>> That works out to about -128 dBu EIN after measuring the full gain, FS
>> reference as sketched above.
>
> That -128 dBu is about the theoretical limit of what you can get. Many
> modern preamps come close to that limit. The three bits noise indicate a
> S/N of 75-80 dB, and that is all you can get from a preamp at maximum
> gain. Adding a microphone, even the best ones, will let the S/N ratio
> drop with several dB.
Right. It's a rare mic, indeed, whose self noise is less than -128 dBu.
Even the NT-1A, a paragon of quietude with 5 dBA self noise and -32 dB
sensitivity re 1V/Pa, yields -119 dBu equivalent self noise out, 9 dB
above the US-122 pre input (the -128 dBu for the US-122 input was also
A-weighted.)
This is an example of why a simple "80 dB dynamic range" specification
for a pre tells you nothing particularly useful. A dBu EIN spec instead
allows a perspective that can be better related to the world.
Arny Krueger
"Bob Cain" <arc...@arcanemethods.com> wrote in message
news:dnm0c...@enews4.newsguy.com...
>
>
> Mike Rivers wrote:
>
>> And the Johnson noise of a 150 ohm resistor at 20 degrees C over a 10
>> kHz bandwidth is about -.15 microvolts, about -155 dB below a
>> reasonable maximum output level of +22 dBu or so. Could this be the
>> -155 of which you speak? That's not EIN.
>
> If it were all of the EIN, it would be -134 dBu.
10 KHz is a bit of an odd bandwidth for audio production, no?
> What noise voltage would be expected from a 3.4k resistor?
.74 microvolts or -120.4 dBu
Bob Cain
Arny Krueger wrote:
> "Bob Cain" <arc...@arcanemethods.com> wrote in message
> news:dnm0c...@enews4.newsguy.com...
>>
>> Mike Rivers wrote:
>>
>>> And the Johnson noise of a 150 ohm resistor at 20 degrees C over a 10
>>> kHz bandwidth is about -.15 microvolts, about -155 dB below a
>>> reasonable maximum output level of +22 dBu or so. Could this be the
>>> -155 of which you speak? That's not EIN.
>> If it were all of the EIN, it would be -134 dBu.
>
> 10 KHz is a bit of an odd bandwidth for audio production, no?
True.
>
>> What noise voltage would be expected from a 3.4k resistor?
>
> .74 microvolts or -120.4 dBu
So that's the best you can get using phantom even if the supply is noise
free, right?
Mike Rivers
Bob Cain wrote:
> > 10 KHz is a bit of an odd bandwidth for audio production, no?
> True.
It's actually not so unreasonable, but if you insist, you can multiply
whatever number I quoted by the square root of 2 and call it noise to
20 kHz.
> > .74 microvolts or -120.4 dBu
>
> So that's the best you can get using phantom even if the supply is noise
> free, right?
Well, a microphone isn't a resistor. If it has a transformer ouptut,
the transformer winding probably has less DC resistance than a 150 ohm
resistor so in theory would contribute less noise, but then there's the
noise of the microphone's electronics to contend with, so you can
really only theorize.
Since there are only a handful of contemporary mic preamps with more
than about 60 dB of gain, if you're looking for a "quiet" microphone to
use with low level sources, what you really want is one with high
sensitivity rather than one with a low self-noise. What good is it if
you can't get it up? <g>
Les Cargill
Time for a new dog, then.
--
Les Cargill
Chris Hornbeck
wrote:
>10 KHz is a bit of an odd bandwidth for audio production, no?
Yes, but it comes with a lot a pedigree. Before the era of
brickwall filtering, conflicting fudge factors settled
somewhere in the vicinity as a close-enough-for-the-
grenades-we're-throwing number.
Still perfectly appropriate as long as everybody's
talking the same-same.
(But maybe a stretch for some of us, or at least, more
sooner than later). And, hopefully, not.
Good fortune,
Chris Hornbeck
Bob Cain
Mike Rivers wrote:
> Bob Cain wrote:
>
>>> 10 KHz is a bit of an odd bandwidth for audio production, no?
>> True.
>
> It's actually not so unreasonable, but if you insist, you can multiply
> whatever number I quoted by the square root of 2 and call it noise to
> 20 kHz.
>
>>> .74 microvolts or -120.4 dBu
>> So that's the best you can get using phantom even if the supply is noise
>> free, right?
>
> Well, a microphone isn't a resistor. If it has a transformer ouptut,
> the transformer winding probably has less DC resistance than a 150 ohm
> resistor so in theory would contribute less noise, ...
Aargh! Of course. The phantom resistors can only serve to reduce the
total resistance seen at the input because they are in parallel with the
source resistance of the mic. Never mind. :-)
Pawel Kusmierek
of Marantz PMD-671 or other Marantz PMD recorders? I admi I am somewhat
lost...
Mike Rivers
I'll be happy to. Just send me one of each. I have test equipment and
curiosity. However off the top of my head, no.
Pawel Kusmierek
But I see this is not easy for others as well...
Maybe someone can at least explain to me, what does it mean:
Noise level Trim Max (22 Hz to 22 kHz): Up to -55 dBu MIC to LINE OUT
Did they measure the noise of the whole signal path from Mic to Line
out? Why not from Mic to digital, which is probably the most important?
-55dBu is measured where?? If it is calculated noise level at the Mic
input, then it makes no sense, because at Trim Max the Mic input level
is -60dBu. If it is at LineOut, it is still only 47.2 dB below the
LineOut "Input level" of -10dbV (=-7.8dBu).
Mike Rivers
Pawel Kusmierek wrote:
> Maybe someone can at least explain to me, what does it mean:
> Noise level Trim Max (22 Hz to 22 kHz): Up to -55 dBu MIC to LINE OUT
> Did they measure the noise of the whole signal path from Mic to Line
> out?
Yes, and that's a pretty honest measurement. Also not a terribly
impressive one. This means that with a presumably properly terminated
input (or a short circuit - my preference) and the trim set to to
maximum gain, there will be hiss coming out the line output at a level
of -55 dBu when it's in the input monitor mode.
Chances are pretty good that if you were to make a test recording of
noise with zero input with the trim all the way up, it would play back
(through the line output) at a level of -55 dBu since the "recorder"
part is almost certainly unity gain, or at least there's a unity gain
setting if there's a playback level control at all. Usually there
isn't, for the line output of a portable recorder.
> Why not from Mic to digital, which is probably the most important?
Because dBu is a voltage and digital is bits. You can't compare these
apples and oranges. You don't know the relationship between the input
level (in this case the front end noise) and what level that digitizes
to. If you knew, for instance, that with the trim at maximum, a signal
of, say -45 dBu produced full scale digital level, then you would have
some significant information, but since we don't know this, we can only
guess.
> -55dBu is measured where?? If it is calculated noise level at the Mic
> input, then it makes no sense, because at Trim Max the Mic input level
> is -60dBu.
No, the gain (if we actually know this - I assume it's on the spec
sheet) of the mic amplifier is 60 dB. The mic input level is whatever
the mic is putting out. At 60 dB gain, whaever comes out the line
output (or wherever they measure that gain through) will be 60 dB
higher than whatever goes in - doesn't matter whether it's noise or
signal. But if you're recording crickets at 100 yards with an SM57,
even with 60 dB of gain, you'll still be recording a very low signal
because crickets at 100 yards might be annoying when you're trying ot
record a vocal in your studio, but in reality, aren't very loud.
This is why the measurement you're looking for is difficult to express
in familiar terms. Because you can't.
Pawel Kusmierek
> Pawel Kusmierek wrote:
> > Did they measure the noise of the whole signal path from Mic to Line
> > out?
>
> Yes, and that's a pretty honest measurement.
Yes, but the main purpose of a portable digital recorder is to record
from an input into digital. Who would buy it and use it as preamp only?
The measurement combines the input (Mic) noise, which is relevant, and
output (Line out) noise, which is irrelevant. I guess that it is the
input side which is mostly responsible for the noise, so my point is
probably not so important. But it's only my guess, and it might be not
true.
>
> > Why not from Mic to digital, which is probably the most important?
>
> Because dBu is a voltage and digital is bits. You can't compare these
> apples and oranges. You don't know the relationship between the input
> level (in this case the front end noise) and what level that digitizes
> to. If you knew, for instance, that with the trim at maximum, a signal
> of, say -45 dBu produced full scale digital level, then you would have
> some significant information, but since we don't know this, we can only
> guess.
I think that the relationship between the input level and the digital
level should be known for such device. I'd love to see specs like:
At minimum trim:
the analog (preamp) clipping point is at A uV (or dBu, or something
like that) input
the digital (ADC) clipping point (0 dBfs) is at B uV input
noise is at C dBfs
At maximum trim:
the analog clipping point is at D uV input
the digital clipping point 0 dBfs is at E uV input
noise is at F dBfs
This would be more usable, and I would know if the unit has been
properly designed, that is, if A is approximately equal to B, and D is
approximately equal to E.
Am I demanding too much?
Mike Rivers
Pawel Kusmierek wrote:
> Yes, but the main purpose of a portable digital recorder is to record
> from an input into digital. Who would buy it and use it as preamp only?
You wouldn't, but you can't listen to digital without following it with
something analog. And you can't measure it either, because you don't
know how much analog noise (that you're measuring) is represented by a
certain digital noise level. There is no such thing as a calibrated D/A
converter.
> The measurement combines the input (Mic) noise, which is relevant, and
> output (Line out) noise, which is irrelevant.
No, it doesn't combine them, other than to add whatever noise the D/A
converter introduces, which is surely much less than the preamp. You're
using the D/A converter as part of your test equipment so you can take
an output reading that's comparable to the input.
How would you propose to test it? And more important, how would you
express the results of your test?
Let me ask you a simpler question, and maybe the answer will help you
to understand my point. Let's say you put a -60 dBu signal into the mic
input, turned the input gain all the way up, and made a recording. You
looked at the digital file and saw that the peaks were at -4 dBFS.
What's the gain? If your answer is 56 dB, you're wroing. You don't know
what the gain is until you play that digital recording back through
something that converts it to analog, so you can measure the voltage
coming out and compute the ratio of output to input voltage.
> I think that the relationship between the input level and the digital
> level should be known for such device. I'd love to see specs like:
That would be nice, but we don't know it. And even if we did know it
for one device, it wouldn't necessarily be the same for another device.
All CD players don't play back the same CD at the same analog level,
for example.
> Am I demanding too much?
No, you're describing a measurement that depends on a certain
calibration factor that isn't universal. Life would be simpler in the
specification lane if all A/D converters had the same input sensitivity
and all D/A converters had the same output level for a given digital
level, but that isn't the way it is. "Pro" converters don't usually
vary by more than about 5 dB one way or another, but if you don't have
the same relationship between analog voltage and digital level both
going in and coming out, any measurement you make will have to be
corrected for that difference. By going on the assumption that the
recorder is a "unity gain" device - that is you get the same level out
as you put in, you can at least measure gain and noise for that
particular device and know your measurements are meaningful.
Pawel Kusmierek
> Pawel Kusmierek wrote:
>
>
> You wouldn't, but you can't listen to digital without following it with
> something analog.
Yes, but typically the recording would be transferred digitally into a
computer or something, processed and then played back, probably using a
very different device. I am interested in how much noise will the
device introduce into my recording.
> Let me ask you a simpler question, and maybe the answer will help you
> to understand my point. Let's say you put a -60 dBu signal into the mic
> input, turned the input gain all the way up, and made a recording. You
> looked at the digital file and saw that the peaks were at -4 dBFS.
> What's the gain? If your answer is 56 dB, you're wroing.
No, I clearly understand the difference between the two. I probably
cannot express my understanding properly in English.
> > I think that the relationship between the input level and the digital
> > level should be known for such device. I'd love to see specs like:
>
> That would be nice, but we don't know it.
Yes, and this is what I am complaining about. My opinion is that we
should be given the data, including the input level to digital level
relationship, by the manufacturer.
> And even if we did know it
> for one device, it wouldn't necessarily be the same for another device.
Do you mean "for another recorder of any kind" or "for another HD-P2"?
> Life would be simpler in the
> specification lane if all A/D converters had the same input sensitivity
> and all D/A converters had the same output level for a given digital
> level, but that isn't the way it is. "Pro" converters don't usually
> vary by more than about 5 dB one way or another,
Again: do you mean "all existing converters" or "converters of the same
type and model and from the same manufacturer"?
Leo Sreto
I got my TASCAM HD-P2 today! :-)
It's a faboulus device! So easy to handle and IMHO the perfect replacement
for my DAT. The sound is much better than what I get from my DAT. So all in
all I'm happy with it. Now we will see how it works in the long run...
Happy New Year to all from
Leo
"Pawel Kusmierek" <bwv...@gazeta.pl> wrote in message
news:1135722543.6...@f14g2000cwb.googlegroups.com...