Topic: Tascam DR-70D 4-channel audio recorder (Part 3)

[noahbickart]

I own a pair of mk6 capsules, a pair of mk22 capsules, a pair of kcy cables (one vark one nbob), and a busman modded tascam. I'll gladly run two identical m/s rigs if someone is willing to lend me another dr-70d...

[Ozpeter]

Compared cheesecadet modded deck to mine stock dr70.

That's a very short test.  Obviously I don't know how actually quiet your room is, but the result could arise from a car passing in the next street at that moment.  And if the mod has, perchance, made some difference to the gain structure, then maybe the playing field isn't level.

What we need to know about is not noise, but signal-to-noise.  So some kind of signal has to be involved in the testing.  One "amateur" method which I rather like is the "clock tick" test.  In your quiet room, place the mic close to a ticking clock, so that the sound of the clock ticks makes the meters move at least a bit.  Record machine A.  Don't move the mic or clock.  Record machine B with its meters showing ballpark similar levels.  Now when you compare the two in your DAW you should be able to fiddle with the gain in the DAW so that you can see that the waveforms from the two recordings are peaking (on the ticks) to the same level.  Now listen to the silence in between the ticks.  You should be able to hear if there is an obvious difference, and you should be able to see it with spectral analysis. 

Of course you can do a similar kind of thing with something which acts as a tone generator - there are mobile phone apps which can be used - and you align the recordings to the same level using the tone, then turn off the tone and record.  I just like the clock method because it gives you an ongoing signal to give your ears a reference point.

[Ozpeter]

As for how much of a real-world problem the system noise is - if you have a quiet room with a hifi or TV in it, and you are familiar with the kind of level settings you use at a concert, record a bit of the hifi with those level settings, setting the hifi playback volume so that you get a typical level on the meters.  (This might annoy your neighbours).   Then stop the hifi playing and record a bit of room silence.  Now play back the recording you've just made, with the playback level set like you'd normally have it.  When you get to the 'silent' part, do you actually hear any system noise (in comparison to the music level you've just been playing)?  Probably not.

Of course if you don't have a quiet playback room, then you can't do this test, but then again you probably shouldn't be worrying about system noise if you don't have a quiet playback room anyway...

[phil_er_up]

if you have a quiet room with a hifi or TV in it, and you are familiar with the kind of level settings you use at a concert, record a bit of the hifi with those level settings, setting the hifi playback volume so that you get a typical level on the meters.  (This might annoy your neighbours).   Then stop the hifi playing and record a bit of room silence.  Now play back the recording you've just made, with the playback level set like you'd normally have it.  When you get to the 'silent' part, do you actually hear any system noise (in comparison to the music level you've just been playing)?  Probably not.

ozpeter this is what I did.

Have a playback system/room with walls treated acoustically with sound panels and bass traps. Outside walls have added insulation. Playing loud music is not a problem.  Volume was around a 100 DB (have sound pressure meter) in room with ear plugs. Preamp has numbers so played it at the same volume number with same CD/song. The 10 seconds of noise I refereed to in the above post was the room noise, I turned off fans, AC unit for the test. The volume level was the same DB for the 10 seconds meaning both 10 seconds were a straight line in wavlab. No dip/peaks in them. The recording were not from the same moment though within 10 minutes of each other and nothing changed in the room except me changing the decks with same mics and position and volume level and song from the CD that I was using. Used wavlab to cut the silence out and used sound forge for the analysis since I think it produces better output numbers that are easy to read.

Though I do not remember if I actually heard any difference and not sure if my ear would be capable detecting that level of noise that is why I ran SF on those 10 seconds.   If I have time still have those files I believe and could see if I hear a difference.

[phil_er_up]

Compared cheesecadet modded deck to mine stock dr70.

Of course you can do a similar kind of thing with something which acts as a tone generator - there are mobile phone apps which can be used - and you align the recordings to the same level using the tone, then turn off the tone and record.  I just like the clock method because it gives you an ongoing signal to give your ears a reference point.

Can I use the tone generator from the dr-70 or SD mixpre-d?

If yes, how would I hook it up?

[johnmuge]

had my first problem with a stock DR-70D.

Three glitches I've counted so far in the Cracker set... None longer than a moment, but all occur in all 4 channels, and don't occur on an AUD recording made with totally separate gear (DPA4061's > D50)

https://archive.org/details/Cracker2015-07-07

Guess I'll be checking to see if it was the Tascam or the SD card (ADATA 16GB SDHC, made in China, Class 6, ser# 1202470917, BL1120122026G, blue gradient sticker)
anyone else noticed anything like this? I was running CH 1-2 unbalanced from the SBD, CH 3-4 phantom directly to the Neumanns, and powering off a Ravpower USB pack, with AA's loaded. I used LOW gain mode for all 4 channels during Cracker.

There was also MAJOR trouble during the opening act, when I ran the mics at MID mode, and had to keep the level knobs VERY close to zero, making me nervous that I'd shut them off completely. Got a bunch of red flashing lights, and some minor "overlevel samples" but the glitches are far more troubling! EVERYTHING past 24:51 in both files is totally corrupted, with signals sounding like they're cutting in and out, or interleaved!? Very bad.

I had the exact problem that you had.  I had the recorder on Mid for the first song and it was too hot so I turned it down too low and it shuts off or just sounds bad.  I switched to Low during the 2nd song and everything sounds great until 30 minutes or so and then it get totally messed up cutting in and out.  The first 30 minutes of the second set is great and then it turns to shit again.  I noticed that I had the power saver mode factory set to 30 minutes so I changed that and I can't get it to screw up again testing it at home.  I'm hoping that was my problem.  I was using  AKG's and Milabs and had phantom power on to all 4 channels. 

[Ozpeter]

Can I use the tone generator from the dr-70 or SD mixpre-d?

If yes, how would I hook it up?

I was just thinking in terms of anything which produces an audible tone in the room at a fixed reproducible level, so you can be sure that the level you are recording at is identical each time you run the test (judged by the meters rather than by input level settings, as the mod might make a difference to the recorded level for any given knob position).

In any test of anything where the outcome is surprising, it's essential to ensure that there's nothing in the test procedure that could possibly have influenced the result (and that can be hard to achieve if you don't have a proper test facility of course).  Having said that, if I'm reading your test procedure right, it should give a reasonably accurate result.  If you were able to post the files it would enable the team to pore over them and see whether there might be any explanation apart from that the mod makes a surprising difference to the noise level.

[Sonus Captor]

Here's an interesting article about measuring mic preamp noise:

http://www.ap.com/kb/show/331#output

[phil_er_up]

Alluded in an above post that I thought the DR70 mod lowered the noise floor -14 db. After more review realize I did not interpret the numbers from SF correctly. Used to wavlab stats and just acquired SF in last few months and not sure the fields in SF were what I thought they were. So think it is best to start over with a clean slate and let you guys make an interpretation. Sorry about the false data. Feel bad about that. Do not want to influence the community one way or another on something that is not FACT.

========================================================================
Ran 3 more test with modded/non-modded deck this week. Non modded deck is 1.0 software of whatever number was installed in the stock unit. Modded deck has 1.11 version.

Side note: When gain pods are max on both units - the signal levels on the recorder are not the same. With low setting, gain pods at maximum gain there was about a ~4db difference between them and I let that go and did not attempt to adjust gains the same because I wanted the pods at max. Not sure if version 1.0 and version 1.11 have same gain values.


==============================================================

Ran tascam on low setting, mic + phantom. Gain pods maximum gain.

3 tests I did:

a) clock
b) song
c) pillow - jon stoppable suggested putting mics in pillow and recording that.

Ozpeter suggested using a clock. Tried that and signal would not register on the recorders levels and recorded it anyway. Did not know how to run the tone generator test so did not attempt that.

Results:
==============================================================
tASCAM dr70 MODDED PILLOW
                                                    Left Channel           Right Channel       
Cursor position (Time)                              00:00:00.000           00:00:00.000       
Sample value at cursor (dB)                         -Inf.                  -Inf.               
Minimum sample position (Time)                      00:00:11.923           00:00:11.923       
Minimum sample value (dB)                           -52.262                -52.288             
Maximum sample position (Time)                      00:00:11.268           00:00:12.310       
Maximum sample value (dB)                           -53.082                -51.279             
RMS level (dB)                                      -65.949                -65.125             
Average value (dB)                                  -90.309                -90.377             
Zero crossings (Hz)                                 3,107.92               2,775.90           
Maximum true peak sample position (Time)            00:00:11.923           00:00:12.310       
Maximum true peak sample value (dB)                 -52.262                -51.279             
Maximum filtered true peak sample position (Time)   00:00:11.923           00:00:12.310       
Maximum filtered true peak sample value (dB)        -52.262                -51.279             

Integrated Loudness (LUFS)      -68.50
Loudness Range (LU)          1.90
Maximum True Peak Loudness (dBTP)   -51.28
Maximum Short-Term Loudness (LUFS)   -67.57
Maximum Momentary Loudness (LUFS)   -66.30

===========================================================
tASCAM dr70 NON-MODDED  PILLOW

                                                    Left Channel           Right Channel       
Cursor position (Time)                              00:00:00.000           00:00:00.000       
Sample value at cursor (dB)                         -Inf.                  -Inf.               
Minimum sample position (Time)                      00:00:03.486           00:00:03.930       
Minimum sample value (dB)                           -58.086                -54.186             
Maximum sample position (Time)                      00:00:00.661           00:00:04.227       
Maximum sample value (dB)                           -58.855                -54.641             
RMS level (dB)                                      -72.812                -69.027             
Average value (dB)                                  -90.275                -90.275             
Zero crossings (Hz)                                 7,563.09               4,808.02           
Maximum true peak sample position (Time)            00:00:03.486           00:00:03.930       
Maximum true peak sample value (dB)                 -58.086                -54.123             
Maximum filtered true peak sample position (Time)   00:00:03.486           00:00:03.930       
Maximum filtered true peak sample value (dB)        -58.086                -54.123             

Integrated Loudness (LUFS)      -66.97
Loudness Range (LU)          2.10
Maximum True Peak Loudness (dBTP)   -54.12
Maximum Short-Term Loudness (LUFS)   -66.66
Maximum Momentary Loudness (LUFS)   -64.63

===========================================================
tASCAM dr70 MODDED CLOCK
                                                    Left Channel           Right Channel       
                                                    Left Channel           Right Channel       
Cursor position (Time)                              00:00:00.000           00:00:00.000       
Sample value at cursor (dB)                         -Inf.                  -Inf.               
Minimum sample position (Time)                      00:00:07.524           00:00:06.637       
Minimum sample value (dB)                           -45.472                -44.633             
Maximum sample position (Time)                      00:00:06.195           00:00:06.194       
Maximum sample value (dB)                           -45.082                -44.568             
RMS level (dB)                                      -67.968                -66.125             
Average value (dB)                                  -90.377                -90.241             
Zero crossings (Hz)                                 5,155.99               3,883.54           
Maximum true peak sample position (Time)            00:00:06.195           00:00:06.194       
Maximum true peak sample value (dB)                 -45.082                -44.568             
Maximum filtered true peak sample position (Time)   00:00:06.195           00:00:06.194       
Maximum filtered true peak sample value (dB)        -45.082                -44.531             

Integrated Loudness (LUFS)      -64.55
Loudness Range (LU)          3.80
Maximum True Peak Loudness (dBTP)   -44.57
Maximum Short-Term Loudness (LUFS)   -62.13
Maximum Momentary Loudness (LUFS)   -60.83







===========================================================
TASCAM dr70 NON-MODDED CLOCK

Cursor position (Time)                              00:00:00.000           00:00:00.000       
Sample value at cursor (dB)                         -Inf.                  -Inf.               
Minimum sample position (Time)                      00:00:17.638           00:00:06.320       
Minimum sample value (dB)                           -50.874                -53.523             
Maximum sample position (Time)                      00:00:04.321           00:00:06.542       
Maximum sample value (dB)                           -52.110                -52.320             
RMS level (dB)                                      -75.116                -74.267             
Average value (dB)                                  -90.309                -90.241             
Zero crossings (Hz)                                 10,421.85              8,275.01           
Maximum true peak sample position (Time)            00:00:17.638           00:00:06.542       
Maximum true peak sample value (dB)                 -50.874                -52.320             
Maximum filtered true peak sample position (Time)   00:00:17.638           00:00:06.542       
Maximum filtered true peak sample value (dB)        -50.874                -52.217             

Integrated Loudness (LUFS)      -69.49
Loudness Range (LU)          0.00
Maximum True Peak Loudness (dBTP)   -50.87
Maximum Short-Term Loudness (LUFS)   -70.36
Maximum Momentary Loudness (LUFS)   -68.92

===========================================================

TASCAM dr70 MODDED SONG
                                                    Left Channel           Right Channel       
Cursor position (Time)                              00:00:00.000           00:00:00.000       
Sample value at cursor (dB)                         -Inf.                  -Inf.               
Minimum sample position (Time)                      00:03:07.083           00:02:25.609       
Minimum sample value (dB)                           -5.693                 -3.863             
Maximum sample position (Time)                      00:03:07.089           00:02:23.598       
Maximum sample value (dB)                           -5.542                 -4.673             
RMS level (dB)                                      -23.495                -21.933             
Average value (dB)                                  -90.309                -90.309             
Zero crossings (Hz)                                 2,070.91               1,927.47           
Maximum true peak sample position (Time)            00:03:07.089           00:02:25.609       
Maximum true peak sample value (dB)                 -5.542                 -3.863             
Maximum filtered true peak sample position (Time)   00:03:07.089           00:02:25.609       
Maximum filtered true peak sample value (dB)        -5.525                 -3.375             

Integrated Loudness (LUFS)      -19.93
Loudness Range (LU)          3.50
Maximum True Peak Loudness (dBTP)   -3.86
Maximum Short-Term Loudness (LUFS)   -17.71
Maximum Momentary Loudness (LUFS)   -14.75

===========================================================

TASCAM dr70 NON-MODDED SONG

                                                    Left Channel           Right Channel       
Cursor position (Time)                              00:00:00.000           00:00:00.000       
Sample value at cursor (dB)                         -Inf.                  -Inf.               
Minimum sample position (Time)                      00:03:20.350           00:03:06.552       
Minimum sample value (dB)                           -12.606                -9.713             
Maximum sample position (Time)                      00:01:55.616           00:03:06.545       
Maximum sample value (dB)                           -12.286                -10.492             
RMS level (dB)                                      -29.664                -27.877             
Average value (dB)                                  -90.343                -90.309             
Zero crossings (Hz)                                 2,551.06               2,298.48           
Maximum true peak sample position (Time)            00:01:55.616           00:03:06.552       
Maximum true peak sample value (dB)                 -12.286                -9.713             
Maximum filtered true peak sample position (Time)   00:01:55.616           00:03:06.552       
Maximum filtered true peak sample value (dB)        -12.286                -9.713             

Integrated Loudness (LUFS)      -26.11
Loudness Range (LU)          3.10
Maximum True Peak Loudness (dBTP)   -9.71
Maximum Short-Term Loudness (LUFS)   -23.58
Maximum Momentary Loudness (LUFS)   -20.87

======================================================================

Do not have time to post the sound files this week. If there is interest and I have time next week then I would go forward with posting them.

Hope made the above clear.
Again sorry about the earlier error and that is not my MO usually.

[willndmb]

So for us lay people, what do those numbers show overall?
Is the mod really doing anything worth getting it done?

[noahbickart]

Hi folks, since this concerns the DR-70d, I'd really love help with this: http://taperssection.com/index.php?topic=173946.0

[voltronic]

Phil,

The maximum LOW setting between those two firmware versions is 10dB higher for the newer version.  The newer firmware also added a MID level.  See the FAQ thread.

Unfortunately, this means your test isn't valid.  There are two solutions:
1. Do the tests again, but with both units set to HIGH or HIGH+.  Those maximum values did not change with the update.  Testing HIGH+ will probably only work with the pillow, and I wouldn't even bother as Jim told me the input transformer noise dominates that setting both before and after the mod.
2. Update to the latest firmware for both units and test as before.  This is the best choice, as the lower limits of the gain pots and thus any settings along the way have changed with the update, meaning that right now you can not level match the two recorders for any setting other than maximums of HIGH and HIGH+.  If you're going to do that, I would test the MID and HIGH settings as well, as those may better show the benefits (or not) of the mod.

EDIT:  One more thing.  I remember Cheesecadet saying on more than one occasion that he did not update the firmware of his modded unit.  What you posted above shows that it is updated.  Is it possible you labeled the units in reverse, or did you update the firmware on that one yourself?

If you're willing to, please post the recordings so we can also analyze with other software.

Sorry for the bad news.  Again, we all appreciate you going through this trouble.

[voltronic]

Hi folks, since this concerns the DR-70d, I'd really love help with this: http://taperssection.com/index.php?topic=173946.0

You linked to the FAQ.  What's your question?

[noahbickart]

Hi folks, since this concerns the DR-70d, I'd really love help with this: http://taperssection.com/index.php?topic=173946.0

You linked to the FAQ.  What's your question?

funny, that link takes *me* to the thread I started in the "ask the tapers" sub forum about some static on a recording that I hope is as a result of a faulty sd card and not the recorder.

[Jonmac]

The way to assess noise level is to feed in a constant 1000Hz tone, set your levels to 0 db, and record 10 seconds of tone, pause the recorder, remove the tone and replace it with a 200 ohm metal film resistor and record 10 seconds of silence.

Then, open the file in a DAW, nomalize the tone to 0db, select a section of the silent part and open it in a spectrum analyser window, ( Audacity has a good one ).

You need to look at the noise levels between 1000 and 10000 Hz, you can do this in audacity by placing the cursor at the required frequency and and reading the noise level on the bottom line.

You can't assess noise levels using the level meter on your DAW, as there is always a lot of very low frequency junk around 10Hz, caused by the internal processor of the recorder, and the level meters will display that, and give a false reading.

Attached is a noise plot of my Tascam DR-40 with the gain set at full.


Audacity can also be used to generate the 1000 Hz tone, but if you are trying to compare two recorders, you will have to make sure that the amplitude of the tone remains exactly the same for both tests.