Recorder

Noise floor of the on-board mics and mic-inputs


Zoom H4, Zoom H4n, Edirol R-4, Sony PCM-D50

The idea


Recently I replaced my Zoom H4 with the new H4n and, of course, I wanted to know the difference between the two, in particular regarding the preamps. Zoom has made strong claims of the new H4n's improved preamps. But how much better are they in reality?
I should add that my Zoom H4 has been modded, i.e. I have added capacitors to reduce the infamous 'beep' sound that occurs when the unit is running on internal batteries.

Because I was fairly unhappy with the results of the new Zoom H4n (and you will know why after reading the text) I am currently replacing it with a Sony PCM-D50. So this recorder as well as my Edirol R-4 (for reference) have also run through all of the tests.

Pricewise, the four units play in somewhat different regions:
Stacks Image 20

Typical European internet price incl. VAT (January 2010)
Price
Zoom H4 (discontinued)

250 €

Zoom H4n

350 €

Sony PCM-D50

580 €

Edirol R-4

1000 €

Setup


I used the three available fixed input settings of the Zoom H4 (Low, Medium, High) as a reference. The internal digital-level setting of this recorder was always set to 100 (flat).
In order to level the other recorders I hooked up a Rode NT4 stereo mic in front of two good speakers, played pretty loud white noise and plugged one channel into the H4 and the other one into either the H4n, the PCM-D50 or the R-4. Then I matched the levels based on the readings of the meters of the different devices as good as possible. This should be accurate to some +/- 1...2 dB.

The Zoom H4n has a record-level range from 1 to 100. The latest firmware (I am using 1.40) offers record-level settings even below 1 (0.9, 0.8, ... , 0.1). However, these settings will just lower the signal on the digital side of the chain. The analog preamps are not reduced below level 1 so all the lower settings are totally useless.

The Sony PCM-D50 has a record-level range from 1 to 10. It also offers a switchable 20 dB pad which I didn't have to use so far.

The Edirol R-4 has four level-knobs which are marked with 11 small bars. I count them clockwise from 1 (coldest setting) to 11 (hottest setting). Close to the XLR-plugs are switches labelled „Mic-Line“. According to the technical specifications they apply an extra 27 dB attenuation.
Record-Level Comparison

Zoom H4

Zoom H4n

Sony PCM-D50

Edirol R-4

Internal Mic

L

10

2

-

Internal Mic

M

45

3.5

-

Internal Mic

H

70

4.5

-

XLR

L

1

2.5

Line-8

XLR

M

30

4

Mic-5

XLR

H

60

5

Mic-7

Compared to the Zoom H4, the new H4n's microphones offer a wider sensitivity range, both on the low end (below 10) and on the high end side (above 70).
When comparing XLR input, the lowest setting is about the same but again the new H4n offers more amplification (above 60). I did not compare the levels of the unbalanced 1/8“ TRS-line inputs.

The Edirol has hotter XLR-inputs than the Zooms. I could not get down to a level comparable to the 'L' setting of the Zoom H4 without using the attenuator switch (Mic-Line) and cranking the preamp up to a setting of 8. That would be about equivalent to a Mic-0 setting if it existed. In the later tests I found out the Edirol preamps would produce quite a lot of noise at this setting (Line-8) so I repeated the tests using the lowest Mic setting (Mic-1) instead.

The Sony PCM-D50 probably accepts the widest signal range. I did not even have to use the 20 dB pad to cover the whole input range of the Zooms. Since the PCM-D50 does not come with XLR inputs I used the unbalanced 1/8" TRS mic input instead.

Tests


I did two series of tests. All were done in 24bit at 44.1kHz which is my favorite recording setup.
Firstly, I placed the internal microphones of the H4 and H4n in a completely silent environment (test#0 -5). What you hear is practically the noise floor only.
The second series of tests (#6-14) are recordings of the XLR (1/8" TRS for the PCM-D50) inputs with either batteries or AC-adapter and 48 V phantom power engaged or disabled:

Test#

Input Gain
Zoom
H4

Rec Level
Zoom
H4n

Rec Level
Sony
PCM-D50

Input Gain
Edirol
R-4

Power
Supply

Phantom
Power

Input
Source

0

L

10

2

-

Battery

-

Mic

1

M

45

3,5

-

Battery

-

Mic

2

H

70

4,5

-

Battery

-

Mic

3

L

10

-

-

AC-Adapter

-

Mic

4

M

45

-

-

AC-Adapter

-

Mic

5

H

70

-

-

AC-Adapter

-

Mic

6

L

1

2

Mic-1

Battery

-

XLR

7

M

30

3,5

Mic-5

Battery

-

XLR

8

H

60

4,5

Mic-7

Battery

-

XLR

9

L

1

-

Mic-1
(Line-8)

AC-Adapter

-

XLR

10

M

30

-

Mic-5

AC-Adapter

-

XLR

11

H

60

-

Mic-7

AC-Adapter

-

XLR

12

L

1

-

Mic-1
(Line-8)

AC-Adapter

48V

XLR

13

M

30

-

Mic-5

AC-Adapter

48V

XLR

14

H

60

-

Mic-7

AC-Adapter

48V

XLR

I used two XLR terminator plugs with 200 ohms impedance and 10 mA phantom current to get comparable results. For the PCM-D50 I also used 200 ohms terminal resistors but, of course, in a 1/8" plug.

Both Zooms showed quite a bit of very low frequency noise, mostly due to SD-card access and probably other internal control circuits. Some recordings also had a notable DC-offset. I applied a 50 Hz 24 dB/oct roll-off filter in Soundtrack Pro to see what difference that would make. It turned out the filtered noise was a lot better. I would definitely recommend using this all the time with the Zooms.

Neither the Sony PCM-D50 nor the Edirol R-4 showed such problems and a low-pass filter makes little difference on these units.

Noise prints were analyzed in Apple's Soundtrack Pro software as well. Each number represents the highest reading of the level meters in dbFS.

unprocessed

50 Hz roll-off

Test#

Zoom
H4

Zoom
H4n

Sony
PCM-D50

Edirol
R-4

Zoom
H4

Zoom
H4n

Sony
PCM-D50

Edirol
R-4

0

-

-79

-81

-83

-

1

-

-64

-73

-73

-

2

-

-62

-60

-58

-

3

-

-

-80

-78

-

-

4

-

-

-64

-70

-

-

5

-

-

-54

-60

-

-

6

-89

-81

-84

-82

7

-80

-79

-83

-76

8

-71

-73

-80

-68

9

-

-80 (-73)

-80

-81

-

-82 (-73)

10

-

-81

-79

-

-76

11

-

-71

-73

-

-67

12

-

-79 (-73)

-81

-80

-

-81 (-73)

13

-

-70

-74

-

-76

14

-

-53

-65

-

-67

All linked files are amplified by 40 dB but otherwise unaltered.

Below is a visual representation of the waveform (top half: left channel, bottom half: right channel) of all tests. All diagrams have the same scale. The horizontal line in the middle of each channel is -infinity. Top and bottom of each channel are -40 dBFS. The length of the clips is about 4 secs.

Each row from left to right shows diagrams for :

Zoom H4

Edirol R-4

Zoom H4n

Test#0: Low gain, internal microphones, battery
Test#1: Mid gain, internal microphones, battery
Test#2: High gain, internal microphones, battery
Test#3: Low gain, internal microphones, AC-adapter
Test#4: Mid gain, internal microphones, AC-adapter
Test#5: High gain, internal microphones, AC-adapter
Test#6: Low gain, XLR-inputs, battery
Test#7: Mid gain, XLR-inputs, battery
Test#8: High gain, XLR-inputs, battery
Test#9: Low gain, XLR-inputs, AC-adapter
Test#10: Mid gain, XLR-inputs, AC-adapter
Test#11: High gain, XLR-inputs, AC-adapter
Test#12: Low gain, XLR-inputs, AC-adapter, 48V/10mA phantom power
Test#13: Mid gain, XLR-inputs, AC-adapter, 48V/10mA phantom power
Test#14: High gain, XLR-inputs, AC-adapter, 48V/10mA phantom power

Conclusions - Part I


The Edirol R-4 is for the most part unaffected by power supply (AC/Battery) or phantom power. However, its noise level grows considerably with increased amplification.

The new Zoom H4n also performs about the same on both AC and battery. But with phantom power switched on it gets much noisier (particularly in the higher sensitivity settings). The same can be said of the Zoom H4 – but keep in mind I am using a modded unit. The original H4 has a severe beep-tone problem when running on battery (a google search will reveal many useful links...). You can still hear the remains of this beep in tests 7 and 8.

One would assume, given its almost three times greater price tag, that the Edirol outperforms the two Zooms easily. Unfortunately this is not the case. The Edirol's noise level is generally not better, especially when you apply a low-cut filter to recordings made with the Zooms. The big difference is the nature of the noise floor. While the Edirol has a uniform, equally distributed high frequency noise which can easily be removed in post production the Zooms both show distinct noise patterns with timely fluctuations. It would be very difficult if not impossible to remove this with a frequency-analysis noise-filter.

In most situations the difference between the old and new Zoom is small. Although the H4n's waveform often looks cleaner, the H4 will become equal as soon as a reasonable low-cut filter is applied. In some situations I liked the sound of the H4's noise better than that of the H4n. There are setups where the H4 clearly has a lower dBFS noise floor reading than its successor.
It is good to know most of the problems the H4 had (beep-sound, phantom power noise) have been resolved in the new unit. But even with the H4n you can still see/hear all SD card accesses and after you've applied a rolloff filter the differences between both devices become almost negligible anyway.

All in all these three devices (Zooms and Edirol) play pretty much in the same league when it comes to preamp noise (except that the Edirol produces a clean hiss and the Zooms produce lots of distinct sounds).
At their hottest settings both the Edirol's and the Zoom H4n's preamps become unusable with a noise floor above -50 dbFS.

The Sony PCM-D50 is a different caliber. With its internal mics the noise floor in absolute terms is not so much different from the Zoom's. But the Sony really shines with an external mic. Then it downright declasses the Edirol R-4 and is also way better than the two Zooms. The Sony PCM-D50 should be perfect for nature and animal/bird sound recordings.

Be sure to read on:
Harmonic distortions. Some pretty interesting results there.