Samsung flash-based mp3-players sound quality issue (YP-T5, T5X, T5H, T5V, YP-55 confirmed, T6, maybe others possible)

I decided to make this research after I tested the sound quality of my newly purchased Samsung Yepp T5X it was almost unbearable. If you find this research is of use to you, then you may use it for your own research

Note: Initially the review was based on the experience with one particular YP-T5X model, but when I saw the exact same result testing the YP-55 player, I decide to address the issue on all Samsung flash-based players. If you find any of my research is incorrect, I would appreciate any input that you feel would be beneficial, in order to rectify errors which may be shown

Table of contents

Description
Research Method
Measurements Results
Results with RMAA Software
Conclusion

Description

I have been using different portable audio devices for some time now. First was Sharp MD-MS701H, minidisc player, then the Eratech Beatsounds EMP-Z, which has a coin-sized shape and claimed to be the smallest in the world (as far as I am aware it is stilll is!). The latter had none of the special features nor even an LCD screen for that matter, so "the smallest", but is also "the simplest".

Having seend the Samsung Yepp YP-T5 (T5V, T5H, T5X), I was impressed by the size and features on the unit the body volume being only twice that of the EMP-Z. Having the body volume just twice of the EMP-Z, and powered by one AAA cell and sported all the modern features of flash-based portable players, such as FM-Tuner, direct MP3 recordings, voice recorder and so on. There was no doubt I would buy it.

Before purchasing my last test being the most important, sound. The maximum volume was not impressive but EMP-Z had similar. The sound was a little strange, lacking in highs. Anyway, this didn't stop me going ahead and buying it. I assumed there would be something in the settting which would allow some adjustments on custom settings.

However swithching to 'Normal equalizer' mode did not solve the problem. I needed something more concrete with confimrations from 'real measurements', rather than allow the matter to stay "as is". The problems appeared to lie in the highter frequences having a lower level than mid and low. I should add that I'm talking about my personal feelings based on previous experience using the same headphones with my new unit.

After some thought my chosen methods of research, which are by no means unique, but will give other customers like myself with few technical skills, a advantage.

Research method

There is a special testing tone, often used for audio equipment testing, tone sweep sometimes found on the testing discs bundled with Hi-Fi magazines. This is special wave of sound which gradually changes frequency from one value to another. Generally the tone from 20 Hz to 20 kHz is being used since this bandwidth is considered virtually audiable by the human ear. When this test is used for acoustic devices, such as acoustic systems or headphones, it allows you to compare what the frequences the device produces and just how well it does it. Because of different hearing conditions and the different hearing abilities of an individual, this technique is probably not the best for an objective analysis.

So, these are the steps I used for my method

Lets generate a tone sweep in any good sound editor, lets assume start frequency will be 50 Hz and end frequency 6000 Hz. For our goal the essential parameter is amplitude, the amplitude should be the same through the whole fragment.
The next step is to encode the wave into MP3 file in order to upload it later to the device. To be very accurate, let the bitrate be 256 kbits.
In all sound devices any equalizing should be turned off, otherwise there's no use of testing at all. For Sharp we switch off Bass Boost. The EMP-Z has no equalizer at all. Samsung is more complex, under the items "Sound" you will find Normal, Classic, Jazz, Rock, DBB (Dynamic Bass Boost), User, SRS, TruBass, WOW. They work as 'exclusive' so one and only one of these items can be selected at any one time. Let's assume the 'Normal' mode is what we need.
After we upload our testing mp3 file to all devices that support mp3 (Beatsounds EMP-Z, Samsung Yepp T5) and record onto the others (minidisc portable player Sharp MD-MS701H and minidisc deck Sony MDS JE510).
One by one we connect our devices to a sound card and record the fragments as digital files.
We compare the result. For a good player the amplitude of the new recorded fragment should stay constant. In other words, we measure the frequency response for each player. If you're not familiar with the term, query the sentence in a search engine, or read an article at pcavtech.com I did the same recently :-)

Note: ADC (Analog-to-Digital Converter) quality. It looks like we would need a very good sound card to to be accurate when converting output sound from the player to the computer, but for our purpose it is only partly correct. A bad soundcard may tend to convert sample values innacurately. If so, we would continue to see a tendency in amplidute change as one particular sample may be incorrectly digitized, either the other samples would then be converted with the same value or all values will have stochastic nature and we will notice this. I may be wrong, anyway I used very good quality equippment, a Sony MDS JE510 deck, connected by optical cable to the computer.

Measurements results

Unfortunately it is not possible to show the waves in original scale since the pictures would be very long in this case. So scaled down picutres will be shown and the ideal result wave I mentioned before should look like a simple rectangle. If there are shifts or deviations, your eyes will notice this as a slightly different figure.

I would appreciate your comments should your findings differ from mine (maxerist@gmail.com).

Sharp MD-MS701H Very good results, the amplitude value �3.05 �� stayed almost the same through the whole fragment. The deviation value is about -0.02 + 0.002 dB

Beatsounds EMP-Z. There is a little gradual increase from lows to highs, but this increase measuered in 0 to 0.06 dB, I'd considered it a good result

Sony MDS-JE510. The only non-portable hero here. Moreover, it is a Analog-to-digital stations for all other tests, so when I decided to measure its values, the Sharp unit became a soundcard. Headphones output was chosen for getting the wave. The deviation is very small too.

Samsung Yepp YP-T5. The main result. Take a look at the shape on the right, can you call it a rectangle? I guess you could. You can see it is exaggerated in the mid frequency and what is interesting, the shape is a very natural outline, not random at all. The difference between the amplitude values at the maximum and minimum points is nearly 1.65 dB. Is it a large or small difference? If your look at the same artice at pcavtech.com, you'll see that the authors considered 1 dB as sometimes a noticable change.

What of the other equalizer items in the Sound menu. May they be used to make the sound more straight? Unfortunaetly the answers is a resounding NO. Look at the shapes in the diagrams below. They show the measured shape for every item I mentioned before (Normal, Classic, Jazz, Rock, DBB (Dynamic Bass Boost), User, SRS, TruBass, WOW). Neither allows you to make the 'high frequences' higher or even the same as middle. The so called 'user preset' has 'Treble' parameter inside, but affects only 'mid frequencies'. Settting Low and Treble to zero will not produce higher 'high frequences'.

A little more accurate results with RMAA software

There is a great software program intended for the testing purposes, called RMAA ( Rightmark Audio Analizer). They use a more accurate approach and so I give you the values that came from the software

RightMark Audio Analyzer v.5.2 test. Testing chain: External loopback (line-out - line-in). Sampling mode: 16-bit, 44 kHz Summary

Test	Sharp MD-MS701	BeatSounds EMP-Z	Samsung Yepp T5X
Frequency response (from 40 Hz to 15 kHz), dB:	+0.61, -0.49	+0.33, -0.32	+1.32, -0.78
Noise level, dB (A):	-91.9	-93.6	-91.7
Dynamic range, dB (A):	91.4	85.5	89.0
THD, %:	0.016	0.0090	0.0061
IMD, %:	0.029	0.044	0.024
Stereo crosstalk, dB:	-79.2	-91.9	-91.1

Frequency response

Spectrum graph

As you may see, RMAA shows a little less values for the frequence response, 1.32 dB. (mine was 1.65), but anyway this result is much worse than those of other players. Just look at the blue line at the image above!

Quick conslusion

My general feeling is that there's something wrong with this player. If there's any modern 3d-sound chip inside or not. I care for one thing: All players should have the ability to produce unchanged sound at the output jack and not left to company standarts.

Some people acutally don't know what the term frequency response means. There are many sites which show "Frequency response 20-20000 Hz", but in this case they had to call it 'Frequency bandwidth'. If you look at the same article at pcavtech, the point is that "Frequency response is specified by the combination of three or four numbers". You will not find other two values for Samsung YP-T5, YP-T5V, YP-T5H, YP-T5X, YP-55 anywhere at their website. And I suppose that Samsung just didn't provide us with the information.

Please let me know if you found this article useful, additionally I will be glad to get any comments on this article (maxerist@gmail.com).

There is also a russian version of this article. I have advised the local representatives of Samsung about this issue and they have said that they 'might' contact the engineers, but I have had no other comments from them at this time. I hope that this article will be read by someone from the Samsung who will take steps to change something in the firmware so the player will have a decent sound and this will be benefitical to the individual and to the reputation of their firm. If Samsung do respond to my observations, I will add their comments to this page and any references to downloadable files or documentation they deem suitable.

I also would like to give my credits to the Beatsounds EMP-Z player. Not only still the 'smallest in the world', but also one having the very respectable sound, good job, Beatsounds! By the way, I have noted many rebranded version of their unit, including Virgin label this facts speak for themselves...

If you want to make a similar test with your player, you can use the same tone, I used. To download a 180k file, click here. It would be great, if you share your results with me, so I may add a new test data to this page (maxerist@gmail.com)

	Sharp MD-MS701H Very good results, the amplitude value �3.05 �� stayed almost the same through the whole fragment. The deviation value is about -0.02 + 0.002 dB
	Beatsounds EMP-Z. There is a little gradual increase from lows to highs, but this increase measuered in 0 to 0.06 dB, I'd considered it a good result
	Sony MDS-JE510. The only non-portable hero here. Moreover, it is a Analog-to-digital stations for all other tests, so when I decided to measure its values, the Sharp unit became a soundcard. Headphones output was chosen for getting the wave. The deviation is very small too.
	Samsung Yepp YP-T5. The main result. Take a look at the shape on the right, can you call it a rectangle? I guess you could. You can see it is exaggerated in the mid frequency and what is interesting, the shape is a very natural outline, not random at all. The difference between the amplitude values at the maximum and minimum points is nearly 1.65 dB. Is it a large or small difference? If your look at the same artice at pcavtech.com, you'll see that the authors considered 1 dB as sometimes a noticable change.