[TooPoorForHiFi]

I just tried it — works for me automatically on Galaxy S9. I connected an OTG cable from the phone to ES100 and it started playing through USB without doing anything else. Maybe make sure that your phone isn't prioritizing Bluetooth connection (try disconnecting Bluetooth or turn it off).
He is trying to use the ES100 as a wired DAC, not Bluetooth.

Do you have your ES100 ON while plugging it in?

 

[rkw]

Do you have your ES100 ON while plugging it in?

It works both ways for me. ES100 could be on or off when I connect the OTG cable. Once I start playing audio on the phone (Galaxy S9), I hear it through the ES100.

 

[TooPoorForHiFi]

It works both ways for me. ES100 could be on or off when I connect the OTG cable. Once I start playing audio on the phone (Galaxy S9), I hear it through the ES100.

Thanks, I think I got it. Phone issue.

Everything works on someone elses Phone.

 

[CactusPete23]

Thanks, I think I got it. Phone issue.

Everything works on someone elses Phone.

Might work better with USB Audio Player Pro (UAPP) App... UAPP gives cleaner signal, and bypasses any android resampling. Can try the free version to see if it works with your phone.

 

[zolom]

Recently (I guess after installing FW version 1.3), when turning my ES100 on, it connects to my LG V30+ BT and can play audio all right, but occasionally, the android application does not seem the recognize the connection. I have to perform "Device Search" and only after selecting the ES100, the app connects. I did reboot, clear cache, to no avail.

Please advise.

Thanks

 

[SubMash]

Recently (I guess after installing FW version 1.3), when turning my ES100 on, it connects to my LG V30+ BT and can play audio all right, but occasionally, the android application does not seem the recognize the connection. I have to perform "Device Search" and only after selecting the ES100, the app connects. I did reboot, clear cache, to no avail.

Please advise.

Thanks

Same story from the beginning. But before it would also not let it connect. Now it connects, although I have no idea which device is which - I have 2 and there are no names. But it still loses connect from time to time.

 

[SubMash]

Multitone is excessive for simple adpcm type codecs.

997 is prime for a reason, it makes a standard test tone that avoids resonances with digital sample rates.

I agree that AptX is particularly bad at HF, definitely worth a sweep test, although we're testing codec noise artifacts, they probably need to be coding more than one sine wave at a time for a chance at a fair comparison.

Information complexity is relevant for a codec test, but purity of signal is required, so it needs to be a generator.

I think both the sweep and multi-tone.

 

[PiSkyHiFi]

Multitone is excessive for simple adpcm type codecs.

I don't know about lossy audio compression dealing with a single tone - a case they literally never have to deal with unless your music taste is unique, but with regular digital compression, such a case where it is a simple pattern repeated, the encoding will reduce the file size to something extremely small and will reproduce perfectly from this massive compression ratio.

It's bad form to test a codec by such a unique case - it probably reveals only a very small number of real world cases and won't give a decent comparison of how the codecs will respond to many waveforms combined, (you know - music)

I think this is fairly obvious, honestly, I don't know why the tests were done that way, it most likely doesn't reveal much about their relative merits as digital audio compressors.

The only thing I managed to glean from these tests is that the absolute noise floor is quite low for all of them - which is cool.

Edit: What do you think a lossy compressor does a DCT on this sine wave is going to look like?
a single peak in the frequency band to then represent.

Your idea of a sweep is better than a single tone, but at least multi tone (which i can confirm this device can test) represents something akin to music.

It might not make a difference because it still might not be enough to trigger the trickier parts of the codec, I imagine the only real way to compare the noise of codecs is to use enough combined waveforms that the codec has to make some touch decisions what to throw away . The comparison will come from how well they can make those tough decisions.

 

[PiSkyHiFi]

Same story from the beginning. But before it would also not let it connect. Now it connects, although I have no idea which device is which - I have 2 and there are no names. But it still loses connect from time to time.

On my V20, this never happens - sometimes, in the car, it cuts out after connecting for a minute (only ever after one minute - never any other time), suddenly playing the audio on my phone while it thinks it's playing on the Bluetooth, but I have a fix for it, which is a usb cable with an inline power switch in the car - allows me to reset it very easily while driving. My phone is in my pocket the whole time, so I put it down to signal strength and a bump in buffer management.

My other one I use for portable headphone use has never had this issue at all - even with the phone in my pocket.

Maybe I'll swap them over for a while.

 

[blomman77]

Hello
I have a question. How is output level estimation supposed to work?
No matter the IEM i connect to Radsone the app says 16 ohm and 105db
and if i try to change the value to the right specification for that IEM i am not able to save that?
Can someone help with this? Thanx

 

[SubMash]

ADPCM is not smart enough to compress better if there is a single tone or lower complexity - it always compresses to the exact same number of bits and doesn't change bandwidth for subbands based on the signal.
It just loses details on exact frequency, but it doesn't care if any other sub-band has any signal or not. So multi-tone for that is useless. Again, all those codecs (except AAC) are Low Power - they are extremely simple algorithms.

Truth is that (as wslee also stated) - you can't compare AAC to ADPCM codecs. AAC looks way worse, but in reality, you can't hear that noise due to how ear and brain work.
But if you compare SBC, APTX, LDAC - that would be fair to just look at the noise only. In those charts, in the app, you can see both - codec noise and hardware harmonics. Harmonics are very hard to hear, btw.

I tested DCT - it adds small masking noise and shifts FR 0.3dB here and there - I'm not a psycho-acoustic expert to say if it really works or not. It does make noise and somebody might like it, but I'm dead sure most people can't hear it in a blind test at all.

I don't know about lossy audio compression dealing with a single tone - a case they literally never have to deal with unless your music taste is unique, but with regular digital compression, such a case where it is a simple pattern repeated, the encoding will reduce the file size to something extremely small and will reproduce perfectly from this massive compression ratio.

It's bad form to test a codec by such a unique case - it probably reveals only a very small number of real world cases and won't give a decent comparison of how the codecs will respond to many waveforms combined, (you know - music)

I think this is fairly obvious, honestly, I don't know why the tests were done that way, it most likely doesn't reveal much about their relative merits as digital audio compressors.

The only thing I managed to glean from these tests is that the absolute noise floor is quite low for all of them - which is cool.

Edit: What do you think a lossy compressor does a DCT on this sine wave is going to look like?
a single peak in the frequency band to then represent.

Your idea of a sweep is better than a single tone, but at least multi tone (which i can confirm this device can test) represents something akin to music.

It might not make a difference because it still might not be enough to trigger the trickier parts of the codec, I imagine the only real way to compare the noise of codecs is to use enough combined waveforms that the codec has to make some touch decisions what to throw away . The comparison will come from how well they can make those tough decisions.

 

[SubMash]

Music connection is rock solid on ES100. But sometimes app shows that device is disconnected and you need to connect to it again. Might have something to do with few audio BT devices available.

On my V20, this never happens - sometimes, in the car, it cuts out after connecting for a minute (only ever after one minute - never any other time), suddenly playing the audio on my phone while it thinks it's playing on the Bluetooth, but I have a fix for it, which is a usb cable with an inline power switch in the car - allows me to reset it very easily while driving. My phone is in my pocket the whole time, so I put it down to signal strength and a bump in buffer management.

My other one I use for portable headphone use has never had this issue at all - even with the phone in my pocket.

Maybe I'll swap them over for a while.

 

[SubMash]

Hello
I have a question. How is output level estimation supposed to work?
No matter the IEM i connect to Radsone the app says 16 ohm and 105db
and if i try to change the value to the right specification for that IEM i am not able to save that?
Can someone help with this? Thanx

Saves for me and changes calcs.

 

[PiSkyHiFi]

ADPCM is not smart enough to compress better if there is a single tone or lower complexity - it always compresses to the exact same number of bits and doesn't change bandwidth for subbands based on the signal.
It just loses details on exact frequency, but it doesn't care if any other sub-band has any signal or not. So multi-tone for that is useless. Again, all those codecs (except AAC) are Low Power - they are extremely simple algorithms.

Truth is that (as wslee also stated) - you can't compare AAC to ADPCM codecs. AAC looks way worse, but in reality, you can't hear that noise due to how ear and brain work.
But if you compare SBC, APTX, LDAC - that would be fair to just look at the noise only. In those charts, in the app, you can see both - codec noise and hardware harmonics. Harmonics are very hard to hear, btw.

I tested DCT - it adds small masking noise and shifts FR 0.3dB here and there - I'm not a psycho-acoustic expert to say if it really works or not. It does make noise and somebody might like it, but I'm dead sure most people can't hear it in a blind test at all.

Well, I don't know about AptX, but if it is ADPCM, which I do know about, then the key word is the first letter of the acronym - Adaptive.

Adaptive means that the difference between samples can adapt in representation bits - it's an adaptive quantiser, plus all the other aspects like predictors feeding back - this will allow a single tone to be efficiently represented, based on only a small or non-existant pattern for the next block in terms of difference from the previous block.

It's actually a lot more complicated than you describe, unless you can show me that AptX is not ADPCM based.

 

[SubMash]

It's in the wiki - "design is based on time domain ADPCM principles without psychoacoustic auditory masking techniques" And bits allocation per subband is fixed.
They only have a dynamic range adjustment so they encode low volume tone without high volume tones more effective.

It's based on this patent https://patents.google.com/patent/EP0398973B1 - if you see here anything that would show that multi-tone test is more useful - let me know, I could be missing something.
In my understanding, 2 tones with the same amplitude in different subbands would be exactly the same.

Well, I don't know about AptX, but if it is ADPCM, which I do know about, then the key word is the first letter of the acronym - Adaptive.

Adaptive means that the difference between samples can adapt in representation bits - it's an adaptive quantiser, plus all the other aspects like predictors feeding back - this will allow a single tone to be efficiently represented, based on only a small or non-existant pattern for the next block in terms of difference from the previous block.

It's actually a lot more complicated than you describe, unless you can show me that AptX is not ADPCM based.