[bangraman]

Did you read my post? Anti-Shock buffers compress the audio in order to fill the buffer. The compressed audio is then de-compressed in real time. Why do this? Check out the buffer size and the anti-shock times for PCDP's and you'll notice very quickly that it doesn't tally at PCM 16-bit 44.1khz data rates.

 

[raif]

thanks for the info Bangraman. I almost sold my ipod and bought an ihp120 untill I read your comments on its digital out capabilities. It is frustrating how much consumer level gear is just straight sub-standard in its construction, but companies get away with it because of the ignorance of the average customer.

 

[pedxing]

I doubt this is a common issue. While I was testing different sound cards at my last job, I discovered that sometimes the digital outputs of sound cards can be rather poor. I guess the physical link can get messed up with low quality and erratically behaving output jacks. Some low quality sound cards would go out of synch at random times with the SPDIF decoding box.

I don't know what type of protocol these digital outs have on the pcdp. Do they have error correction and error detection schemes?

 

[bangraman]

Quote:

Originally Posted by raif thanks for the info Bangraman. I almost sold my ipod and bought an ihp120 untill I read your comments on its digital out capabilities. It is frustrating how much consumer level gear is just straight sub-standard in its construction, but companies get away with it because of the ignorance of the average customer.

It was a bit of a shame. Notwithstanding all my other issues with the machine, the slightly muddy and 'dead' sound that the iHP exhibited from the headphone and line out seems to be rooted in the digital output capability of the iHP. Once again, the difference between it and the D-E555 in the digital domain was major. I figure a simple thing: If I'm going to use a $600+ digital client, then I should go for a portable that doesn't cramp it's style regardless of the fact that it's a PCDP or DAP, since the whole point of carrying such a digital client around will be for the better sound. Otherwise I might just as well stick to taking the HD25-1 unamped.


It would be good if iRiver could improve the core audio quality of the H series, do some work on that operating system and cure some of it's more glaring physical design problems. Then it has every chance to be a true iPod killer in terms of listening to quality music.

 

[Dane]

Quote:

Originally Posted by bangraman Did you read my post? Anti-Shock buffers compress the audio in order to fill the buffer. The compressed audio is then de-compressed in real time. Why do this? Check out the buffer size and the anti-shock times for PCDP's and you'll notice very quickly that it doesn't tally at PCM 16-bit 44.1khz data rates.

On my PCDP the anti-shock can be disabled (to preserve battery I assume), does this correct the problem you describe? Also, is the compression method used for the anti-shock buffer lossy?

EDIT: OK, reading your posts more closely I assume the problem goes away if ESP is disabled. And from what I understand the compression/decompression is effectively lossy especially on old school PCDP with anti-shock enabled (due to their limited number crushing capabilities). Conclusion: You get the best digital out from an old school PCDP with anti-shock disabled. Right..?

 

[ppl]

Quote:

Originally Posted by bangraman While differences in the abilities of decks to output digital data may be much, much smaller, the ability of the portable transport or the codec chip to output the uncompressed data over optical would seem to contribute most to the change in sound. G-protection for example compresses the data that it's sticking into the buffer on the fly, so with CD players with G-protection permanently enabled (i.e. modern ones) I'm given to understand that you get a reconstituted compressed data stream out of the optical socket. Old-school transports clearly couldn't extract the data cleanly or fast enough, because there's a high degree of digital artifacts from the D-555. The D-E555's ESP turns off automatically when you plug in an optical client so you're getting an uncompressed data stream from the CD. The iHP, iMP, D-NE900, etc are all outputting reconstituted compressed digital data to a degree. This would seem to explain the difference in performance.

It is my understanding that the IHp-120 is bit perfect so no over sampling is done if this is really the case. Since the source if that info is a trusted one and My IHP-120 via aos PPA DAC sounded as good as the coax from my M Audio Audiophile 2496 with the same wave file. also it sounded identical or dam close to the optical output of a Classic JVC Redbook CDP. The JVC did have slightly more loudness. I think when i get some time ill play a MTS audio from DVD onto the IHP hard drive and see if the file plays this would indicate no resembling if it plays. Woodgy has some info on this a while ago

 

[bangraman]

Quote:

Originally Posted by ppl It is my understanding that the IHp-120 is bit perfect so no over sampling is done if this is really the case. Since the source if that info is a trusted one and My IHP-120 via aos PPA DAC sounded as good as the coax from my M Audio Audiophile 2496 with the same wave file. also it sounded identical or dam close to the optical output of a Classic JVC Redbook CDP. The JVC did have slightly more loudness. I think when i get some time ill play a MTS audio from DVD onto the IHP hard drive and see if the file plays this would indicate no resembling if it plays. Woodgy has some info on this a while ago

I expected what you say as well with WAV playback, but I wasn't convinced when comparing with other portable and fixed digital sources. Certainly in terms of compressed audio playback, even with high bitrate files the level of occlusion and 'deadness' became far more of an issue but I also had enough qualms with uncompressed playback through the optical out to feel that it was of debatable merit sticking a $650 upsampling phone on it. The ATH-D1000 does it's own mucking around with the signal so it could have been a freak case of disagreement. I'd be interested in those tests too.

 

[Notorious P.I.G.]

Quote:

Originally Posted by bangraman Did you read my post? Anti-Shock buffers compress the audio in order to fill the buffer. The compressed audio is then de-compressed in real time. Why do this? Check out the buffer size and the anti-shock times for PCDP's and you'll notice very quickly that it doesn't tally at PCM 16-bit 44.1khz data rates.

I apologize. I misunderstood your post to mean that the optical out was sending compressed data.

Yes, makes sense now.

 

[bangraman]

Quote:

Originally Posted by Dane On my PCDP the anti-shock can be disabled (to preserve battery I assume), does this correct the problem you describe? Also, is the compression method used for the anti-shock buffer lossy? EDIT: OK, reading your posts more closely I assume the problem goes away if ESP is disabled. And from what I understand the compression/decompression is effectively lossy especially on old school PCDP with anti-shock enabled (due to their limited number crushing capabilities). Conclusion: You get the best digital out from an old school PCDP with anti-shock disabled. Right..?

Is the compression lossy? I don't know*. But (and this is pure speculation) it is something of a coincidence that ESP came on to the market as Minidisc was gradually being rolled out.


*EDIT: As P.I.G. says below... and in thinking about it, not just what he says but I don't think a mathematical lossless compression in realtime would have yielded the compression rates required for this use given the availablility of ASICS / DSP's for this purpose back then.

 

[Notorious P.I.G.]

Quote:

Originally Posted by Dane On my PCDP the anti-shock can be disabled (to preserve battery I assume), does this correct the problem you describe? Also, is the compression method used for the anti-shock buffer lossy?

Given that the sampling is done at just above the Nyquist rate, then the compression would almost have to be lossy.

 

[Dane]

Quote:

Originally Posted by Notorious P.I.G. Given that the sampling is done at just above the Nyquist rate, then the compression would almost have to be lossy.

Errm, not sure I follow. It seems like you're talking about analog to digital conversion...? I strongly assume that the path from disc to digital out on a PCDP is purely digital. The question is whether the information on the disc is somehow modified in this process - I never thought that this could happen with a disc in relatively good condition (PCDP error correction capabilities is probably horrible).

The buffer stuff is concerning although I dont quite understand how a buffer mechanism could loose bits - for power saving reasons I suspect the compression method used to be quite simple (something like saving sample diffs instead of absolute values).

On the other hand, to fill the buffer, the transport must sometimes run above normal speed, which could introduce more errors. That way it isn't the buffer itself that causes the errors but rather the way it's filled... blah, blah, .... OK, I'm rambling

 

[Coalescent]

bangraman,

In your listening tests involving the iRiver iMP-550's optical out, did you have the anti-shock system disabled?

It's part of the subsystem that switches between anti-skip, etiquette, and normal (uncompressed, presumably) functionality. It is toggled by pressing the MODE/CD<->FM/STEREO dial to the right using what the manual calls a "long press." It changes a small icon on the display to indicate one of the three modes.

Thanks,

-Chad

 

[bangraman]

Ah. Now that you mention it, no I didn't. I'll figure out whether that mode is unbuffered / compressed.

 

[AdNauseam]

Quote:

Originally Posted by Notorious P.I.G. Given that the sampling is done at just above the Nyquist rate, then the compression would almost have to be lossy.

I don't understand what you mean by this. The PDCP shouldn't resample the signal for digital out and probably doesn't do so for compression. Also, oversampling can be used to prevent errors from being introduced.

Dane:

In order to make the buffer on a CD player last longer the data is compressed. To get a good level of compression without using a lot of processing power the player pretty much has no choice but to use lossy compression and discard data. The buffer itself doesn't lose data(assuming it's working correctly) but it does contain compressed data.

 

[Notorious P.I.G.]

Quote:

Originally Posted by AdNauseam I don't understand what you mean by this. The PDCP shouldn't resample the signal for digital out and probably doesn't do so for compression. Also, oversampling can be used to prevent errors from being introduced..

There isn't any resampling. The original sampling is done at just above the Nyquist criterion--thus, data loss from a dynamically sampled signal is inevitable.

Quote:

Originally Posted by AdNauseam Dane: In order to make the buffer on a CD player last longer the data is compressed. To get a good level of compression without using a lot of processing power the player pretty much has no choice but to use lossy compression and discard data. The buffer itself doesn't lose data(assuming it's working correctly) but it does contain compressed data.

Lossy compression=lost data (aka lossy)