MLXXX

Hi grammefriday, you are aware that two different DACs are involved. However, it is possible some other readers are struggling to follow aspects of this thread. _ _ _ Scenario 1. The RCA output of the CD player is an analogue signal from the DAC inside the player. From what gammefriday says, it appears to be at a greater volume level than his particular amplifier actually needs to produce an adequate listening volume. However, other, less sensitive, amplifiers might need that relatively high signal level. Scenario 2. The coax output of the CD player is a digital stream coming from the transport inside the player. (The DAC inside the player is not used for this connection.) The amplifier coax input detects the stream coming in through the coaxial cable, synchronises to it, and sends the synchronised stream to the internal DAC of the amplifier. (The same process occurs if a CD player has an optical spdif output and that is connected via optical cable to an amplifier optical spdif input.) The fact that the amplifier is giving different volume levels in scenarios 1 and 2 is simply because the signals are being delivered to the amplifier in different formats, one analogue and the other digital. And it turns out that the particular amplifier has gain to spare with its analogue input.

My impressions also, some years ago. However the DAB+ bitrate and/or the streaming bitrate have improved in some cases. As noted earlier in this thread, ABC Classic FM on DAB+ has gone up from 80kbps nominal as it was years ago to 120kbps nominal* today, a very significant improvement. I was a bit surprised to discover this morning how good the streaming rate is now for the Brisbane classical station 4MBS. It is now at 127kbs (AAC, 44.1kHz):- http://online-radio.eu/radio/1028-4mbs-classic-fm It can be nice to listen to a local classical music station with music compiled by people one may know, and presented by people one may know. And hearing news about local concerts and artists. And it can be convenient when driving around town to listen via a car radio tuned to FM, or with later model cars DAB+ (if preferred). The current streaming rate for the national classic music station ABC Classic has also increased compared with what it was years ago. Checking it live a moment ago I found it to be at about 128kbps (AAC, 44.1kHz). It ought to sound slightly better than its DAB+ version broadcast at 120kbs nominal* (AAC, 48kHz). ___________ * The true audio stream bitrate for DAB+ is somewhat lower than the service bitrate you see displayed on your DAB+ radio. Part of the service bitrate is allocated to encoding redundancy, to enable error-free reception to continue in the presence of interference. And, usually, another part of the service bitrate is allocated to low-resolution slides, that some DAB+ radios are designed to display.

ABC Classic for many years was at only 80kbps, but part of that nominal 80kbps was set aside for data for the slide show that some DAB+ radios are able to display. The HE-AAC codec provided a clean sound but of course left out certain detail, which is what psychoacoustic codecs do. The higher the bitrate, the less such a codec leaves out. People vary greatly in their subjective responses. For my hearing, 80kbps (nominal) was not good enough for classical music. It robbed violins and certain vocalists of beauty and excitement. A bit of background hiss and a bit of THD/IMD with FM radio reception were small prices to pay, for me. A 120kbps (nominal) rate is a huge step up. It means that SBR (spectral band replication: bursts of artificial treble) is not needed, just the normal AAC codec can be used. I strongly applaud the decision that was taken to increase the bitrate of ABC Classic. So I'd be inclined to agree that ABC Classic at 120kbps (nominal) could indeed compare quite favourably with the sound from an FM tuner. However I note it would still be less than YouTube audio quality. (I think YouTube have the right compromise - around 130kbps (actual, not nominal) with an efficient codec. The difference between that and lossless is very slight indeed; inaudible for some listeners.) As for 64kbps DAB+ radio in Australia, I hear too many sprays of treble (e.g. when a commentator speaks), and too much loss of body and vitality in the overall sound. It is good enough for talk back radio, marginal for rock music, and seriously lacking for classical music. That's for my ears.

Around town (in Brisbane) I find car radio FM a better experience than DAB+ for a DAB+ service bitrate around 64kbps nominal and definitely better than DAB+ at 48kbs nominal. Having said that, there are times when multipath reception severely compromises the FM reception quality. Out on the highway at greater distances from transmitters hiss increasingly becomes an issue with car radio FM but as you say there is not the problem of sudden complete muting which happens with DAB+ when the signal strength falls below a critical level. There are certainly some excellent streaming stations available for audio quality. Some very high bitrates and/or very efficient codecs. However where an FM radio service in Australia is also available via internet streaming I've found that the bitrate and codec in combination are typically not able to provide as good audio quality as that obtainable from an FM tuner.

I've noticed that modern FM tuners/receivers can deliver a pretty good audio signal to noise ratio even with a weak signal. In the past, an outside antenna was needed to avoid distracting hiss. If you are not too far from an FM transmitter you might be able to get away without an exterior FM antenna. Another option to consider is connecting to an existing external TV antenna. Even if not optimised for the FM band it may give enough signal for a modern receiver to work with. Streaming will give an extremely good SNR (no audible hiss) but as others have commented the audio quality in other respects will typically not be all that good because of the typical use of limited bitrates. DAB+ radio in Australia typically suffers in the same way. However some DAB+ stations offer much higher than usual bitrates and can equal or exceed FM radio subjective sound quality.

Yes I don't think there is too much dispute in relation to a 4K physical Blu-ray disc that it can be expected to hold its own against (if not outperform) 4K streaming, in terms of the picture quality, with current typical 4K streaming rates. But @betty boop has now introduced into this thread the subject matter of how 4K streaming compares with a Full HD Blu-ray. The suggestion that 4K streaming is typically inferior in picture quality to watching a 1080p Blu-ray is quite a contentious suggestion. Arguably, that subject matter merits a separate thread, if there is interest in pursuing it in detail.

That article is almost disinformation, in my opinion. In some cases what is says may hold, but typically it would be wrong. I don't know why there is so much apparent bias against streaming by some writers, whether that bias be due to ignorance, or intentional. I find that in practice with Australian Netflix, 4K streaming looks better than a 1080p Blu-ray for drama feature films, where bitrate is not such an important issue. I watch a great deal of 4K Netflix and I find the added clarity and the vividness of the colour very appealing. This is with a 70'" flat screen TV at about a 3 metre viewing distance. (Our cable NBN has plenty of bitrate capacity, and is quite reliable, so the stream remains at 4K; it does not fall back to lower resolutions and bitrates.) I also note that 4K provides the possibility of HDR whereas 1080p Blu-rays lack that capability. The article does actually make a concession for HDR where it states: One area where 4K streaming has a distinct advantage over standard Blu-ray movies is when it comes to dynamic range. Only 4K UHD Blu-rays support HDR, but it's becoming common with 4K streaming. A good HDR 4K stream will knock the boots of a standard 1080p Blu-ray without HDR. At least when it comes to this particular feature.

I didn't make an assumption. I relied on my memory of what I had read. Netflix don't simply decode an existing Blu-ray and re-encode it at a lower bitrate (a process known as transcoding). They get source material direct. A quick search a moment ago of a Netfix blog yielded this: "To ensure that we have high quality output streams, we need pristine video sources. Netflix ingests source videos from our originals production houses or content partners. ... Our preferred source type is Interoperable Master Format (IMF). " ..." You may be aware that public cinemas don't simply play UHD Blu-ray discs. They have received digital movies in DCP format.

It's a while since I last looked into this but my understanding is that streaming services like Netflix on occasion (or even normally) have access to higher bitrate 4K masters than the versions disseminated on 4K discs. And so when they (or their supplier) encodes for a lower average bitrate for 4K streaming (say 15Mbps) they are working from a higher bitrate source. They are not re-encoding a 4K physical disc. In a lot of movies there are relatively static scenes involving dialogue. In such scenes, bitrate is not as critical. And so streaming can look very similar to the picture from a 4K Blu-ray disc. In action movies, bitrate can be critical, of course. It's interesting watching older action movies where part of the action was created with cgi. You can often notice a blurring and a lot of use of subdued colour, e.g. teal. This was because the computer hardware of the era would have taken too long to render complex action scenes in full detail.

A quick Google search came up with the comparison video below. A lot of the time it shows full frames (sometimes with the left side the 4K streaming version, and the right hand side the 4K Blu-ray version). I'd recommend this video. Although some 4 years old, it's very well done, with plenty of comparisons. The streaming source was Apple TV 4K. A lot of the time there are quite noticeable differences in the side by side comparisons, though as the presenter says this can in some instances amount to no more than a "pixel peeping" exercise, rather than identifying something that would necessarily be noticeable in general viewing. Sometimes differences are not noticeable even when pixel peeping, e.g. for Mission: Impossible – Fallout . I have set the YouTube video below (which is available in 4K resolution) to start where this is presented (at 4 minutes into the video):-

Al, where does this guy describe his methodolgy? Where does he provide objective information about the capture methods employed? For example, what fraction of the original frame widths is he showing at the top in this video? He talks a great deal about his subjective impressions. Looking at this third video even with YouTube set to HD I see screen captures displayed at the top that appear radically different. On the top left is a brighter looking image. On the top right is a darker and softer image. Relative brightness can be a question of HDR settings. But the softness shouldn't be as obvious. It very hard to reconcile with the video you linked to in your opening post. There the differences were quite subtle according to the 3 person panel of observers. (albeit that they did not have side by side comparisons available). I'd be much more impressed with full size still frame captures. Over the years, other people have done this comparing DVDs with Full HD Blu-rays. and with 4K Blu-rays I'd mention that the quality I get on Australian 4K Netflix is nowhere near as bad as this third video would seem to suggest, for a 4K streaming service. (I have to assume the presenter was using 4K services for his comparisons in this third video.) The presenter is obviously passionate, but this video of his woefully lacks technical rigour, in my humble opinion. Perhaps he has posted full size still images somewhere else: there may be copyright exceptions that would apply to allow selected frames to be published on the internet.

You use the words "while clearly not comparing the audio side would have to be a blind freddy i would say not to pick up the quite obvious differences in picture between physical disc vs streamed". Why do you say "streamed"? Is that what the video discussion in the second video you linked to is primarily about? I just sampled a little bit of the video discussion and I heard reference to a Full HD Blu-ray. (There is also reference to a "digital version". I dimly recall low bitrate digital versions of movies being available for playing on mobile phones.) What different versions does the second video compare exactly, @betty boop? It seems that on the one hand is a premium 4K Blu-ray release. It is not apparent to me what the other version(s) is/are. As I say, I have not watched the whole of the second video. * * * My take on the video in the opening post is that it suggested it was very hard to see a difference between live viewing of a 4K blu-ray of a title and 4K streaming of that title. I don't know which streaming service was used for that first video. I am assuming it would have to have been a 4K service streaming a 4K source.

At 4m 55s the woman on the left says comparison would be easier to figure out with the two versions side by side as the differences for her were so minute. That in my experience is also the case with audio. (Quick A B comparisons between a few seconds of a high bitrate lossy version of a recording, and the lossless version of those few seconds, can reveal minor differences that general listening might not.) It's interesting that some directors seem to go out of their way to utilise HDR capability, so you can find part of the streamed picture with shadow detail and another part of the picture with intense lighting. (These sorts of scenes tend to crop up only occasionally if at all. They would never have been attempted in years gone by.) I watch a lot of Netflix and every now and again I can say "aha" that seems to be in 4K (possibly with HDR). And I might then change the picture settings to expose the video bitrate and resolution info in the top left hand corner of my TV set, to confirm that it is indeed 4K. Although I was an early adopter of 4K Blu-ray I find little motivation in 2024 to buy 4K Blu-ray discs, given that 4K streaming comes so close in video quality, and there is such a wide variety of streaming material from which to choose.

Hi Steff, I see that no one else has commented on this post of yours, so I''ll have a go. It does seem to me an odd premise: to make microphone recordings of high level systems for people to listen to with YouTube. Given that YouTube necessarily involves use of a lossy codec at only a moderate bitrate, and also that most people would use their own streaming to watch YouTube (rather than downloading it first before watching it), how could it be expected that subtle differences in streaming performance with the equipment being demonstrated would be noticeable? Only relatively major differences would be noticeable. Are there any differences for your hearing? If so could you advise which of the videos and where? Someone could then extract the YouTube audio at relevant points, for careful A B comparison. There's another thing: to your knowledge were lossless recordings made available as a support for the YouTube presentations? If so, these would be the sources to use for careful A B comparisons. Have I missed something? I'm really at a loss to understand how it was expected that the videos would succeed in demonstrating differences for people watching the videos on their own equipment.

We took delivery of an Echo Studio today (ordered during the Boxing Day sales). Very decent sound for a portable smart speaker, and certainly streets ahead of a small Echo model we already had. The Echo Studio came with a 3.5mm port for connection of either an analogue cable or a mini Toslink optical cable. Initially the Toslink connection wouldn't work. The reason: it wouldn't accept spdif at 96kHz. After changing the source to 48kHz spdif it worked fine! Quite loud too.

I haven't forgotten about providing further detail on the above!! I hope to open a separate topic in coming days, where I would provide links to comparative versions of the same short extract of music (as downloaded from a lossless source, as downloaded from YouTube, and as captured with an ADC from one of Terry's systems).

This week I came across an example of a movie with a very impressive dubbed in language option. It was apparent that great care must have been taken in the choice of the actors who supplied their voices. The movie, Renfield, has recently been released on Netflix Australia It's a vampire movie, with Count Dracula played by Nicholas Cage, and his assistant (Renfield) played by Nicholas Hoult. Cage is a menacing, sometimes whimsical, Dracula. He plays the role with a primarily British, sometimes American, and sometimes European (Transylvanian?) accent. Hoult plays the role with a light, relaxed, British accent. There's lots of blood and gore; and the special effects in the fight scenes are well done. However, I first watched with the dialogue set to French and that offered up a distinctly different experience. Dracula's menacing character is played to perfection. The voice actor gives full expression to the mood swings Cage put into the visual performance, but with different nuances. And Renfield is played very sympathetically (in French). So, I guess I'm reiterating that if you are wanting some experience in hearing a foreign language (perhaps for study purposes, or perhaps just for the sheer entertainment value), there is definitely some quality material on offer by way of alternative language dialogue tracks on video streaming platforms, such as Netflix.

It's a real plus for listening tests with Ethernet networking that the second point Dave mentions above, "same SPL between samples", tends to be much easier to achieve than in many other situations. For example, comparing different DACs usually raises a tricky SPL matching issue because different model DACs typically do not deliver exactly the same output levels.

Sometimes it is easier to keep it simple - it is a title for a discussion thread not academic report. Thanks for responding to this and other points I raised in my last post. I can readily understand the room, the speakers, and the amplifier (with room correction DSP?) being personal choices that audiophiles might well wish to explore. As for the Ethernet network also being an important part of such exploration (in order to avoid or reduce audible impairment from noise), the jury is out on that for many audiophiles.

Yes, your opening post is very good for giving background and providing links for people to follow up. The thread does not provide actual examples of basic Ethernet system,s other than details provided in passing by some of the posters along the way. Technically yes, because you present the following explicit assumption in your opening post: "Assumption There is noise in your audio system that is unintended and it diminishes sound quality. ..." If someone takes those words to heart, they will not query the assumption. They will only write posts that are consistent with the assumption! However the thread title gives the appearance of a much broader scope: "Ethernet System for Audio: putting it all together". Perhaps a different title would have made the assumption more prominent, something like: "Ethernet System for Audio: how to reduce the noise that diminishes sound quality". For links for 4, would you envisage articles in online audio magazines? And perhaps link to videos such as Paul McGowan video? I think though that a lot of the people who write for magazines or make videos merely explain possible mechanisms for noise creation, and possible vulnerabilities of a DAC to such noise or to clock disturbances. * * * @dbastin, If your intention for the thread was for it be focused on bare examples of techniques to combat noise, then perhaps what would satisfy you now would be a consolidated list of hardware and techniques that people have suggested in the course of this thread for noise reduction. (There would be no need for any links to technical material.) You might be well placed to prepare such a consolidated list.

Dbastin, I think key info would include these main categories: 1. At a broad level of explanation, what types of hardware and software are needed to access music streamed from an external source (the internet) and optionally from a local source (notably, Network Attached Storage). 2. At a more detailed level, what are some typical easy-to-set-up and easy-to-use solutions for 1? 3. What special measures can be taken to avoid sound quality degradation? This is more usually expressed in a positive way, e.g., "How can the sound quality be improved?". 4. Explanations why the steps in 3 are usually necessary and usually work. 5. Explanations why the steps in 3 are usually not necessary. I think a newcomer would mostly be interested in category 2 above. @dbastin would you be in a position to prepare a summary of 2 above, given your experience and the fact you started up the thread? I think it could be a really useful practical guide. * * * However, this and similar threads seem to concentrate on 3 above. Also people who offer suggestion for 3, often go on to provide explanations that fall into category 4 above. And such category 4 explanations tends to trigger explanations in category 5!! It is the category 4 and 5 explanations that tend to get highly repetitious. Category 4 explanations could be briefly summarised as follows: Ethernet switches and their power supplies generate electrical hash (and timing irregularities). The hash is transmitted along the Ethernet cable and disturbs the streamer box or similar device by upsetting the steady clocking out of audio buffer contents to the DAC (a form of jitter) and/or directly polluting the analogue output of the DAC with spurious noise. These disturbances are typically audible and typically require special measures to be taken to overcome them. (For example: an audiophile-grade switch, a special noise isolating Ethernet cable or filter, or special low noise power supplies.) It is accepted by many or most audiophiles that the streaming audio buffer can be expected to contain the correct zeros and ones. However some audiophiles believe that the buffer contents retain a history of the noise present in the Ethernet network when the data was loaded into the buffer, so that if the Ethernet network is suddenly disconnected, the adverse effect of Ethernet noise will continue to be audible as the remaining contents of the buffer play out. Category 5 explanations could be briefly summarised as follows: Ethernet networking is a mature technology with government mandated compliance regulations regarding levels of spurious emissions. A well-designed streamer box or similar device can be expected to operate perfectly well in its clocking out of audio buffer contents, irrespective of the typical levels of hash from everyday internet switches and Ethernet cables preceding it in the chain. Also, timing irregularities upstream of the streamer box should not have any audible impact on the analogue output of the DAC. Incoming audio data is error checked before being accepted and added to buffer contents. The memory chips used to store buffer contents are digital and lack the capacity to store "low level background noise": they can only store zeros and ones. As a test, if the Ethernet network is disabled (unplugged and powered down) what continues to play from the audio buffer will be the corruption-free zeros and ones. I hope some of the above may possibly help, Cheers

Terry, thank you again for your hospitality today. You have some truly amazing equipment! * * * Just a quick post to mention that I spent 3 hours today at Terry O's place on the Gold Coast, with Cafad (a very pleasant chap!) in attendance as well. Terry afforded me the opportunity to record a balanced XLR feed from an amplifier in his front room. It may be a few days before I have the opportunity to analyse and report properly on that. (On one occasion the sound seemed louder to my ears but making a quick check a moment ago of my recording, it seems the level didn't actually change! Interesting.) The real delight for me was hearing Terry's setup in his home theatre style room: very quiet and with no distracting reverberation. I had never before in my life heard such a clearly defined sound stage! (I could hear where instruments of the symphony orchestra were located, from left to right, and from to back as well. ) Terry and Cafad were able to hear differences when changes were made to the Ethernet cabling but I must confess I was not sure I could hear any differences. What I can be sure of is that sound quality was extraordinarily good. I was very impressed.

Not in the sense of delivering wrong bits. That is right. Only in the sense of delivering noise pollution at the same time. And if there is an everyday [i.e. non-audiophile optimised] wired Ethernet connection involved for accessing streamed audio, the noise pollution will probably be serious enough to have an audible adverse effect on the DAC output, or so it is said. It is posited that the noise disturbs the evenness of the clocking out (a jitter effect) and/or finds its way directly into the analogue output of the DAC.

What you have outlined amounts to bits not being delivered "correctly". The bits delivered are correct of themselves (no zeros changed to ones or ones changed to zeros), but their delivery --- it is argued --- is accompanied by a disturbance of some kind and that disturbance is attributed to something upstream in the network. Such disturbances, it is reported in threads like this one, commonly lead to an audible impairment in the sound. And so the goal is to reduce upstream disturbances to the point where there is no audible adverse effect on the analogue sound delivered by the DAC. (The well known competing school of thought is that a well-designed buffer will clock out the data bits evenly and without spurious noise, even if upstream the packets of data arrive sporadically, or are clocked unevenly, or if there is electrical noise upstream.)

Just to be clear, the Nordost switch is not designed to alter the audio content, is it? The audio buffer in the streamer box or similar device will end up being fed exactly the same bits, wlll it not?